Unit 4 Tutorials: Statistical Inference

INSIDE UNIT 4

Foundations of Hypothesis Testing

Sample Selection
Point Estimation
Sampling Distributions
Sampling Error
Confidence Intervals - Means
Confidence Intervals - Proportions

Hypothesis Testing

Hypothesis Testing Basics

One-Sample Hypothesis Testing - Means
One-Sample Hypothesis Testing - Proportions
Two-Sample Hypothesis Testing for Means: Two-Tailed Test
Two-Sample Hypothesis Testing for Means: Right-Tailed Test
Two-Sample Hypothesis Testing for Means: Left-Tailed Test
ANOVA

Sample Selection
by Sophia

 WHAT'S COVERED

In this lesson, you will learn about the four main types of sampling that can be applied in business data
analytics scenarios. Specifically, this lesson will cover:
1. Introduction to Sampling
1a. Random Sampling

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 1
 1b. Systematic Sampling
1c. Stratified Sampling
1d. Cluster Sampling
1e. Difference Between Stratified and Cluster Sampling

1. Introduction to Sampling
When you are working with data, it is often impossible or impractical to collect information from every single
person or item in a group. This is where sampling comes in handy. Sampling is like taking a small piece of a big
cake to understand what the whole cake tastes like. In business data analytics, sampling helps us make
predictions and decisions without needing to look at every single piece of data.

There are several reasons why sampling is used in business data analytics, such as:

Time and Cost Efficiency: Collecting data from everyone can be very expensive and time-consuming.
Sampling saves time and money.
Manageability: Handling a smaller set of data is easier and more manageable.
Accuracy: If done correctly, a sample can give us a good estimate of the whole group.

Let’s start exploring several widely used sampling techniques.

1a. Random Sampling

Random sampling is a way to select a small group of items or people from a larger group, also known as the
population, in such a way that every item or person has an equal chance of being chosen. This method is very
important in business data analytics, because it helps us make fair and accurate conclusions about the whole
population without having to look at every single item or person.

Imagine you want to know what kind of products customers like the most in a big store. Checking every
customer’s preference would take too much time and effort. Instead, you can use random sampling to pick a
smaller group of customers to ask. If done correctly, this smaller group will represent the preferences of all
customers.

Suppose you were studying customer preferences and wanted to create a random sample. Follow these steps:

 STEP BY STEP

1. List the Population: First, make a list of all the items or people in the population. For example, if you are
studying customer preferences, your list might include all the customers who visited the store in the
past month.
2. Assign Numbers: Give each item or person a unique number. This helps in keeping track of who or
what is being selected.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 2
 3. Use a Random Method: Use a random method to pick numbers from your list. This can be done with
Excel.
4. Select Your Sample: The items or people that correspond to the randomly picked numbers form your
sample. This sample should be a good representation of the whole population (that is, all customers).

Let’s take a look at how you could create a random sample in Excel.

EXAMPLE Suppose you have 100 customer transactions from a store. The first five observations are
provided below.

To take a random sample of 10 customers from this data set, you can use the Excel RAND() function. The
RAND() function in Excel generates a random decimal number between 0 and 1. Each row in your dataset
gets a random number, ensuring that every row has an equal chance of being selected. Once each row has
a random number, sort the dataset by that column and then select the top 10 rows as your random sample.
A possible random sample is shown below:

 TRY IT

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 3
 Use Excel to find a random sample of 10 customer transactions from the previous example.

Using the data in the customer_preferences_random_sample.xlsx file, perform the following:

1. Create a new column by entering Random_Number in cell E1.

2. Enter the following formula in cell E2:

=RAND()

You should see a random number from 0 to 1 populate cell E2. Hover over the small black plus sign (Auto-
Fill Handle) at the bottom-right corner of cell E2. Click and hold the left mouse button, then drag the handle
down to the last cell where you want the formula (cell E101). Each row will receive a random number
between 0 and 1.

3. Create a new column by entering Random_Number1 in cell F1. Copy the values in the Random_Number
column. Go to the Random_Number1 column and paste the values into the Random_Number1 column by
selecting Paste Special --> Values Only.

The RAND() function recalculates every time Excel performs a calculation, including when you copy and
paste data or sort the data. When you perform this copy and paste step, the values in the Random_Number
column will change from their original values.

4. Sort the Random_Number1 column in ascending order (smallest to largest) by selecting the
Random_Number1 column. In the Home menu, select the drop-down arrow beside the Sort and Filter icon
and select Sort Ascending. Select Expand and Sort to apply the sort to all rows.

Then, you can select the top 10 rows to obtain the random sample.

When you work through these steps, you will obtain a different random sample from what is shown in the
example, due to the random nature of the RAND() function.

 TRY IT

In the Excel file customer_service_reviews.xlsx, there are 30 customer satisfaction scores and reviews for a
local restaurant.

Take a random sample of 5 customer reviews. +

One possible solution might look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 4
 The customers selected for your random sample will be different because of the random numbers
generated by the RAND() function each time the function is used.

Random sampling helps ensure that every member of the population has an equal chance of being selected,
which makes the sample more representative of the entire population. Results from a random sample can often
be generalized to the entire population, providing valuable insights for decision-making.

1b. Systematic Sampling

Systematic sampling is a method of selecting a sample from a larger population in a systematic, orderly way.
Instead of randomly picking individuals, you select every kth individual from a list or sequence. The value of k is
determined by dividing the population size by the desired sample size. For example, if you have a population of
1,000 and you want a sample of 100, you would select every 10th individual.

Follow these steps to create a systematic sample.

 STEP BY STEP

1. Define the Population: Identify the entire group you want to study.
2. Determine the Sample Size: Decide how many individuals you need in your sample.
3. Calculate the Sampling Interval (k): Divide the population size by the sample size.
4. Select a Random Starting Point: Choose a random starting point within the first (k) individuals.
5. Select Every kth Individual: From the starting point, select every kth individual until you reach your
sample size.

Let’s look at an example of how you can create a systematic sample in Excel:

 TRY IT

Your company wants to analyze customer satisfaction based on survey responses. You have a list of 1,000
customers who completed the survey, and you want to select a sample of 100 customers for detailed
analysis.

Create a systematic sample size of 100 using the customer_satisfaction_systematic_sample.xlsx data file
by performing the following:

1. Create a new column by entering Systematic_Sample in cell E1.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 5
 2. Calculate the sampling interval by dividing the total number of customers by the desired sample size.

3. Generate a random starting point. In an empty cell, enter the formula:

=RANDBETWEEN(1,10)

This Excel formula will generate a random number between 1 and 10. Suppose the random number is 7. This
will be your starting point.

4. Select every kth customer. Starting from the 7th customer, select every 10th customer. In cell E8 (this is the
7th observation), enter the formula:

=MOD(ROW(A7)-7,10)

Let’s break down this Excel formula.

ROW(A7): This function returns the row number of cell A7. In this case, it returns 7 because A7 is in the
7th row.

ROW(A7) - 7: This part subtracts 7 from the row number. For cell E8, it calculates .
MOD(ROW(A7) - 7, 10): The MOD() function returns the remainder after dividing the first argument by the
second argument.

The Excel function =MOD(ROW(A7)-7,10) will resolve to =MOD(0,10) and this will equal 0. The MOD() function
in Excel returns the remainder after a number is divided by a divisor. When you use =MOD(0,10), you are
asking for the remainder of 0 divided by 10.

When you divide 0 by 10 the result is 0 because dividing 0 by any number always gives 0. Since

there’s nothing left over after the division, the remainder is also 0.

For the next row, cell E9 (the 8th observation), the Excel function =MOD(ROW(A8)-7,10) will resolve to
=MOD(1,10) and this will equal 1. The Excel function =MOD(1,10) will return the remainder of 1 divided 10

10 cannot go into 1 even once, so the entire 1 is left as the remainder.

Modulo arithmetic deals with the remainder when one number is divided by another. For example, in

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 6
 modulo arithmetic is 2, because 7 divided by 5 leaves a remainder of 2. In Excel, the =MOD() function

works the same way, so =MOD(7,5) would return 2, just like the long division example.

You will see in the 7th row, and then for every 10 rows, there is a 0 indicating that is the row that should be
selected for the sample.

5. To obtain the systematic sample, filter the data for rows where the value in the Systematic_Sample
column is 0. Select the Systematic_Sample column. In the Home menu, select the drop-down arrow beside
the Sort and Filter icon and select Filter. Select the drop-down arrow beside the Systematic_Sample column
name in cell E1, deselect the Select All checkbox, and select the 0 checkbox. Select Apply.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 7
 The worksheet will update to only include the rows where the column Systematic_Sample is 0. Remember
when you filter in Excel, the rows that do not meet the filtering criteria are hidden. It is necessary to ensure
you are only working with values where Systematic_Sample is 0. Next, you will copy and paste these values
to a new worksheet.

6. After applying the filter, select all the visible rows (including the column names). Right-click and choose
Copy, or press Ctrl + C.

7. Add a new worksheet by clicking the + icon at the bottom of the Excel window.

8. Click on the first cell (A1) in the new worksheet. Right-click and select Paste Special --> Values Only.

The first five rows of the systematic sample are shown below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 8
 9. Check your work by comparing your systematic sample to the one provided in this example.

 TRY IT

Using the banking_transactions_systematic_sample.xlsx file, create a systematic sample of 25

transactions.

Solution:

Suppose the random number generated from the RANDBETWEEN() function was 15. Starting at the 15th
transaction, select every 20th transaction. In cell E16, enter the formula:

=MOD(ROW(A15)-15, 20)

The first five rows of the final systematic sample are shown below.

 WATCH

Check out this video about creating a systematic sample size of 25.

 TERM TO KNOW

Systematic Sample
A type of sampling method where every kth element is selected.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 9
 1c. Stratified Sampling
Stratified sampling is a sampling method used to ensure that different groups within a population are
represented fairly in a sample.

Imagine you are a business analyst working for a company that sells products to different age groups. You want
to understand how each age group feels about your products. If you just pick a random sample of customers,
you might end up with too many people from one age group and not enough from another. This is where
stratified sampling comes in handy.

Follow these steps to create a stratified sample:

 STEP BY STEP

1. Divide the Population into Strata: First, you divide your entire population into different groups, or
strata. Each group should have something in common. For example, you could divide your customers
into age groups: 18–25, 26–35, 36–45, and so on.
2. Sample from Each Stratum: Next, you take a random sample from each group. The size of the sample
from each group can be proportional to the size of the group in the population. For example, if 20% of
your customers are in the 18–25 age group, then 20% of your sample should come from that group.
3. Combine the Samples: Finally, you combine the samples from each group to form your overall sample.

EXAMPLE Suppose you have 1,000 customers and you want to survey 100 of them. You divide them
into four age groups:

18–25: 200 customers

26–35: 300 customers
36–45: 250 customers
46–55: 250 customers

You decide to take 10% of each group:

18–25: 20 customers
26–35: 30 customers
36–45: 25 customers
46–55: 25 customers

By doing this, you make sure that each age group is fairly represented in your survey.
Let’s now look at how you would construct this stratified sample in Excel.

EXAMPLE Using a similar scenario as described in the previous example, you are going to construct a
stratified sample in Excel.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 10
 The Excel file customer_data_age_group.xlsx contains information about 20 customers divided into four
age groups: 18-25, 26-35, 36-45, 46-55.

Use the customer_data_age_group.xlsx Excel file and perform the following to construct a stratified sample.

1. Ensure the data is sorted by Age_Group. For this data, it is sorted by Age_Group. Always check to make
sure the column that contains the group/strata that you want to sample from is sorted.

2. Decide how many samples you want from each group. Let’s sample two customers per age group.

3. Create a new worksheet in your Excel file named stratified_sample. Create the same three columns that
are in the customer_data worksheet. You are creating a blank worksheet that will contain your stratified
sample when you are done.

4. Return to the customer_data worksheet and select the Age_Group column.

5. Under the Home tab, select the dropdown arrow to the right of the Sort and Filter tool. Select Filter.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 11
 6. A dropdown box will appear on the right side of the Age_Group column. Select the drop-down box.
Choose Select All to remove all checks from the checkboxes, then choose the checkbox next to the 18-25
age category. Select the Apply button.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 12
 7. Only the rows for the 18-25 age category will be displayed in the worksheet. Copy the first two rows of
the filtered data. Then, paste these rows into the stratified_sample worksheet. These are the first two
selected rows from the first age group.

You are now going to repeat these steps for each age group.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 13
 8. To revert the data in the customer_data worksheet back to the non-filtered data, select the drop-down
box to the right of the Age_Group column header and select Clear Filter from ‘Age_Group’.

9. Repeat steps 6-8 for the remaining four age groups. Your final systematic sample that will be in your
stratified_sample worksheet should look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 14
 Now, try to construct a stratified sample in Excel on your own!

 TRY IT

The product_groups.xlsx Excel file contains information about transactions across different product
categories. Construct a stratified sample in which you select three transactions from each product category.

Solution: Your final stratified sample should have 18 rows and look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 15
  WATCH

Check out this video about creating a stratified sample.

 DID YOU KNOW

In this tutorial, you used a manual method to construct a stratified sample (you selected the top 2 or 3 rows
from each group). This approach ensures that everyone obtains the same results, which is important for
learning and consistency. However, it is important to understand that in real-world applications, samples
should be selected randomly from each group or stratum to ensure the sample is representative of the
entire population.

 TERMS TO KNOW

Stratified Sampling
A type of sampling method that involves dividing a population into smaller groups, known as strata, that
share similar characteristics.

Strata
Subgroups or segments of a population that share similar characteristics.

1d. Cluster Sampling

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 16
 Cluster sampling is a sampling method used to make data collection easier and more efficient, especially when
dealing with large populations. This technique is particularly useful in business data analytics when you need to
gather information from a large group but want to save time and resources.

Think of yourself as a business analyst working for a company that wants to survey its customers across the
country. Instead of trying to survey every single customer, which would be very time-consuming and expensive,
you can use cluster sampling to make the process simpler.

Follow these steps to create a cluster sample.

 STEP BY STEP

1. Divide the Population into Clusters: First, you divide the entire population into smaller groups, called
clusters. Each cluster should be a mini representation of the whole population. For example, you could
divide your customers by city or region.
2. Randomly Select Clusters: Next, you randomly select a few clusters to survey. Instead of surveying
customers from every city, you might choose a few cities at random.
3. Survey Everyone in the Selected Clusters: Finally, you survey all the customers within the chosen
clusters. This way, you still get a good mix of data without having to survey everyone.

EXAMPLE Let’s say you have 10,000 customers spread across 5 cities. You want to survey 1,000
customers. Here’s how you could use cluster sampling:

Divide into Clusters: Each city is a cluster.

Randomly Select Clusters: Randomly pick 10 cities.
Survey Everyone in Selected Clusters: Survey all customers in those 10 cities.

By doing this, you gather data from different parts of the country without having to reach out to every single
customer.
Let’s now look at how you would construct this cluster sample in Excel.

EXAMPLE Using a similar scenario as described in the previous example, you are going to construct a
cluster sample in Excel.

The customer_cluster_data.xlsx Excel file contains customers that are spread across five different cities.
Each city represents a cluster. In this example, we have 5 clusters: CityA, CityB, CityC, CityD, and CityE.
Each row in the data represents an individual customer who lives in one of the cities.

Use the customer_cluster_data.xlsx Excel file and perform the following to construct a cluster sample.

1. Create a new worksheet in your Excel file named cluster_sample. Create the same three columns that are
in the customer_data worksheet. You are creating a blank worksheet that will contain your cluster sample
when you are done.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 17
 2. Return to the customer_data worksheet and create two new columns named City_List and Number.
Create City_List in column E and create Number in column F. List the five cities in the City_List column and
assign a number 1 to 5 to each city.

3. Use the RANDBETWEEN() function to randomly select numbers corresponding to the cities.

In cell H2, enter the formula:

=RANDBETWEEN(F2, F6)

This will generate a random number between 1 and 5. Suppose it generates 3.

In cell H3, enter the formula:

=RANDBETWEEN(F2, F6)

This will generate another random number between 1 and 5. Suppose it generates 4.

If both cells H2 and H3 have the same number, regenerate one of them.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 18
 You will obtain different random numbers because the function RANDBETWEEN() is designed to generate a
new random number within the specified range each time it is recalculated.

So that your results will be consistent with this example, use 3 and 4 as your randomly selected values from
RANDBETWEEN().

4. Match the numbers 3 and 4 to the cities. These random values correspond to CityC and CityD. Filter the
data so that all customers from CityC and CityD are selected.

5. Select the City column. Under the Home tab, select the dropdown arrow to the right of the Sort and Filter
tool. Select Filter. A dropdown box will appear at the right side of the City column. Select the drop-down
box. Choose Select All to remove all checks from the checkboxes, then choose the checkboxes next to
CityC and CityD. Select the Apply button.

6. Only the rows for CityC and CityD will be displayed in the worksheet. Copy all the rows. Then, paste
these rows into the cluster_sample worksheet. Your final cluster sample that will be in your cluster_sample
worksheet should look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 19
 Now, try to construct a cluster sample in Excel on your own!

 TRY IT

The regions_cluster.xlsx Excel file contains information about customers across different regions of the
country. Construct a cluster sample in which you select customers from three regions of the country.

To obtain consistent results, assign the numbers 1 to 4 to each region. That is, the North region should be
assigned the number 1, South should be assigned the number 2, East should be assigned the number 3,
and West should be assigned the number 4. Assume that using the RANDBETWEEN() function, you obtain
the random numbers 1 and 3.

Your final cluster sample should have 10 rows and look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 20
  WATCH

Check out this video on creating a cluster sample.

Cluster sampling is a useful tool in business data analytics. It helps you collect data efficiently and effectively,
making it easier to analyze and draw conclusions. Remember, the key is to ensure that each cluster is a good
representation of the whole population.

 TERM TO KNOW

Cluster Sampling
A type of sampling method where entire clusters are selected and all elements within the cluster are
sampled.

1e. Difference Between Stratified and Cluster Sampling

Understanding the difference between stratified and cluster sampling is crucial for selecting the appropriate
method for your business data analytics needs, as each technique offers unique advantages and applications
depending on the structure and goals of your project.

Below is a table that highlights the key differences between cluster sampling and stratified sampling, along with
business data analytics examples for each method.

Aspect Stratified Sampling Cluster Sampling

Definition Divides the population into strata, then randomly Divides the population into clusters, then
samples from each stratum. randomly selects entire clusters to
sample.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 21
 Purpose To ensure representation of specific subgroups To simplify data collection and reduce
within the population. costs, especially for large, dispersed
populations.

Population Strata are based on specific characteristics (for Clusters are usually based on natural
Division example, age, income, department). groupings (for example, geographical
areas).

Sampling Randomly sample from each stratum Randomly select a few clusters and
Process proportionally to its size in the population. include all members within those clusters.

Homogeneity Strata should be homogeneous (similar within Clusters should be heterogeneous

each stratum). (diverse within each cluster).

Example Customer Segmentation: A bank segments Market Research: A retail chain analyzes
customers by income levels and samples from sales data from randomly selected stores
each segment to analyze spending behavior. to identify regional trends.

 SUMMARY

In this lesson, you learned that sampling is essential in business data analysis because it allows
analysts to draw conclusions about a larger population without examining every individual. Sampling
methods like random sampling, systematic sampling, stratified sampling, and cluster sampling are
used to select a subset of individuals from a larger population. These methods help ensure that the
sample is representative of the population, which is crucial for making accurate inferences about a
population. For each sampling method, you gained hands-on practice in constructing each type of
sample in Excel. These Excel exercises connected the concept of each sampling method with real-
world applications, guiding you through how these sampling methods are constructed in practice.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Cluster Sampling
A type of sampling method where entire clusters are selected and all elements within the cluster
are sampled.

Strata
Subgroups or segments of a population that share similar characteristics.

Stratified Sampling
A type of sampling method that involves dividing a population into smaller groups, known as strata,
that share similar characteristics.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 22
 Systematic Sample
A type of sampling method where every kth element is selected.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 23
 Point Estimation
by Sophia

 WHAT'S COVERED

In this lesson, you will learn about point estimation for means and proportions, with a review of sample
statistics and population parameters. Specifically, this lesson will cover:
1. Introduction to Point Estimation
1a. Key Concepts for Point Estimations
1b. Point Estimations for Means and Standard Deviations
1c. Point Estimators for Proportions
1d. Proportions vs. Means

1. Introduction to Point Estimation

In the last tutorial, you learned that samples can provide good estimates of the whole group. This means that a
well-chosen sample can reflect the characteristics of the entire population. This is where point estimators come
into play.

A point estimator is a statistic that provides an estimate of a population parameter. For example, if you want to
know the average height of students in a school, you might take a sample of 50 students and calculate the
average height of those 50 students. This average is a point estimator of the true average height of all students
in the school.

Sampling is linked to point estimators in several ways, such as:

Representativeness: A good sample should be representative of the population. If the sample is

representative, the point estimator (like the sample mean) will be close to the true population parameter
(like the population mean).
Unbiasedness: An unbiased point estimator means that, on average, it hits the true population parameter.
Random sampling methods, like simple random sampling, help ensure that our point estimators are
unbiased.
Precision: The larger and more representative the sample, the more precise the point estimator will be.
Precision refers to how much the estimates vary from sample to sample. Larger samples tend to give more
precise estimates.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 24
 Imagine you want to estimate the average spending of customers in a store. You cannot ask every customer, so
you take a random sample of 100 customers and find their average spending. This average spending is your
point estimator. If your sample is representative, this point estimator will be a good estimate of the average
spending of all customers.
By using proper sampling methods, as you learned previously, you can ensure that the point estimators are
accurate and reliable. This allows us to make informed decisions based on the data collected from a sample,
rather than needing to survey the entire population.

Let’s now look at several common point estimators.

 TERM TO KNOW

Point Estimators
Statistics used to provide a single best guess or estimate of an unknown population parameter based
on sample data.

1a. Key Concepts for Point Estimations

Distinguishing between population and sample information is a review of content covered earlier in the course,
but these are key concepts in understanding point estimators. As a reminder:

Population vs. Sample:

Population: The entire group you are interested in.

Sample: A subset of the population used to make inferences about the whole group.

Parameter vs. Statistic:

Parameter: A value that describes a characteristic of a population.
Statistic: A value that describes a characteristic of a sample.

1b. Point Estimations for Means and Standard Deviations

A point estimator for the population mean, μ, is a statistic that estimates the entire population's true mean based
on a sample. The most common point estimator for the population mean is the sample mean, .

estimates μ

To estimate the population standard deviation, σ, use the sample standard deviation, s.

s estimates σ

EXAMPLE You are a data analyst at a financial firm. Your manager has asked you to analyze the daily
closing prices of a particular stock to understand its performance and volatility. You are going to use point
estimators to provide this information.

You take a random sample of 30 daily closing stock prices for stock XYZ from last year. You find

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 25
 and

Mean closing price: The mean closing price of stock XYZ over the past month is estimated to be $151. This
value is a point estimate of the population mean, μ, which represents the average closing price of the stock
in the entire population of trading days. The sample mean, is used as a point estimator because it
provides a single, best guess of the population mean based on the sample data.

The point estimate of the mean closing price provides a concise summary of the stock’s average
performance over the past year. It helps investors understand the typical price level of the stock.

Standard Deviation: The standard deviation of the closing prices is approximately $5.88. This value is a
point estimate of the population standard deviation, σ, which measures the volatility of the stock’s closing
prices in the entire population of trading days. The sample standard deviation, s, is used as a point
estimator because it provides a single, best guess of the population standard deviation based on the
sample data.

The point estimate of the standard deviation gives insight into the stock’s volatility. A higher standard
deviation indicates greater variability in the stock prices, suggesting higher risk. Conversely, a lower
standard deviation indicates more stable prices.

 TRY IT

Using the stock_prices_sample.xlsx Excel file, develop point estimators for the population mean and
standard deviation closing stock price.

You can use the AVERAGE() and STDEV.S() Excel functions. Below are the formulas to create the point
estimators in Excel.

Remember, estimates μ; that is, the sample average, , is the point estimator for the population mean, μ.
By calculating the average of the stock prices in this sample using the AVERAGE() function in Excel, you are
estimating what the average closing price would be if you had data for every possible closing stock price in
the population. This point estimator is useful because it provides a single value that estimates the central
tendency of the stock prices in the population.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 26
 Likewise, s estimates σ; that is, the sample standard deviation, s, is the point estimator for the population
standard deviation, σ. By calculating the standard deviation of the stock prices in your sample using
the STDEV.S() function in Excel, you are estimating what the standard deviation of the closing stock prices
would be if you had data for every possible closing stock price in the population. This point estimator is
useful because it provides a single value that summarizes the variability of the stock prices in the
population.

1c. Point Estimators for Proportions

A point estimator for a population proportion uses the proportion from a sample to estimate the true proportion
of the entire population. Let’s think about this concept in simple terms. A proportion is a part of a whole. For
example, if you have a survey of 100 customers and 60 of them say they are satisfied with your product, the
proportion of satisfied customers is 60 out of 100, or 60%. A point estimator is a single value that gives us an
estimate of a population parameter. In this case, the population parameter is the true proportion of satisfied
customers in the entire customer base.

To estimate the population proportion, denoted as p, you use the sample proportion, denoted as . The sample
proportion is calculated using the formula:

where x is the number of favorable outcomes and n is the total number of observations in the sample.
Suppose you surveyed 200 customers and 150 of them said they were satisfied with your product. The sample
proportion is found with the following calculation:

This means that 75% of the surveyed customers are satisfied. 75% is the point estimate of the true proportion of
satisfied customers in the entire population, p.
Understanding point estimators for proportions helps businesses make informed decisions. For example, if a
high proportion of customers are satisfied, the business might continue with its current strategies. If the
proportion is low, the business might look for ways to improve customer satisfaction. The point estimator for the
proportion helps businesses make educated guesses about the larger population proportion based on sample
data.

 DID YOU KNOW

A proportion and a probability are closely related concepts, but they are not the same. Below are some key
differences and how they relate to point estimation.

Proportion: A proportion is a measure that represents the fraction of a sample or population that has a
particular characteristic. For example, if 150 out of 200 surveyed customers are satisfied, the proportion
of satisfied customers is 75%. When you use the sample proportion, , to estimate the population
proportion, p, you are using a point estimator.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 27
 Probability: Probability is a measure of the likelihood that a particular event will occur. For example, if
you randomly select a customer from the surveyed group, the probability that the customer is satisfied
is 75%, assuming the sample is representative of the population.

Let’s take a look at an example that distinguishes between the concepts of proportion and probability.

EXAMPLE You are a data analyst at a healthcare clinic that is concerned about the high rate of patient
no-shows for scheduled appointments. To address this issue, you collected data from a sample of 500
scheduled appointments over the past month.

Proportion:
After analyzing the data, you found that 75 out of the 500 scheduled appointments resulted in no-shows.
The proportion of no-show appointments is:

This means that based on your sample, you estimate that 15% of all scheduled appointments at the clinic
result in no-shows. This proportion is a point estimate of the population proportion, p.

Probability:
If you randomly select one appointment from the 500 appointments you sampled, the probability that this
appointment will be a no-show is 15%. This probability assumes that your sample is representative of the
entire population of scheduled appointments at the clinic.

Why This Matters:

Understanding the proportion of no-show appointments helps the clinic assess the overall impact of no-
shows on its operations. A high proportion of no-shows can lead to inefficiencies, lost revenue, and longer
wait times for other patients. By identifying the extent of no-shows, the clinic can implement strategies to
reduce them, such as sending appointment reminders, offering flexible scheduling, or implementing a no-
show policy.

On the other hand, knowing the probability helps in making predictions about individual appointments. For
instance, if the clinic wants to predict the likelihood that a randomly chosen appointment will be a no-show,
they can use the probability derived from the sample data.
Let’s look at an example of how a point estimate of a proportion can be used in a real-world business data
analytics scenario.

EXAMPLE You are a data analyst at a large corporation. The company is concerned about the amount
of food waste generated in its corporate cafeteria. To address this, you collected data from a sample of 100
meals served in the last month in the cafeteria and noted whether each meal resulted in significant food
waste.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 28
 The point estimate for the proportion of meals that result in significant food waste is 25%. This means that
based on your sample, you estimate that 25% of all meals served in the corporate cafeteria result in
significant food waste.

Understanding the proportion of food waste is crucial for the corporation. A high proportion of food waste
indicates inefficiencies and potential cost savings opportunities. By identifying the extent of food waste, the
company can implement strategies to reduce waste, such as better portion control, improved menu
planning, or employee education programs.

 TRY IT

Using the data in the food_waste.xlsx file, determine the point estimator for the proportion of food waste
for the entire company. A snapshot of the data is provided below. There are two columns: a unique
identifier for the meal, Meal_ID, and an indicator variable, Food_Waste, that specifies whether the meal had
significant waste or not.

Perform the following.

1. In cell D3, enter Total Meals. In cell E3, enter the following formula:

=COUNT(A2:A101)

The COUNT() function will count the total number of meals.

2. In cell D4, enter Meals with Food Waste. In cell E4, enter the following formula:

=COUNTIF(B2:B101,”Yes”)

The COUNTIF() function will count the total number of meals with significant food waste.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 29
 3. In cell D5, enter Proportion of Food Waste. In cell E5, enter the following formula:

=E4/E3

This is the point estimator for the proportion of food waste for the company, The Excel worksheet
with the formulas used is shown below.

 TRY IT

Using the hospital_readmissions.xlsx Excel file, determine the point estimator for the proportion of all
patients that were readmitted. This file contains a random sample of patients that have been readmitted to
a hospital.

Solution: The Excel worksheet with the formulas used is shown below.

1d. Proportions vs. Means

Ensuring you understand the difference between means and proportions is important in data analytics because
it will ensure that you select the appropriate analytical technique going forward. Below is a table that
summarizes the differences between proportions and means.

Aspect Proportions Means

Type of Categorical (yes/no, success/failure, regions of Quantitative data (test scores, sales)
Data the country)

Purpose Understand the relative frequency of a Measures the central tendency (average) of
category quantitative data

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 30
 Examples Defect rates: proportion of defective items in a Sales figures: average sales per day
batch
In business data analytics, point estimation helps in making informed decisions by providing single best
estimates of population parameters based on sample data. For example, estimating the proportion of satisfied
customers can guide improvements in customer service, while estimating average sales can help in forecasting
and inventory management.

While point estimators are useful for providing a single value estimate, they do not convey the uncertainty or
variability around the estimate. You are going to learn about other methods that can be used to understand the
range within which the true population parameters are likely to fall.

 SUMMARY

In this lesson, you were introduced to the concept of point estimators. A point estimator is a statistic
used to provide a single best guess or estimate of an unknown population parameter based on sample
data. Point estimators allow companies to be efficient in their data decision making because they do
not need to survey an entire population. Instead, they can rely on a well-chosen sample to provide
accurate estimates of key metrics. This approach saves time and resources while still delivering
valuable insights. This efficiency is crucial in fast-paced environments where timely decisions can lead
to competitive advantages. Point estimators differ based on whether you are dealing with means or
proportions. Point estimators for means typically use sample averages to estimate the population
mean, while point estimators for proportions use sample proportions to estimate the population
proportion.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Point Estimators
Statistics used to provide a single best guess or estimate of an unknown population parameter
based on sample data.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 31
 Sampling Distributions
by Sophia

 WHAT'S COVERED

In this lesson, you will learn about sampling distributions for means and proportions. Specifically, this
lesson will cover:
1. Introduction to Sampling Distributions
1a. Importance of Sampling Distributions
1b. Sampling Distribution of the Mean (x̄)
1c. Estimating Population Mean Using Sampling Distributions: A Practical Approach
1d. Sampling Distribution of the Proportion (p̂)

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational idea of sampling distributions.

1. Introduction to Sampling Distributions

A sampling distribution is a way to understand how a statistic (like an average) from a sample can vary. Imagine
you want to know the average amount of money customers spend at a store. Instead of asking every customer,
you take several small samples and calculate the average for each sample. The collection of these averages
forms a sampling distribution.

Sampling distributions help us understand the variability of our sample statistics. This is important because it
allows businesses to make predictions and decisions based on sample data, rather than needing to survey an
entire population.

Let’s say a company wants to know the average amount of time customers spend on their website. Instead of
tracking every single visitor, they take multiple samples of 100 visitors each and calculate the average time
spent for each sample. These averages might be slightly different, but together they form a sampling
distribution.

Imagine you take 10 different samples of 100 visitors each and calculate the average time spent on the website
for each sample. You might get averages like 5 minutes, 5.2 minutes, 4.8 minutes, etc. If you plot these
averages on a graph, you’ll see a distribution of these sample averages. This is your sampling distribution. The

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 32
 histogram below shows a sampling distribution of the average time spent on a website using 10 different
samples of 100 visitors each.

Understanding sampling distributions helps businesses make better decisions. For example, if the sampling
distribution shows that most sample averages are around 5 minutes, the company can be confident that the
true average time spent on the website is close to 5 minutes.

 KEY CONCEPT

Population: The entire group you’re interested in, like all visitors to the website.
Sample: A smaller group selected from the population, like 100 visitors.
Sample Statistic: This is a number that describes a sample, like the average time spent on a website.
Sampling Distribution: The distribution of a statistic (like the average) from many samples.

 TERM TO KNOW

Sampling Distribution
Distribution of a given statistic based on a random sample, showing how the statistic varies from sample
to sample.

1a. Importance of Sampling Distributions

Sampling distributions are crucial in business data analytics for several reasons:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 33
 1. Making Predictions: Sampling distributions help businesses make predictions about a population based on
sample data. For example, if a company wants to know the average amount customers spend, they can use
a sampling distribution to estimate this average without surveying every customer.
2. Understanding Variability: They show how much a sample statistic (like an average) can vary from sample
to sample. This helps businesses understand the reliability of their data. For instance, if the average time
spent on a website varies little between samples, the company can be more confident in its estimate.
3. Decision Making: Businesses often make decisions based on sample data. Sampling distributions provide a
way to measure the uncertainty of these decisions. For example, a company might use a sampling
distribution to decide whether a new marketing strategy is effective, based on sample sales data.
4. Quality Control: In manufacturing, sampling distributions are used to monitor product quality. By taking
samples from production lines and analyzing the distribution of defects, companies can identify and
address quality issues more efficiently.
5. Cost Efficiency: Surveying an entire population can be expensive and time-consuming. Sampling
distributions allow businesses to gather insights and make decisions based on smaller, more manageable
samples, saving time and resources.

EXAMPLE A hospital wants to know the average wait time for patients in the emergency room (ER).
Instead of tracking every patient, they take several samples of 100 patients each and calculate the average
wait time for each sample. The distribution of these averages forms a sampling distribution.

The key aspects are:

Population: All patients visiting the ER.

Sample: Groups of 100 patients each.
Sample Statistic: Average wait time in the ER.
Sampling Distribution: The distribution of the average wait times from multiple samples.

The hospital took 50 samples of 100 patients each, calculated the average wait time for each patient, and
plotted these 50 averages, as shown in the histogram below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 34
 By analyzing the sampling distribution, the hospital can estimate the true average wait time for all patients.
This helps in making decisions about staffing, resource allocation, and process improvements.

The sample averages are between 28 and 32 minutes, which suggests that the average wait time for
patients in the ER is likely around 30 minutes, with some variation. This histogram helps the hospital
understand the typical wait time and how much it can vary.

1b. Sampling Distribution of the Mean (x̄)

Suppose you want to understand the average spending of customers at a store. Instead of asking every single
customer, you take several smaller samples and calculate the average spending for each sample. The collection
of these sample averages forms what is known as the sampling distribution of the mean ( ).

The table below shows the collection of these sample means, which forms the sampling distribution of

Sample Sample Size Mean (x̄)

1 100

2 100

3 100

... ... ...

30 100

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 35
 The histogram below is a plot of all the sample means from the table above. The histogram visualizes the
distribution of the sample means. The histogram helps you understand the central tendency (average) and
variability of the sample means.

The graphic below illustrates the construction of another sampling distribution of Even though the population
mean is 82.5, the image visually demonstrates how the means of samples drawn from that population can vary.
This sampling distribution of consists of six samples, each with a sample size of 20.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 36
 Source: Sampling Variability - MathBitsNotebook(A2)

Just as with other distributions we have studied, the sampling distribution of has its own mean, standard
deviation, and shape.

Returning to the example of understanding the average spending of customers at a store, no information is
available regarding the original population's spending patterns. However, this is acceptable due to the Central
Limit Theorem, a statistical concept that offers insights into the sampling distribution of the sample mean ( ).

The Central Limit Theorem (CLT) provides you with information about the characteristics of the sampling
distribution of The CLT tells us:

Regardless of the shape of the population distribution (whether skewed, uniform, etc.), the sampling
distribution of will be approximately normal (bell-shaped) if the sample size is large enough (at least 30 for
most distributions).
The sampling distribution of has its own mean and standard deviation.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 37
 The sampling distribution of will have a mean equal to the population mean.

The standard deviation of the sampling distribution of is the population standard deviation divided by the
square root of the sample size (this is called the standard error).

In practice, the true value of σ is typically unknown and cannot be directly calculated. Therefore, σ is estimated
using the sample standard deviation, denoted as s. Consequently, the sample standard error is defined as
follows:

Let’s see how you might use the Central Limit Theorem to tell you something about a sampling distribution of
the

EXAMPLE A retail chain wants to understand its average monthly sales across all its 250 stores to
make informed business decisions. However, analyzing the average sales across all its stores is time-
consuming and resource-intensive. Instead, the company asks you to use a sample-based approach to
estimate the average monthly sales for all its 250 stores.

You create 50 samples of sales from 10 stores and find the average of each sample. That is, you create the
sampling distribution of shown in the spreadsheet below. The first five rows of the sampling distribution
for sales_CLT.xlsx are shown below.

A description of each column is:

Sample_Number: Identifies the sample (1 to 50).

Store_1 to Store_10: The sales figures for each of the 10 stores in that particular sample.
Sample_Means: The average sales calculated from the 10 stores in that sample. The sample means in
column L is the sampling distribution of

For this sampling distribution, the number of samples is 50 and the sample size (n) is 10.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 38
 Based on the known properties of the sampling distribution of , you can provide the company with an
estimate of the average monthly sales across all 250 stores by using:

For this problem, can be found by taking the mean of the sample means in column L. In Excel, in cell N2,
enter:

=AVERAGE(L2:L51)

You find that the mean of the sampling distribution of is $4,948. This gives the retail chain a reliable
estimate of the average monthly sales across all stores. The mean of the sample means is a more reliable
estimate of the population mean because it aggregates information from multiple samples, reducing the
impact of any anomalies in individual samples. This is why the retail chain uses the mean of the sample to
estimate the overall average monthly sales for all its stores.

You can also find the standard error by entering the following formula in cell N3:

=STDEV.S(AF2:AF51)/SQRT(50)

The standard error is $26. Since the population standard deviation is not known, the Excel formula that is
being used to calculate standard error is:

This helps the retail chain understand the precision of their estimate of the population mean, .
That is, this value helps the retail chain understand how much the mean of the sample means ($4,948) is
expected to vary from the true population mean. A smaller standard error indicates a more precise estimate
of the population mean, μ.

 WATCH

Check out this video on using a sample-based approach to estimate the average monthly sales.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 39
  TERMS TO KNOW

Sampling Distribution of x̄
A distribution of all possible sample means of a given size from a population, showing how the sample
mean varies from sample to sample.

Central Limit Theorem (CLT)

A property that states the sampling distribution of approaches a normal distribution as the sample size
becomes large, regardless of the original population’s distribution.

Standard Error
Measure of the variability of a sample statistic (such as ) from the population, calculated as the
standard deviation of the sampling distribution of that statistic.

1c. Estimating Population Mean Using Sampling Distributions: A Practical Approach

The mean of the sampling distribution of is theoretically equal to the population mean, μ. However, in practice,
there are several reasons why they might not be exactly the same. For instance:

Sampling Variability. Each sample you take is different, and the sample mean will vary from sample to
sample. This variability is natural and expected. The more samples you take, the closer the mean of the
sampling distribution will get to the population mean, but with a limited number of samples, there will
always be some difference.
Sample Size. The size of each sample affects the accuracy. Smaller samples tend to have more variability,
which can lead to a greater difference between the sample mean and the population mean. Larger samples
tend to produce sample means that are closer to the population mean.
Random Error. Random errors can occur due to various factors, such as measurement errors or random
fluctuations in the data. These errors can cause the sample mean to differ slightly from the population
mean.
Finite Number of Samples. In the previous example, you took 50 samples from 10 stores. While 50 is a
good sample size, it is finite. If you were to take an infinite number of samples, the mean of the sampling
distribution of would converge to the population mean, μ. With a finite number of samples, there will
always be some small difference.

The histogram below shows that the mean of the sampling distribution of is not exactly equal to the
population mean, due to some of the factors discussed above.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 40
 1d. Sampling Distribution of the Proportion (p̂)
Imagine you’re working for a company that wants to know the proportion of employees who prefer working
remotely versus in the office. Instead of asking every employee, you decide to survey a smaller group, or
sample, of employees. The proportion of employees in your sample who prefer working remotely is called .
Now, if you took many different samples from your company, each sample would give you a different . The
distribution of all these values from different samples is called the sampling distribution of the proportion (p̂).

The table below shows the collection of these sample proportions, which forms the sampling distribution of

Sample Sample Size Sample Proportion (p̂)

1 100

2 100

3 100

... ... ...

30 100

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 41
 The histogram below is a plot of all the sample proportions from the table above. The histogram visualizes the
distribution of the sample proportions. The histogram helps you understand the central tendency (average) and
variability of the sample proportions.

The histogram above was created with 30 samples of size 100. If you were to take 100,000 samples of size 100,
the histogram would look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 42
 The reason the histogram becomes smoother and looks more like a normal distribution with more samples is
due to the law of large numbers and the Central Limit Theorem working together. As you take more samples,
the average of the sample proportions will converge to the population proportion. This means that with more
samples, the overall distribution of sample means will better represent the true population mean.

The Central Limit Theorem states that the distribution of will approach a normal distribution as the sample
size increases. However, this also implies that with a larger number of samples, the variability in the sample
means decreases, making the distribution appear more normal.

Below are some key points about the sampling distribution of

Random Sampling. Each sample should be randomly selected to ensure that every employee has an equal
chance of being chosen. This helps make your sample representative of the whole company.

Sample Size. The number of employees in each sample affects the sampling distribution. Larger samples
tend to give more accurate estimates of the true proportion of employees who prefer working remotely.

Shape of the Distribution. The sampling distribution of will approximate a normal distribution as the
sample size becomes large.

For the sampling distribution of , a common rule of thumb is that the sample size, n, should be large
enough such that both np and are greater than 5, where p is the population proportion. This helps

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 43
 ensure that the normal approximation is valid.

Mean of the Distribution. The mean of the sampling distribution of is equal to the true population
proportion, p. This means that, on average, will be a good estimate of p. The formula for the mean of the
sampling distribution of is:

In the previous example, if you knew the true population proportion, p, was 60%, the mean of the sampling
distribution of would be close to p.

Standard Deviation (Standard Error): The standard deviation of the sampling distribution of is called the
standard error. It measures how much varies from sample to sample. The formula for the standard error of
is:

The key reason for studying the sampling distributions of (the sample mean) and (the sample proportion) is
that these distributions help assess how close the sample mean or proportion is to the true population mean or
proportion. In other words, they indicate how accurate is as an estimate for μ (the population mean) and how
accurate is as an estimate for p (the population proportion). These topics will be explored in more detail in
future tutorials.

 TERMS TO KNOW

Sampling Distribution of p̂
A distribution of all possible sample proportions of a given sample size from a population, showing how
the sample proportion varies from sample to sample.

Law of Large Numbers

A property that states as the size of a sample increases, the sample mean will become closer to the
population mean and the sample proportion will become closer to the population proportion.

 SUMMARY

In this lesson, you learned about sampling distributions. A sampling distribution is the distribution of a
given statistic (such as the sample mean or sample proportion) based on a random sample. It shows
how the statistic varies from sample to sample. A practical reason to study sampling distributions in
business data analytics is to make accurate and reliable inferences about a population based on
sample data. This tutorial focused on two specific sampling distributions: sampling distribution of the
sample mean and sample proportion. The sampling distribution of the sample mean represents the

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 44
 sample means of all possible samples of a given size from a population. As the sample size increases,
the sampling distribution of the mean approaches a normal distribution, regardless of the population’s
distribution. The sampling distribution of the sample proportion represents the sample proportions of
all possible samples of a given size from a population.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Central Limit Theorem (CLT)

A property that states the sampling distribution of approaches a normal distribution as the sample
size becomes large, regardless of the original population’s distribution.

Law of Large Numbers

A property that states as the size of a sample increases, the sample mean will become closer to the
population mean and the sample proportion will become closer to the population proportion.

Sampling Distribution
Distribution of a given statistic based on a random sample, showing how the statistic varies from
sample to sample.

Sampling Distribution of p̂
A distribution of all possible sample proportions of a given sample size from a population, showing
how the sample proportion varies from sample to sample.

Sampling Distribution of x̄
A distribution of all possible sample means of a given size from a population, showing how the
sample mean varies from sample to sample.

Standard Error
Measure of the variability of a sample statistic (such as ) from the population, calculated as the
standard deviation of the sampling distribution of that statistic.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 45
 Sampling Error
by Sophia

 WHAT'S COVERED

In this lesson, you will learn about sampling error and how it affects the outcomes of your data analysis.
Specifically, this lesson will cover:
1. Introduction to Sampling Error
1a. Sampling Error for Means
1b. Sampling Error for Proportions

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational idea of sampling error.

1. Introduction to Sampling Error

When you are analyzing data in business, it is important to understand that the data you work with often comes
from samples, not the entire population. This means there can be differences between the sample data and the
actual population data, which is known as sampling error. These errors can affect the accuracy of your analysis
and the decisions you make based on that analysis.

Suppose you are trying to figure out the average amount of money people spend at a store. If you randomly
select 10 customers and calculate their average amount spent, the sample mean might differ significantly from
the true average amount of money spent by all customers at the store. Asking a small group of people about
their spending habits might not perfectly represent everyone who shops there. This difference is a sampling
error. If you increase the sample size to 100 customers, the sample mean is likely to be closer to the true
population mean. The larger and more diverse your sample, the more likely it is to accurately reflect the whole
population, reducing the impact of sampling errors.

 TERM TO KNOW

Sampling Error
The difference between the results obtained from a sample and the actual values in the population from
which the sample was drawn.

1a. Sampling Error for Means

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 46
 Suppose you are analyzing the average amount of money customers spend at a store. Let’s simulate a
population of customer spending amounts and draw samples of different sizes to visualize how the sample
mean varies and how the standard error changes.

Presume you know that the average amount spent by all customers is $500 and the standard deviation among
all the customer spending amounts is $100. That is, and

You construct two sampling distributions of The first is 1,000 samples with a sample size of 30, and the
second is 1,000 samples with a sample size of 300. You construct two histograms for each sampling distribution
and overlay a normal distribution over the histogram, as shown below.

The histogram for the smaller sample size (30) shows more spread and variability in the sample means
compared to the histogram for the larger sample size (300). This indicates that the sample means from smaller
samples are more dispersed around the population mean.

The histogram for the larger sample size (300) is smoother and more bell-shaped, closely approximating a
normal distribution. This is due to the Central Limit Theorem, which states that the sampling distribution of will
be approximately normal for large sample sizes.

The standard error, which measures the variability of the sampling distribution of , is smaller for the larger
sample size (300). This is evident from the narrower spread of the histogram. A smaller standard error means
that the sample mean is a more precise estimate of the population mean. Standard error is a metric used to
measure sampling error.

You can calculate the standard error for each distribution to numerically confirm that the amount of variability in
the sampling distribution of with a sample size of 300 is less than the amount of variability in the sampling

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 47
 distribution of with a sample size of 30.

Sample Size Number of Samples Standard Error

30 1000

300 1000

In summary, the standard error provides a measure of how much the sample mean (or other statistic) is
expected to vary from the true population mean due to sampling error. A smaller standard error indicates that
the sample mean is likely to be closer to the population mean, while a larger standard error suggests more
variability and thus more sampling error. You can think of standard error as a measure of accuracy of which a
sample represents a population.

 KEY CONCEPT

Larger sample sizes result in smaller standard error, which means that is a more accurate and precise
point estimate of μ.

Smaller sample sizes result in a larger standard error, which means that is a less accurate and less precise
point estimate of μ.

1b. Sampling Error for Proportions

The concepts of sampling error, standard error, and the Central Limit Theorem apply equally to both sample
means and sample proportions. Whether estimating a population mean or a population proportion, the
variability of sample estimates, the shape of their sampling distributions, and the precision of these estimates
(as measured by the standard error) follow the same rules. This consistency ensures that these foundational
statistical concepts can be reliably applied across different types of data, supporting accurate analytical
outcomes. Let’s now explore the sampling error for proportions motivated by an example.

Suppose you are a business analyst at an online retail company. You want to estimate the proportion of
customers who make a purchase after visiting the website. Instead of surveying every visitor, you take samples
of visitors to estimate this proportion.

Presume you know that 40% of all customers make a purchase after visiting the website, that is, you know the
true population proportion p.

You construct two sampling distributions of The first is 1,000 samples with a sample size of 50, and the
second is 1,000 samples with a sample size of 500. You construct two histograms for each sampling distribution
and overlay a normal distribution over the histogram, as shown below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 48
 The histogram for the smaller sample size (50) shows more spread and variability in the sample proportions
compared to the histogram for the larger sample size (500). This indicates that the sample proportions from
smaller samples are more dispersed around the population proportion.

The histogram for the larger sample size (500) is smoother and more bell-shaped, closely approximating a
normal distribution. This is due to the Central Limit Theorem, which states that the sampling distribution of will
be approximately normal for large sample sizes.

The standard error, which measures the variability of the sampling distribution of is smaller for the larger
sample size (500). This is evident from the narrower spread of the histogram. A smaller standard error means
that the sample proportion is a more precise estimate of the population proportion.

You can calculate the standard error for each distribution to numerically confirm that the variability in the
sampling distribution of with a sample size of 500 is smaller than the variability in the sampling distribution of
with a sample size of 50. This provides validation that larger sample sizes result in less variability in the
sampling distribution.

Sample Size Number of Samples Standard Error

50 1000

500 1000

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 49
  KEY CONCEPT

Larger sample sizes result in smaller standard error, which means that is a more accurate and precise
point estimate of p.

Smaller sample sizes result in a larger standard error, which means that is a less accurate and less precise
point estimate of p.

 SUMMARY

In this lesson, you learned that sampling error represents the difference between a sample statistic and
the actual population parameter, arising from the natural variability in selecting a sample. Sampling
error is not a mistake but a natural part of the sampling process. For the sampling error for the mean,
the standard error is calculated as the population standard deviation divided by the square root of the
sample size, indicating how much the sample mean is expected to vary from the true population mean.
For the sampling error for proportions, the standard error is calculated using the population proportion
and sample size. Increasing the sample size reduces both types of sampling errors, leading to more
accurate estimates of the population parameters like the mean of the population and the proportion of
the population.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Sampling Error
The difference between the results obtained from a sample and the actual values in the population
from which the sample was drawn.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 50
 Confidence Intervals - Means
by Sophia

 WHAT'S COVERED

In this lesson, you will learn about how to construct and interpret a confidence interval for a population
mean that can be applied in a business analytics environment. Specifically, this lesson will cover:
1. Introduction to Confidence Intervals for Means
1a. t-distribution
1b. Confidence Intervals for Means
1c. Confidence Level and Percentage
1d. Calculating and Interpreting Confidence Intervals for Means

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational idea of confidence intervals for means.

1. Introduction to Confidence Intervals for Means

Suppose you want to know the average amount of money customers spend at a retail store, but you cannot ask
every single customer. Instead, you take a sample of customers and calculate the average spending from that
sample. However, how confident can you be that this sample average is close to the true average spending of
all customers?

This is where confidence intervals become useful. A confidence interval provides a range of values that is likely
to contain the true average spending of all customers. For example, if you calculate a 95% confidence interval,
it indicates that you are 95% confident the true average spending lies within that range.

Confidence intervals help you assess the precision of your point estimate and provide a way to express the
uncertainty in your data. They are a valuable tool in business data analytics because they allow you to make
informed decisions based on sample data while accounting for the natural variability inherent in sampling.

 TERM TO KNOW

Confidence Intervals
A range of values used to estimate an unknown parameter of a population.

1a. t-distribution

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 51
 The t-distribution is a type of probability distribution that is symmetric and bell-shaped, similar to the normal
distribution and standard normal but with heavier tails. This means it is more prone to producing values that fall
far from its mean. The t-distribution is particularly useful when dealing with small sample sizes (typically less
than 30) or when the population standard deviation, σ, is unknown.

A t-distribution has a quantity known as the degrees of freedom that affects the shape of the distribution. The
degrees of freedom are calculated as the sample size minus 1 ( ). Below is a graph of several t-distributions
with varying degrees of freedom. As the degrees of freedom increase, the t-distribution curves become closer
to the standard normal distribution.

 TERMS TO KNOW

t-distribution
A type of probability distribution that is symmetric, bell-shaped, and used when estimating population
parameters, particularly when the sample size is small and the population standard deviation is
unknown.

Degrees of Freedom
Value that determines the shape of a t-distribution and is calculated as the sample size minus one.

1b. Confidence Intervals for Means

When constructing a confidence interval for a mean, you create a range around the sample mean ( ) that is
likely to contain the true population mean (μ). In essence, a confidence interval helps you gauge how close your
sample estimate is to the true average by providing a range of plausible values. The way you construct the
range of values for μ is by using:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 52
 The margin of error consists of two components.

1. Critical value
2. Standard error

The formula for constructing a confidence interval is given by:

The t value is known as a critical value and is a value from a t-distribution. The critical value will determine how
wide the confidence interval will be. Think of the critical value as a multiplier that adjusts a range based on how
confident you want to be that the confidence interval contains the true value of μ. The critical value helps you
make sure our confidence interval is wide enough to account for the uncertainty in the sample. Uncertainty
refers to how much you do not know about the true average of the entire population, μ, based on the sample
data (sampling error of the mean).
You use the t-distribution to calculate a confidence interval for the mean when:

You have a small sample size (usually less than 30).

You do not know σ, the population standard deviation (which is often the case).

In most real-world business analytics scenarios, you are not going to know the population standard deviation, σ,
so you will use the t-distribution. However, when the sample size is large (sample size greater than 30), the
degrees of freedom in a t-distribution are also large. Degrees of freedom refer to the number of values in a
calculation that are free to vary. “Free to vary” means that certain values in a calculation can change while still
maintaining the required relationships among the data values.
For example, suppose you have a sample of 5 numbers, and you know their average. If you know the average,
4 of those numbers can be any value, but the 5th number is fixed to ensure the average remains the same. This
means only 4 numbers are “free to vary.”

In the context of a t-distribution, degrees of freedom are calculated as the sample size minus 1 (n-1). This
subtraction accounts for the fact that one value is not free to vary, because it is determined by the other values
to maintain the sample’s average.

So, if you have a sample size of 5, the degrees of freedom would be . This adjustment is important
because it affects the shape of the t-distribution, making it more accurate for small sample sizes.

As the degrees of freedom increase, the t-distribution becomes very similar to the standard normal distribution.
This means that for large samples, the critical values from the t-distribution and the standard normal distribution
converge, making the difference negligible. Therefore, even though you start with the t-distribution, you can use
the standard normal distribution (z-scores) for critical values when the sample size is large.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 53
 The z-scores that you learned about in the previous Challenge are used as the critical values for constructing a
confidence interval for the mean. Below is a table of critical values for the standard normal distribution when the
sample size is large. In this case, the critical values are equivalent to those of the standard normal distribution.

Confidence Level z Critical Value Excel Formula

90% 1.645 =[Link](0.95)

95% 1.96 =[Link](0.975)

99% 2.576 =[Link](0.995)

The Excel function [Link]() returns the inverse of the standard normal cumulative distribution. This
function helps you find the z critical value for the specified confidence level.

Using the standard normal critical values simplifies calculations without sacrificing accuracy. For large samples,
the slight difference between the critical values from the t-distribution and the standard normal distribution does
not significantly impact the confidence interval’s precision.

The second part of the margin of error is the standard error that was introduced in the last tutorial. Recall that
the standard error tells you how much the sample mean is likely to vary from the true population mean. A
smaller standard error means a more precise estimate.

The standard error directly affects the width of the confidence interval. A larger standard error results in a wider
confidence interval, indicating more uncertainty about the population mean. Conversely, a smaller standard
error results in a narrower confidence interval, indicating more precision.

 TERMS TO KNOW

Margin of Error
A numerical value that expresses the range within which the true value of the population parameter is
expected to fall.

Critical Value
A factor used to calculate the upper and lower bounds of a confidence interval.

1c. Confidence Level and Percentage

The confidence level is the percentage that tells us how sure we are that the true population parameter (like the
mean) lies within our confidence interval. Common confidence levels are 90%, 95%, and 99%.

When you select a confidence level, such as 95%, it means that if you were to take many samples and construct
a confidence interval for each, approximately 95% of those intervals would include the true population mean (μ).

The confidence level determines the critical value, t, used in the confidence interval formula. For example, a
95% confidence level corresponds to a t value of 1.96 for large samples.

If you calculate a 95% confidence interval for the average height of students and get a range from 5’4" to 5’8", it
means you are 95% confident that the true average height of all students falls within this range. If you repeated
this process many times, 95% of the intervals you calculated would contain the true average height.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 54
 The confidence level helps you understand the reliability of your point estimate, For example:

Higher Confidence Level (e.g., 99%): This results in a wider interval. While it increases the likelihood that
the interval contains the true population mean, μ, it also reflects more uncertainty about the exact value of
μ.
Lower Confidence Level (e.g., 90%): This results in a narrower interval. Although the interval is more
precise, there is more certainty that it contains the true population mean, μ.

Let’s look at an example to see how changing the confidence level impacts the width of the confidence interval.

EXAMPLE Suppose we have a sample of 100 users from a streaming service, and we want to estimate
the average number of hours they watch TV per week. You find from the sample data that hours and
hours.

The table below shows three different confidence intervals using different confidence levels.

Confidence T Critical Standard Margin of Error Lower Limit Upper Limit

Level Value Error (MOE) x-bar - MOE x-bar + MOE

90% 1.645

95% 1.96

99% 2.576

As you can see, increasing the confidence level from 90% to 99% results in a wider interval, reflecting more
uncertainty that the interval contains the true population mean.

1d. Calculating and Interpreting Confidence Intervals for Means

Let’s look at an example of how you would calculate and interpret a confidence interval in a real-world scenario.

EXAMPLE You are a data analyst at StreamFlix, a popular streaming service. You want to estimate the
average number of hours subscribers watch StreamFlix TV content per week and construct a 95%
confidence interval for this estimate. You select a sample of 50 users from the StreamFlix database

Using the data in the TV_hours_watched.xlsx file, construct a 95% confidence interval for the mean
number of TV hours watched by performing the following:

1. From the Home menu, select the XL Miner Analysis Toolpak and select the Descriptive Statistics option.

2. Put your cursor in the Input Selection box and select the data in column B, Hours_Watched.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 55
 3. Put your cursor in the Output Range box and select column D2. This will be where the output will be
placed in the Excel worksheet.

4. Deselect the options for the Kth Largest and Kth Smallest checkboxes. Select OK.

The option Confidence Level for Mean is where you can specify the type of confidence level you wish. If
you wanted a 90% confidence level, you would enter ‘90’ for this option.

You obtain the following output in your Excel worksheet.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 56
 Everything you need to construct the confidence interval for the mean is in the above output. Remember
that the formula for the confidence interval is:

Both values are in the Excel output. The point estimate is the sample mean, , and is labeled Mean, and the
number that is labeled Confidence Level (95%) is the margin of error. So, all you need to do is the addition
and subtraction of these values to find the lower and upper limits of the 95% confidence interval.

5. Enter Lower confidence interval limit in cell D19. In cell E19, enter the following formula:

=E4-E17

6. Enter Upper confidence interval limit in cell D20. In cell E20, enter the following formula:

=E4+E17

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 57
 The 95% confidence interval for the average number of TV hours watched by all StreamFlix subscribers is
between 12.80 and 17.52 hours.

You can be 95% confident that the average number of TV hours watched by all StreamFlix subscribers per
week lies between 12.80 and 17.52 hours.

The confidence interval plays a crucial role in the business implications for StreamFlix by providing a range
within which the true average number of TV hours watched is likely to fall. Below is how the confidence
interval relates to the various business implications:

Improved User Retention

A narrower confidence interval gives StreamFlix more confidence in their engagement metrics,
allowing them to make more precise adjustments to retain users.
If the confidence interval shows a wide range, it might indicate the need for targeted interventions to
understand and address the variability in user engagement.
Optimized Content Investment
Obtaining a range of the true average viewing hours for all subscribers allows StreamFlix to allocate
resources more effectively. For example, a wider confidence interval might suggest higher
uncertainty, prompting a more cautious approach to content investment.
Now, try constructing a confidence interval for another applied problem.

 TRY IT

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 58
 A tech company wants to analyze the average time spent (in minutes) by users on their new mobile app in a
week. They have collected data from 50 users related to their weekly usage time. Calculate a 99%
confidence interval of the population mean of weekly usage time on a new mobile app using the
weekly_usage_time.xlsx Excel file. Interpret the confidence interval.

Solution: Using the Descriptive Statistics option in the XL Miner Analysis Toolpak, select a 99% confidence
interval to be constructed using the Weekly_Usage_Time column.

Once you calculate the upper and lower limits of the 99% confidence interval, the Excel output should look
like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 59
 The 99% confidence interval for the true mean weekly usage time users spend on the app is between 14
(rounded) and 18 (rounded) minutes.

Interpretation: You can be 99% confident that the mean weekly usage time for all users of the app is
between 14 and 18 minutes.

 WATCH

Check out this video analyzing the average time spent (in minutes) by users on their new mobile app.

 TRY IT

You are a data analyst for TrendyTech, a startup that develops a popular fitness app. You need to estimate
the average number of steps all users take per day. Construct a 90% confidence interval for this estimate
using data in the number_of_steps.xlsx Excel file. The Excel file has the number of steps recorded for 50
users for a particular day.

Solution: The 90% confidence interval is 4,709 to 6,124 steps. You can be 90% confident that the true
average number of steps taken per day by all TrendyTech app users lies somewhere between 4,709 and
6,124 steps.

The Excel output should look like the following:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 60
  SUMMARY

In this lesson, you learned that confidence intervals provide a range of values likely to include the true
population mean, helping to quantify the precision of your estimate while accounting for natural
variability in sampling. A higher confidence level results in a wider interval, increasing the likelihood of
containing the true mean. This tutorial focused on constructing and interpreting a confidence interval
for a population mean, using an example from a streaming service to estimate the average number of
TV viewing hours per week. The confidence interval offered a reliable estimate for all subscribers and
provided insights to guide business decisions on user retention and content investment.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 61
 Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Confidence Intervals
A range of values used to estimate an unknown parameter of a population.

Critical Value
A factor used to calculate the upper and lower bounds of a confidence interval.

Degrees of Freedom
Value that determines the shape of a t-distribution and is calculated as the sample size minus one.

Margin of Error
A numerical value that expresses the range within which the true value of the population parameter
is expected to fall.

t-distribution
A type of probability distribution that is symmetric, bell-shaped, and used when estimating
population parameters, particularly when the sample size is small and the population standard
deviation is unknown.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 62
 Confidence Intervals - Proportions
by Sophia

 WHAT'S COVERED

In this lesson, you will learn about how to construct and interpret a confidence interval for a population
proportion that can be applied in a business analytics environment. Specifically, this lesson will cover:
1. Introduction to Confidence Intervals for Proportions
1a. Confidence Intervals for Proportions
1b. Critical Values for Confidence Intervals for Proportions
1c. Calculating and Interpreting Confidence Intervals for Proportions

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational ideas of critical z-value and how to construct a confidence interval
for a proportion.

1. Introduction to Confidence Intervals for

Proportions
Suppose you want to find out how many individuals in the company you work for prefer online shopping rather
than physical stores. You ask 80 of your co-workers and 50 of them tell you they prefer shopping online. The
proportion of the sample of your co-workers that you surveyed who prefer shopping online is 62.5%, but that is
based on a sample, not the population. Due to sampling error, we cannot infer this to be true of the population.
To make inferences about the population, we must calculate a confidence interval. A confidence interval gives
you a range of values that is likely to contain the true proportion of a population. In this case, it helps us
estimate the true proportion of all individuals at your company who prefer online shopping.

1a. Confidence Intervals for Proportions

Just like how you create a confidence interval for a population mean (μ), you can also develop one for a
population proportion (p). The idea is the same: you're estimating a range of likely values for the true proportion
of a population based on a sample. This range, called the confidence interval, gives an estimate of where the
true population proportion is likely to be. The way you construct the range of values for the population
proportion, p, is:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 63
 The margin of error consists of two components:

1. Critical value
2. Standard error

The formula for constructing a confidence interval for the population proportion, p, is given by:

The z-value, also called the critical value, comes from the standard normal distribution. It works in a similar way
to the t critical value used in confidence intervals for the mean. The z-value tells you how many standard
deviations the sample proportion is from the true population proportion (p). This helps calculate the margin
of error, which determines the range of plausible values for the population proportion. For example, for a 95%
confidence interval, the z critical value would be 1.96, meaning you are looking at the range within 1.96 standard
deviations of the sample proportion, when constructing an interval for the population proportion, p.
The table below shows the z critical values for the three commonly used confidence intervals. Excel functions
for how to find these values are also provided.

Confidence Level z Critical Value Excel Formula

90% 1.645 =[Link](0.95)

95% 1.96 =[Link](0.975)

99% 2.576 =[Link](0.995)

The Excel function [Link]() returns the inverse of the standard normal cumulative distribution. This
function helps you find the z critical value for the specified confidence level.

The second part of the margin of error is the standard error for a proportion, which was introduced in a previous
tutorial. The standard error shows how much the sample proportion is likely to differ from the true
population proportion (p). A smaller standard error means the estimate is more precise.

1b. Critical Values for Confidence Intervals for Proportions

When you use the [Link]() function in Excel to find the z critical value for a 95% confidence level, you use
the argument 0.975. Here’s why:

1. Confidence Level: A 95% confidence level means you want to capture the middle 95% of the data in a
normal distribution. This leaves 5% of the data outside this range, split equally between the two tails (the
ends) of the distribution.
2. Splitting the Tails: Since the normal distribution is symmetrical, you split the 5% into two equal parts: 2.5%
in the lower tail and 2.5% in the upper tail.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 64
 3. Cumulative Probability: The [Link]() function requires a cumulative probability, which is the total area
under the curve up to a certain point. For the upper tail, we need to include the middle 95% plus the lower
2.5%, which adds up to 97.5% or 0.975.
4. Finding the z critical value: By using 0.975 as the argument in [Link](0.975), you are asking Excel to
find the z-score that corresponds to the point where 97.5% of the data lies below it. This z-score is 1.96,
which is the critical value for a 95% confidence level.

You can visualize this concept by examining a bell curve (standard normal distribution) as shown below, where:

The middle 95% of the data is shaded.

The remaining 5% is split into two tails: 2.5% on the left and 2.5% on the right.
The point that separates the middle 95% from the upper 2.5% is at 97.5% of the cumulative distribution.

By using 0.975 in the [Link]() function, you find the z critical value that marks this point.

The same concept applies to the 90% and 99% confidence intervals for the argument for the [Link]()
Excel function.

1c. Calculating and Interpreting Confidence Intervals for Proportions

Let’s look at an example of how you would calculate and interpret a confidence interval in a real-world scenario.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 65
 EXAMPLE You are a data analyst working for a social media manager for a new app. You survey 100
users and find that 60 of them love the new feature that has recently been added to the app.

Using the data in the new_app_feature.xlsx file, construct a 95% confidence interval for the true proportion
of all users who love the new feature in the app, p, by performing the following.

1. In cell D1, enter n (sample size). In cell E1, enter 100.

2. In cell D2, enter p-hat (sample proportion). In cell E2, enter the following formula:

=COUNTIF(B2:B101,”Yes”)/E1

3. In cell D3, enter critical value. In cell E3, enter the following formula:

=[Link](0.95)

4. In cell D4, enter standard error. In cell E4, enter the following formula:

=SQRT(((E2)*(1-E2))/E1)

5. In cell D6, enter Lower confidence interval limit. In cell E6, enter the following formula:

=E2 - ((E3)*(E4))

6. In cell D7, enter Upper confidence interval limit. In cell E7, enter the following formula:

=E2 + ((E3)*(E4))

The completed Excel worksheet along with the Excel formulas are shown in the screenshots below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 66
 Using Excel, you find a 95% confidence interval to be 50.4% and 69.6%.

You can be 95% confident that the true proportion of all users who love the new feature falls within this
range. This means that if we were to take many different samples of 100 users and calculate the confidence
interval for each sample, 95% of those intervals would contain the true proportion of users who love the
feature.

The confidence interval plays a crucial role in the business implications for this social media company by
providing a range within which the true proportion of all users who love the new feature is likely to fall.
Below is how the confidence interval relates to the various business implications:

User Satisfaction: The confidence interval indicates a significant portion of users are satisfied with the
new feature, which is a positive sign for the app’s development team.
Marketing Strategies: With a significant portion of users liking the feature, marketing campaigns can
highlight this feature to attract new users. Knowing that up to almost 70% of users might love the
feature provides a strong selling point.

 THINK ABOUT IT

What are some other practical business implications of the confidence interval that you could provide to the
social media manager?

 TRY IT

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 67
 You are a data analyst working for a bank that has recently launched a new mobile banking app. The bank
wants to know how satisfied customers are with the app. You surveyed 150 customers, and 105 of them
report that they are satisfied with the app. The data is in the Excel file named bank_mobile_app.xlsx.
Construct and interpret a 99% confidence interval for all the bank’s customers who are satisfied with the
new mobile banking app.

Solution: You can be 99% confident that the true proportion of all banking customers who are satisfied with
the new mobile app falls between 60.36% and 79.64%.

For guidance, the Excel formulas are shown in the screenshot below.

 WATCH

Check out this video on how to construct and interpret a 99% confidence interval for satisfied bank
customers.

 SUMMARY

In this lesson, you learned how to construct and interpret confidence intervals for a proportion. This
tutorial outlined the formula to calculate the confidence interval for a proportion. The formula contained
a critical value and the standard error. You learned that the critical value is a z-value from the standard
normal distribution. Using an example from a social media company implementing a new feature in

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 68
 their mobile app, you used Excel to construct and interpret a confidence interval for the proportion of
all users who love the new mobile app feature. In addition, practical implications were provided for how
the social media company could use the confidence interval to provide insights into the company’s user
retention efforts and marketing strategies.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 69
 Hypothesis Testing Basics
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to formulate a hypothesis test. Specifically, this lesson will cover:
1. Introduction to Hypothesis Testing
1a. Key Terms in Hypothesis Testing
1b. Steps in Hypothesis Testing
1c. Types of Hypothesis Tests
1d. Create Hypothesis Tests

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational ideas of one-tailed and two-tailed hypothesis testing.

1. Introduction to Hypothesis Testing

Hypothesis testing is a method used to make decisions or inferences about a population based on a sample of
data. Imagine you want to know if a new marketing strategy is more effective than the old one. Instead of
testing the strategy on every single customer, you can test it on a smaller group and use hypothesis testing to
draw conclusions about the entire customer base.

 TERM TO KNOW

Hypothesis Testing
A statistical method used to make decisions about a population based on sample data.

1a. Key Terms in Hypothesis Testing

1. Hypothesis: A statement that can be tested. There are two types:

a. Null Hypothesis (H₀): The default assumption that there is no effect or no difference.
b. Alternative Hypothesis (H₁): The statement you want to test, suggesting there is an effect or a difference.
Sometimes the alternative hypothesis is denoted as Ha.
2. Significance Level (α): The probability of rejecting the null hypothesis when it is true. Commonly set at 0.05
(5%) or 0.10 (10%).

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 70
 3. Test Statistic: A standardized value that is calculated from sample data during a hypothesis test. It is used
to decide whether to reject or fail to reject the null hypothesis.
4. P-value: The probability of obtaining test results at least as extreme as the observed results, assuming the
null hypothesis is true. A small p-value (typically ≤ 0.05) indicates strong evidence against the null
hypothesis.

 TERMS TO KNOW

Null Hypothesis
A statement used in statistical hypothesis testing that proposes there is no effect or no difference in a
population.

Alternative Hypothesis
A statement in statistical hypothesis testing that proposes there is an effect or a difference in the
population.

P-value
A measure used in statistical hypothesis testing to help determine the significance of your results.

1b. Steps in Hypothesis Testing

1. State the Hypotheses: Clearly define the null and alternative hypotheses.
2. Gather the Data: Gather data in a way that is designed to test the hypotheses.
3. Choose the Significance Level: Decide on the α level (for example, 0.05 or 0.10).
4. Perform a Statistical Test: Use an appropriate statistical test to analyze the data. Common tests include t-
tests, z-tests, and ANOVA. You will use a t-test for hypothesis testing for means.
5. Make a Decision: Based on the test results, decide whether to reject or fail to reject the null hypothesis.
This decision is guided by a p-value, which indicates the probability of observing the data if the null
hypothesis is true.
6. Interpret the Results: Explain the results of the hypothesis test in the context of the business problem.

 KEY CONCEPT

You apply the following rule when deciding whether to reject or fail to reject the null hypothesis.

If p-value (level of significance), reject the null hypothesis.

If p-value (level of significance), fail to reject the null hypothesis.

 DID YOU KNOW

In hypothesis testing, you never decide to accept the null hypothesis. The only two choices are to:

Reject the null hypothesis.

Fail to reject the null hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 71
 Below is a list of key reasons as to why you do not conclude to accept the null hypothesis.

1. You start by assuming the null hypothesis (H₀) is true. Hypothesis testing aims to find out if there is
sufficient evidence to challenge and reject this assumption.
2. Failing to reject the null hypothesis does not mean it is true; it just indicates there is insufficient
evidence to support the alternative hypothesis. It is like a “not guilty” verdict in court, which does not
confirm innocence, but rather that there is not enough proof of guilt.
3. Hypothesis tests aim to find evidence that contradicts the null hypothesis. If the data does not provide
strong enough evidence, you fail to reject the null hypothesis. This does not mean the null hypothesis
is true; it just means there is not enough evidence to disprove it.

 TRY IT

A data analyst is studying the impact of a new marketing campaign on monthly sales. The null hypothesis is
that the marketing campaign has no effect on monthly sales. The analyst conducts a study and obtains a p-
value of 0.042.

If the significance level is 0.05, should the analyst reject the null hypothesis? +

Yes, the analyst should reject the null hypothesis, since the p-value is less than or equal to the level of
significance. This means there is statistically significant evidence to suggest that the marketing
campaign has an effect on monthly sales.

 TRY IT

A financial data analyst is analyzing whether a new automated invoicing system reduces the average time
to process invoices. The null hypothesis is that the new system has no effect on the average processing
time. The analyst conducts a hypothesis test and obtains a p-value of 0.12.

If the significance level is 0.10, should the analyst reject the null hypothesis? +

No, the analyst should fail to reject the null hypothesis, since the p-value is greater than the level of
significance. This means there is not enough statistical evidence to conclude that the new automated
invoicing system reduces the average processing time.

1c. Types of Hypothesis Tests

Hypothesis tests are statistical methods used to make inferences about population parameters, such as the
population mean (μ) or population proportion (p). These tests help determine whether there is enough evidence
in a sample to infer that a certain condition holds for the entire population.

There are three main types of hypothesis tests.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 72
 Two-Tailed Test: This test checks for any significant difference from the null hypothesis in either direction
(greater than or less than). It’s used when the alternative hypothesis is that the population parameter is not
equal to a specific value

Right-Tailed Test: This test checks for a significant increase from the null hypothesis. It is used when the
alternative hypothesis is that the population parameter is greater than a specific value

Left-Tailed Test: This test checks for a significant decrease from the null hypothesis. It is used when the
alternative hypothesis is that the parameter is less than a specific value

In the next section, you will gain some practical experience developing each of these tests.

 TERMS TO KNOW

Two-Tailed Test
Type of hypothesis test used when the research question is to determine if the population parameter is
different from the null hypothesis value.

Right-Tailed Test
Type of hypothesis test used when the research question is to determine if the population parameter is
greater than the null hypothesis value.

Left-Tailed Test
Type of hypothesis test used when the research question is to determine if the population parameter is
less than the null hypothesis value.

1d. Create Hypothesis Tests

Let’s take a look at a couple of different scenarios and define the null and alternative hypotheses for each
situation. For each scenario provided, a statement of the null and alternative hypotheses is given, and whether
the test is two-tailed, left-tailed, or right-tailed.

EXAMPLE A company wants to determine if the average annual salary of its employees has changed
from the industry average of $63,500.

H₀: (average annual salary of the company’s employees is equal to $63,500)

H₁: (average annual salary of the company’s employees is not equal to $63,500)

This is a two-tailed test, because the company is interested in any significant difference, whether the salary
is higher or lower than $63,500.

EXAMPLE A financial auditor is investigating whether the average number of days to collect accounts
receivable has increased from the previous average of 30 days.

H₀: (average number of days to collect accounts receivable is equal to 30 days)

H₁: (average number of days to collect accounts receivable is greater than 30 days)

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 73
 This is a right-tailed test, because the auditor is specifically looking for an increase in the collection period.

EXAMPLE A CFO (Chief Financial Officer) wants to test if the implementation of a new expense tracking
system has decreased the average monthly operating costs, which were previously $200,000.

H₀: (average monthly operating costs are equal to $200,000)

H₁: (average monthly operating costs are less than $200,000)

This is a left-tailed test, because the CFO is specifically looking for a decrease in the average monthly
operating costs.

 SUMMARY

In this lesson, you learned that hypothesis testing is a statistical method used to make decisions about
a population based on sample data. It involves formulating two hypotheses: the null hypothesis (H₀),
which assumes no effect or difference, and the alternative hypothesis (H₁), which suggests there is an
effect or difference. The tutorial guided you through the steps in hypothesis testing, such as defining
the null and alternative hypotheses, selecting a significance level (α), gathering and analyzing sample
data, making a final decision using the results of the hypothesis test, and interpreting the hypothesis
test back in the context of the business problem. There are three types of hypothesis tests: two-tailed,
right-tailed, and left-tailed. For the three types of tests, you were provided with real-world business
scenarios, and you created a hypothesis test for each.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Alternative Hypothesis
A statement in statistical hypothesis testing that proposes there is an effect or a difference in the
population.

Hypothesis Testing
A statistical method used to make decisions about a population based on sample data.

Left-Tailed Test
Type of hypothesis test used when the research question is to determine if the population
parameter is less than the null hypothesis value.

Null Hypothesis
A statement used in statistical hypothesis testing that proposes there is no effect or no difference
in a population.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 74
 P-value
A measure used in statistical hypothesis testing to help determine the significance of your results.

Right-Tailed Test
Type of hypothesis test used when the research question is to determine if the population
parameter is greater than the null hypothesis value.

Two-Tailed Test
Type of hypothesis test used when the research question is to determine if the population
parameter is different from the null hypothesis value.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 75
 One-Sample Hypothesis Testing - Means
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to perform and interpret three different types of hypothesis tests for
population means. Specifically, this lesson will cover:
1. Introduction to One-Sample Hypothesis Testing for a Mean
1a. Two-Tailed Hypothesis Test for a Mean
1b. Excel [Link]() Functions for Calculating P-Values for Hypothesis Tests for Means
1c. Right-Tailed Hypothesis Test for a Mean
1d. Left-Tailed Hypothesis Test for a Mean

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational idea of one-sample hypothesis tests for population means.

1. Introduction to One-Sample Hypothesis Testing

for a Mean
In various business data analytics applications, determining whether a single group’s mean differs from a known
value is often necessary. For instance, a company might want to check if its average salary differs from the
industry average. This can be assessed using a one-sample hypothesis test for the mean.

A one-sample hypothesis test for the mean is a statistical method used to compare the mean of a single sample
to a known population mean. This test helps determine if the observed sample mean is significantly different
from the hypothesized population mean, or if the difference could be due to random variation.

Recall in the last tutorial, steps were provided to perform a hypothesis test.

1. State the Hypotheses: Clearly define the null and alternative hypotheses.
2. Gather the Data: Gather data in a way that is designed to test the hypotheses.
3. Choose the Significance Level: Decide on the α level (for example, 0.05 or 0.10).
4. Perform a Statistical Test: Use an appropriate statistical test to analyze the data. Common tests include t-
tests, z-tests, and ANOVA.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 76
 5. Make a Decision: Based on the test results, decide whether to reject or fail to reject the null hypothesis.
This decision is guided by a p-value, which indicates the probability of observing the data if the null
hypothesis is true.
6. Interpret the Results: Explain the results of the hypothesis test in the context of the business problem.

This tutorial is going to focus on performing a statistical test to help you make informed business decisions. You
will be guided through the process of constructing a test statistic and using the p-value to decide whether to
reject or fail to reject the null hypothesis for a hypothesis test for a mean.
There are three types of hypothesis tests for a mean. Each one is discussed in the upcoming sections.

1a. Two-Tailed Hypothesis Test for a Mean

Let us walk through a practical example of performing a two-tailed hypothesis test for a mean.

EXAMPLE Tech Innovators Inc. is a rapidly growing technology company known for its innovative
products and dynamic work environment. The company prides itself on offering competitive salaries to
attract top talent. Recently, the HR department has raised concerns that the average annual salary of
employees might have deviated from the industry average of $63,000. To address this concern, the
company has decided to conduct a hypothesis test to determine if there has been a significant change in
the average annual salary of its employees.

As a business data analyst, you have been tasked with performing a hypothesis test using the salary data of
Tech Innovators Inc. to determine if the average annual salary has changed from the industry average of
$63,000.

In performing this hypothesis test, you will complete the following steps.

Step 1: State the Hypotheses

H₀:
H₁:

Step 2: Gather the Data

In Excel, you have a sample of 100 Tech Innovators employee salaries. Using the data in Excel, you find:

(sample mean; average value of the employee salaries in the sample)

(sample standard deviation; amount of variability in the sample of employee salaries)
(sample size; number of observations in the sample)

Step 3: Choose a Significance Level

Use a level of significance,

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 77
 Step 4: Perform a Statistical Test

Since the population standard deviation, σ, is not known, you will use a t-test for this hypothesis test.

The t-test requires a test statistic to be computed. The test statistic is given by where μ₀ is the

hypothesized mean in the null hypothesis, $63,500 in this example. This test statistic tells you how many
standard deviations the sample mean is from the hypothesized mean of $63,500.

A test statistic of 2.16 means that the sample mean is 2.16 standard deviations above the hypothesized
population mean. This large value suggests that the difference between the sample mean and the
hypothesized population mean, μ₀, is quite substantial. In other words, the sample mean is far enough from
the hypothesized mean that it suggests there may be a meaningful difference between the two.

The test statistic directly relates back to the sampling distribution of The green histogram below
represents the distribution of sample means if you repeatedly sampled from the population with a mean of
The sampling distribution of is centered around the hypothesized mean of

and has a standard deviation (standard error) of

The blue dashed line at $64,480.76 represents the sample mean, The area to the right of the blue
dashed line represents the p-value, which is the probability of obtaining a test statistic as extreme as, or
more extreme than, the observed test statistic (2.16) if the null hypothesis were true.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 78
 The t-distribution is used to standardize the sample mean, to account for the sample size and variability.
The test statistic (t) is a point on the t-distribution, indicating how many standard deviations is from the
hypothesized mean, You use the t-distribution to find the p-value for the hypothesis test. The
graph below shows the t-distribution for a two-tailed hypothesis test. The orange dashed lines at
and show the observed test statistics. The orange shaded areas represent the p-value, indicating
the probability of observing test statistics as extreme as or more extreme, under the null hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 79
 In a two-tailed test, you look for extreme values on both ends of the distribution, because you want to see if
the sample mean is significantly different from the hypothesized mean (either higher or lower).

So, you have two critical values:

One at for the high end.

One at for the low end.

You will now use Excel to find the p-value for this two-tailed hypothesis test. Using the data in the
employees_salaries.xlsx file, perform the following.

1. In cell C2, enter hypothesized mean. In cell D2, enter 63500.

2. In cell C3, enter sample size. In cell D3, enter 100.

3. In cell C4, enter sample mean (x-bar). In cell D4, enter the following formula:

=AVERAGE(A2:A101)

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 80
 4. In cell C5, enter sample standard deviation (s). In cell D5, enter the following formula:

=STDEV.S(A2:A101)

5. In cell C7, enter standard error. In cell D7, enter the following formula:

=D5/SQRT(D3)

6. In cell C8, enter t test statistic. In cell D8, enter the following formula:

=(D4-D2)/D7

7. In cell C9, enter p-value (two-tailed test). In cell D9, enter the following formula:

=[Link].2T(D8,99)

In cell D9, you should obtain a p-value of 0.0332.

The [Link].2T() Excel function is used to calculate the two-tailed p-value of the t-distribution. The first
argument is the value of the test statistic and the second argument is the degrees of freedom for the t-
distribution, which is

Your Excel spreadsheet should contain these values for all of the computations.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 81
 Using Excel, you find the p-value to be 0.0332.

Step 5: Make a Decision

Since the p-value (level of significance), you can reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.0332) is less than the significance level (0.05), you reject the null hypothesis. This
means there is sufficient evidence to conclude that the mean salary of Tech Innovators employees is
significantly different from $63,500.

For guidance, the Excel formulas are shown in the screenshot below.

 TRY IT

An e-commerce company wants to ensure that the average delivery time for their orders is consistent with
their target delivery time of 3 days. They collect a random sample of delivery times (in days) for recent
orders to test if the average delivery time has deviated from the target.

The random sample is in the Excel file named e-commerce_delivery_times.xlsx. The delivery times are
measured in days using a decimal form. The decimal part of the number of days is the number of hours on
that particular day. For example, 3.2 means 3 days and 2 hours.

Using the data in the e-commerce_delivery_times.xlsx Excel file, perform a hypothesis test to determine if
the average delivery time for all orders at the e-commerce company is significantly different from the target
delivery time of 3 days. Use a level of significance of 0.05. Interpret the results of the hypothesis test.

Solution:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 82
 Null and alternative hypotheses:

H₀:

H₁:

Using Excel, you find the p-value for this test to be 0.2702. The Excel worksheets with the values of the test
statistic and p-value are provided below.

You notice that the test statistic is a negative value for this problem, -1.12, meaning that the sample mean of
2.92 days is 1.12 standard deviations below the hypothesized mean of 3 days.

If the test statistic is negative, you will need to enclose the t test statistic in the ABS() function (absolute
value) in Excel when using the [Link].2T() function to find the p-value.

The Excel worksheet with values and formulas is shown below:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 83
 Interpretation: Since the p-value (0.2702) is greater than the significance level (0.05), you fail to reject the
null hypothesis. Based on the sample of delivery times collected, you do not have enough statistical
evidence to conclude that the average delivery time for all packages is different from 3 days.

 WATCH

Follow along with this video on analyzing average delivery time with a two-tailed hypothesis test.

 TERM TO KNOW

Test Statistic
A standardized value calculated from sample data for a hypothesis test that measures how much your
sample data deviates from the null hypothesis.

1b. Excel [Link]() Functions for Calculating P-Values for Hypothesis Tests for Means
This table provides a guide on which Excel [Link]() function to use for calculating p-values in different types of t-
tests. It includes functions for two-tailed, right-tailed, and left-tailed tests, specifying the appropriate function
and a brief description of each. The value of t in the table represents the test statistic.

Type of Test Excel Function Description

Calculates the two-tailed p-value for the t-distribution

[Link].2T(t, df)
Two-tailed when the t test statistic is positive

[Link].2T(ABS(t), df) P-value calculation when test statistic is negative

Right-tailed [Link](t, df) Calculates the right-tailed p-value for the t-distribution

Left-tailed [Link](t, df, TRUE) Calculates the left-tailed p-value for the t-distribution

1c. Right-Tailed Hypothesis Test for a Mean

Let’s walk through a practical example of a right-tailed hypothesis test for a mean.

EXAMPLE Financial Solutions Inc. is a well-established financial services company that prides itself on
efficient operations and strong financial health. One key performance metric the company monitors closely
is the average number of days it takes to collect accounts receivable. Historically, the company has
maintained an average collection period of 30 days. Recently, there have been concerns that this period
might have increased, potentially impacting cash flow and operational efficiency.

As a business data analyst, you have been tasked with analyzing the accounts receivable collection data to
determine if the average number of days to collect accounts receivable has increased from the historical
average of 30 days.

In performing this hypothesis test, you will complete the following steps.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 84
 Step 1: State the Hypotheses

H₀:
H₁:

Step 2: Gather the Data

In Excel, you have a sample of the number of days it took to collect accounts receivable for 100 different
accounts. Using the data in Excel, you find:

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Perform a Statistical Test

Once again, since σ is not known, you will use the t-test to conduct this test. The test statistic is given by

where μ₀ is the hypothesized mean in the null hypothesis, 30 days in this example.

A test statistic of 1.32 means that the sample mean is 1.32 standard deviations above the hypothesized
population mean.

You use the t-distribution to find the p-value for the right-tailed hypothesis test. The graph below shows the
t-distribution for a right-tailed hypothesis test. The orange dashed line at shows the observed test
statistic. The orange shaded area represents the p-value, indicating the probability of observing the test
statistic, or something more extreme, if the null hypothesis is true.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 85
 In a right-tailed test, you look for extreme values to the right end of the distribution, because you want to
see if the sample mean is significantly greater than the hypothesized mean.

You will now use Excel to find the p-value for this right-tailed hypothesis test. Using the data in the
accounts_receivable_collection_days.xlsx file, perform the following.

For guidance, the Excel formulas are shown in the screenshot below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 86
 1. In cell C2, enter hypothesized mean. In cell D2, enter 30.

2. In cell C3, enter sample size. In cell D3, enter 100.

3. In cell C4, enter sample mean (x-bar). In cell D4, enter the following formula:

=AVERAGE(A2:A101)

4. In cell C5, enter sample standard deviation (s). In cell D5, enter the following formula:

=STDEV.S(A2:A101)

5. In cell C7, enter standard error. In cell D7, enter the following formula:

=D5/SQRT(D3)

6. In cell C8, enter t test statistic. In cell D8, enter the following formula:

=(D4-D2)/D7

7. In cell C9, enter p-value (right-tailed test). In cell D9, enter the following formula:

=[Link](D8,99)

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 87
 The [Link]() Excel function is used to calculate the p-value for this right-tailed test. The first argument is
the value of the test statistic and the second argument is the degrees of freedom for the t-distribution,
which is

Your Excel spreadsheet should contain these values for all of the computations.

Step 5: Make a Decision

Since the p-value (level of significance), you fail to reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.0946) is greater than the significance level (0.05), you fail to reject the null hypothesis.
This means there is not enough evidence to conclude that the mean number of days to collect accounts
receivable is greater than 30 days.

 TRY IT

You are working as a data analyst for an insurance company. The company wants to ensure that the
average claim amount for a specific type of insurance policy does not exceed $5,000 for all their policies.
They collect a sample of claim amounts (in dollars) for recent claims to test if the average claim amount for
all their policies has exceeded this threshold.

Using the data in the claim_amounts.xlsx Excel file, perform a hypothesis test to determine if the average
claim amount for these policies exceeds $5,000. Use a level of significance of 0.05. Interpret the results of
the hypothesis test.

Solution:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 88
 Null and alternative hypotheses:

H₀:

H₁:

Using Excel, you find the p-value for this test to be 0.0129. The Excel worksheets with the values of the test
statistic and p-value are provided below.

Interpretation: Since the p-value (0.0129) is less than the significance level (0.05), you reject the null
hypothesis. Based on the sample of claim amounts, you have evidence to conclude that the average claim
amount for all claims for the insurance company is more than $5,000.

 WATCH

Follow along with this video on performing a right-tailed hypothesis test on average claim amount.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 89
 1d. Left-Tailed Hypothesis Test for a Mean
Let’s walk through a practical example of a left-tailed hypothesis test for a mean.

EXAMPLE GreenTech Solutions, a leading company in sustainable technology, recently implemented a

new expense tracking system aimed at reducing operating costs. The CFO (Chief Financial Officer) is keen
to evaluate whether this new system has effectively decreased the average monthly operating costs, which
were previously $200,000. To assess the impact, the CFO has decided to conduct a hypothesis test.

As a business data analyst, you have been tasked with analyzing the monthly operating costs data to
determine if the implementation of the new expense tracking system has significantly decreased the
average monthly operating costs from the previous average of $200,000.

In performing this hypothesis test, you will complete the following steps.

Step 1: State the Hypotheses

H₀:
H₁:

Step 2: Gather the Data

In Excel, you have a sample of 100 months of operating costs. Using the data in Excel, you find:

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Perform a Statistical Test

Once again, since σ is not known, you will use the t-test to conduct this test. The test statistic is given by

where μ₀ is the hypothesized mean in the null hypothesis, $200,000 in this example.

A test statistic of -1.71 means that the sample mean is 1.71 standard deviations below the hypothesized
population mean.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 90
 You use the t-distribution to find the p-value for the left-tailed hypothesis test. The graph below shows the t-
distribution for a left-tailed hypothesis test. The orange dashed line at shows the observed test
statistic. The orange shaded area represents the p-value, indicating the probability of observing the test
statistic or something more extreme if the null hypothesis is true.

In a left-tailed test, you look for extreme values to the left end of the distribution, because you want to see if
the sample mean is significantly less than the hypothesized mean.

You will now use Excel to find the p-value for this left-tailed hypothesis test. Using the data in the
monthly_operating [Link] file, perform the following.

For guidance, the Excel formulas are shown in the screenshot below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 91
 1. In cell C2, enter hypothesized mean. In cell D2, enter 200000.

2. In cell C3, enter sample size. In cell D3, enter 100.

3. In cell C4, enter sample mean (x-bar). In cell D4, enter the following formula:

=AVERAGE(A2:A101)

4. In cell C5, enter sample standard deviation (s). In cell D5, enter the following formula:

=STDEV.S(A2:A101)

5. In cell C7, enter standard error. In cell D7, enter the following formula:

=D5/SQRT(D3)

6. In cell C8, enter t test statistic. In cell D8, enter the following formula:

=(D4-D2)/D7

7. In cell C9, enter p-value (right-tailed test). In cell D9, enter the following formula:

=[Link](D8,99,TRUE)

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 92
 The [Link]() Excel function is used to calculate the p-value for this left-tailed test. The first argument is the
value of the test statistic and the second argument is the degrees of freedom for the t-distribution, which is
The third argument is a logical value that specifies to calculate the cumulative probability up to the
test statistic, which is the left-tailed p-value for the t-distribution.

Step 5: Make a Decision

Since the p-value (level of significance), you reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.0452) is less than the significance level (0.05), you reject the null hypothesis. This
means there is enough evidence to conclude that the mean operating cost is less than $200,000.

 TRY IT

You are a data analyst at a company that is concerned about employee work-life balance. The company
believes that employees should work an average of 40 hours per week. To ensure this, you have collected
data on the number of hours worked per week by 50 employees. You want to test if the average work
hours per week is less than 40 hours for all employees.

Using the data in the employee_work_hours.xlsx Excel file, perform a hypothesis test to test if the average
number of hours worked for this sample of employees is less than 40 hours per week. Use a level of
significance of 0.05. Interpret the results of the hypothesis test.

Solution:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 93
 Null and alternative hypotheses:

H₀:

H₁:

Using Excel, you find the p-value for this test to be 0.0765. The Excel worksheets with the values of the test
statistic and p-value are provided below.

Interpretation: Since the p-value (0.0765) is less than the significance level (0.05), you fail to reject the null
hypothesis. There is not enough statistical evidence to conclude that the average work hours per week for
all employees is less than 40 hours.

 WATCH

Check out this video on performing a left-tailed hypothesis test on the average amount of hours worked.

 SUMMARY

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 94
 In this lesson, you were provided with a comprehensive guide on performing and interpreting
hypothesis tests for population means. The focus was on three types of tests: two-tailed, right-tailed,
and left-tailed tests for a population mean. You were provided an outline for performing a hypothesis
test for a population mean, such as defining the null and alternative hypotheses, gathering data,
choosing a significance level, performing the statistical test, making a decision, and interpreting the
results of the test. Practical examples were provided for each test type of test including a two-tailed test
for employee salaries, a right-tailed test related to accounts receivable, and a left-tailed test performed
on operating costs for a company. The tutorial also included instructions for using Excel’s [Link]()
function to calculate the p-values for each of these tests and provided hands-on exercises to reinforce
performing each of the three types of hypothesis tests for population means.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Test Statistic
A standardized value calculated from sample data for a hypothesis test that measures how much
your sample data deviates from the null hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 95
 One-Sample Hypothesis Testing - Proportions
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to perform and interpret three different types of hypothesis tests for
population proportions. Specifically, this lesson will cover:
1. Introduction to One-Sample Hypothesis Testing for a Proportion
1a. Types of Hypothesis Tests for Proportions
2. Hypothesis Testing for Proportions: Two-Tailed, Right-Tailed, and Left-Tailed Tests
2a. Two-Tailed Hypothesis Test for a Proportion
2b. Excel [Link]() Functions for Calculating P-Values
2c. Right-Tailed Hypothesis Test for a Proportion
2d. Left-Tailed Hypothesis Test for a Proportion

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational idea of one-sample hypothesis tests for population proportions.

1. Introduction to One-Sample Hypothesis Testing

for a Proportion
Hypothesis testing is a method used to make decisions or draw conclusions about a population based on
sample data. In the context of hypothesis testing for proportions, the focus is on the percentage of a certain
characteristic within a group. For instance, if a company wants to determine if more than 50% of its customers
are satisfied with their service, hypothesis testing can be employed. This process involves setting up a
hypothesis, collecting data, and using statistical methods to analyze whether the data supports the hypothesis.

In a similar fashion as in the last tutorial, the steps to perform a hypothesis test are as follows:

1. State the Hypotheses: Clearly define the null and alternative hypotheses.
2. Gather the Data: Gather data in a way that is designed to test the hypotheses.
3. Choose the Significance Level: Decide on the α level (for example, 0.05 or 0.10).
4. Perform a Statistical Test: Use an appropriate statistical test to analyze the data. You will use a z-test for
hypothesis testing for proportions.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 96
 5. Make a Decision: Based on the test results, decide whether to reject or fail to reject the null hypothesis.
This decision is guided by a p-value, which indicates the probability of observing the data if the null
hypothesis is true.
6. Interpret the Results: Explain the results of the hypothesis test in the context of the business problem.

This tutorial is going to focus on performing a statistical test to help you make informed business decisions. You
will be guided through the process of constructing a test statistic and using the p-value to decide whether to
reject or fail to reject the null hypothesis for a hypothesis test for a proportion.

1a. Types of Hypothesis Tests for Proportions

Just like with population means, there are three types of hypothesis tests for population proportions, based on
the direction of the test. A description of each type of test is provided below.

1. Two-Tailed Test:

Purpose: To determine if the sample proportion is significantly different from the hypothesized population
proportion, either higher or lower.
Hypotheses:
Null Hypothesis H₀: (the population proportion is equal to a hypothesized population proportion)
Alternative Hypothesis H₁: (the population proportion is not equal to a hypothesized population
proportion)

EXAMPLE A company wants to test if the proportion of satisfied customers is different from 50%.
2. Right-Sided (One-Tailed) Test:

Purpose: To determine if the sample proportion is significantly greater than the hypothesized population
proportion.
Hypotheses:
Null Hypothesis H₀: (the population proportion is equal to a hypothesized population proportion)
Alternative Hypothesis H₁: (the population proportion is greater than a hypothesized population
proportion)

EXAMPLE A company wants to test if more than 60% of its customers are satisfied with their service.
3. Left-Sided (One-Tailed) Test:

Purpose: To determine if the sample proportion is significantly less than the hypothesized population
proportion.
Hypotheses:
Null Hypothesis H₀: (the population proportion is equal to a hypothesized population proportion)
Alternative Hypothesis H₁: (the population proportion is less than a hypothesized population
proportion)

EXAMPLE A company wants to test if less than 40% of its customers are dissatisfied with their service.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 97
 2. Hypothesis Testing for Proportions: Two-Tailed,
Right-Tailed, and Left-Tailed Tests
In the upcoming sections, you will explore the different types of hypothesis tests used in hypothesis tests for
proportions: two-tailed, right-tailed, and left-tailed tests. You will discover the purpose of each test, how to
perform the test, and how to interpret the results. Whether you are analyzing financial metrics or customer
satisfaction surveys, mastering these tests will enhance your analytical skills and decision-making capabilities.

2a. Two-Tailed Hypothesis Test for a Proportion

Let us walk through a practical example of performing a two-tailed hypothesis test for a proportion.

EXAMPLE SavvyShoppers, a retail company, has recently launched a new customer loyalty program
aimed at increasing customer retention and sales. The program offers various incentives, such as discounts,
exclusive offers, and reward points for frequent purchases. The company wants to evaluate whether the
new loyalty program has significantly changed the customer conversion rate compared to the historical
conversion rate.

You need to conduct a two-tailed hypothesis test to determine if the new loyalty program has had a
significant impact on the customer conversion rate (percentage of customers who make purchases after a
new loyalty program is implemented).

In performing this hypothesis test, you will complete the following steps:

Step 1: State the Hypotheses

H₀: (conversion rate is equal to the historical conversion rate of the hypothesized
proportion, )
H₁: (conversion rate is not equal to the historical conversion rate of )

Step 2: Gather the Data

In Excel, you have a sample of 100 customers that indicates whether the customer converted or not (that is,
made a purchase or not) after the loyalty program was introduced. Using the data in Excel, you find:

(the sample proportion of customers who converted (made a purchase) after the loyalty
program was implemented)

Step 3: Choose a Significance Level

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 98
 Use a level of significance,

Step 4: Perform a Statistical Test

For hypothesis tests for proportions, you use a test statistic from a standard normal distribution, a z-test
statistic.

where is the hypothesized proportion in the null hypothesis, which is 0.50, and is

the sample proportion, which is 0.44 in this example. The test statistic tells you how many standard
deviations the sample proportion is from the hypothesized proportion of 0.50.

A test statistic of -1.2 means that the sample proportion is 1.2 standard deviations below the hypothesized
population proportion.

Just like for hypothesis tests for the mean, the test statistic directly relates back to the sampling distribution
of the proportion. The green histogram below illustrates the sampling distribution of showing how the
proportions of buyers from numerous samples of 100 people each form a distribution. The distribution is
centered around the hypothesized population proportion, and has a standard error calculated
using:

The blue dashed line at 0.44 represents the sample proportion of 0.44. The area under the distribution to
the right of the blue dashed line represents the p-value, which is the probability of obtaining a test statistic
as extreme as, or more extreme than, the observed test statistic of -1.2, if the null hypothesis were true.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 99
 The standard normal distribution is used to standardize the sample proportion. The test statistic (z)
represents a specific point on the standard normal distribution. It shows how many standard deviations the
sample proportion, is from the hypothesized population proportion, You use the
standard normal distribution to find the p-value for the hypothesis test. The graph below shows the
standard normal distribution for a two-tailed hypothesis test. The orange dashed lines at and
show the observed test statistics. The orange shaded areas represent the p-value, indicating the
probability of observing test statistics as extreme as or more extreme, under the null hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 100
 In a two-tailed test, you look for extreme values on both ends of the distribution, because you want to see if
the sample proportion is significantly different from the hypothesized proportion, either higher or lower.

So, you have two critical values:

One at for the high end.

One at for the low end.

You will now use Excel to find the p-value for this two-tailed hypothesis test, using the data in the
conversion_rate.xlsx file. The column Converted in the data represents if a particular customer made a
purchase after the loyalty program was implemented; 1 means a purchase was made and 0 means no
purchase was made.

Perform the following.

1. In cell C3, enter sample size. In cell D3, enter 100.

2. In cell C4, enter sample proportion (p-hat). In cell D4, enter the following formula:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 101
 =COUNTIF(A2:A101,1)/COUNT(A2:A101)

This formula calculates the proportion (or percentage) of cells in the range A2:A101 that contain the value 1.

3. In cell C6, enter standard error. In cell D6, enter the following formula:

=SQRT((D2*(1-D2))/D3)

4. In cell C7, enter z test statistic. In cell D7, enter the following formula:

=(D4-D2)/D6

5. In cell C8, enter absolute value z test statistic. In cell D8, enter the following formula:

=ABS(D7)

6. In cell C9, enter p-value (two-tailed test). In cell D9, enter the following formula:

=2 * (1 - [Link](D8, TRUE))

In cell D8, you should obtain a p-value of 0.2301.

Let’s Break Down the Excel P-Value Calculation

Let’s explain the components of this Excel formula for finding the p-value.

The [Link]() function in Excel is used to work with the standard normal distribution, which is a
special case of the normal distribution with a mean of 0 and a standard deviation of 1. The [Link]()
function returns the probability that a standard normal random variable is less than or equal to a given value
(z). This is useful for finding probabilities and p-values in hypothesis testing.

The syntax for the function is:

[Link](z, cumulative) where:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 102
 z: The z-test statistic (or z-score) for which you want to find the probability. The z-value represents the
number of standard deviations the sample proportion is from the hypothesized proportion, in the
context of hypothesis testing for proportions.
cumulative: A logical value (TRUE or FALSE) that determines the type of probability you want to
calculate. When you set cumulative to TRUE, the function calculates the probability of observing a test
statistic less than the observed z-test statistic. This means it gives you the total area under the curve to
the left of the z-test statistic. Essentially, it tells you the likelihood of observing a value up to and
including the z-test statistic.

For guidance, the Excel formulas are shown in the screenshot below.

Now, let’s break down the components of the Excel formula used to calculate the two-sided p-value piece
by piece.

=2 * (1 - [Link](1.2, TRUE))

[Link](1.2, TRUE):
Purpose: This part of the formula calculates the probability that a standard normal variable (z) is less
than or equal to 1.2.
Explanation: It gives the area under the standard normal curve to the left of the z-test statistic 1.2.
Essentially, it tells us how much of the distribution lies to the left of 1.2.

1 - [Link](1.2, TRUE):
Purpose: This part calculates the probability that a standard normal variable (z) is greater than 1.2.
Explanation: Since the total area under the curve is 1 (representing 100% probability), subtracting the
area to the left of 1.2 from 1 gives us the area to the right of 1.2.

2 * (1 - [Link](1.2, TRUE)):
Purpose: This part calculates the p-value for a two-tailed test.
Explanation: In a two-tailed test, you are interested in extreme values on both ends of the
distribution. Therefore, you double the area in one tail to account for both tails. This gives the total

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 103
 probability of observing a value as extreme as 1.2, or more extreme, in either direction (both positive
and negative).

Using Excel, you find the p-value to be 0.2301.

Step 5: Make a Decision

Since the p-value (level of significance), you fail to reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.2301) is greater than the significance level (0.05), you fail to reject the null hypothesis.
There is not enough evidence to conclude that the new loyalty program has significantly changed the
customer conversion rate from the historical conversion rate of 50%.

 TRY IT

A bank has recently introduced a new premium banking service aimed at attracting more customers.
Historically, the opt-in rate for new services has been 40%. The bank wants to evaluate whether the new
premium service has significantly changed the opt-in rate.

The customer_opt_in_banking_service.xlsx Excel file contains a column named Opted_In, which indicates
whether a customer opted in for the new premium banking service or not. A 1 means the customer opted in
for the new service, and a 0 means the customer did not opt in for the new service.

Using the data in the customer_opt_in_banking_service.xlsx Excel file, conduct a hypothesis test to
determine if the opt-in rate for the new banking service is significantly different from the historical rate of
40%. Use a significance level of 0.05. Interpret the results of the hypothesis test.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 104
 Solution:

Null and alternative hypotheses:

H₀:

H₁:

Using Excel, you find the p-value for this test to be 0.0433. The Excel worksheets with the values of the test
statistic and p-value are provided below.

Interpretation: Since the p-value (0.0433) is less than the significance level (0.05), you reject the null
hypothesis. Based on the sample of customers, there is sufficient statistical evidence to conclude that the
opt-in rate for the new banking service is significantly different from the historical rate of 40%. Therefore, we
can infer that the new premium banking service has had a significant impact on the customer opt-in rate.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 105
  WATCH

Check out this video on conducting a two-tailed hypothesis test to analyze the opt-in rate.

2b. Excel [Link]() Functions for Calculating P-Values

This table provides a guide on which Excel [Link]() function to use for calculating p-values in different
types of z-tests. It includes functions for two-tailed, right-tailed, and left-tailed tests, specifying the appropriate
function and a brief description of each. The value of z in the table represents the test statistic.

Type of Test Excel Function Description

Two-Tailed Test =2 * (1 - [Link](ABS(z), TRUE)) This function calculates the p-value for a two-
tailed test by doubling the area in one tail.
ABS(z) ensures the z-test statistics is positive,
and [Link](z, TRUE) gives the cumulative
probability up to the z-test statistic.

Right-Tailed Test =1 - [Link](z, TRUE) This function calculates the area to the right of
the z-test statistic for a right-tailed test.
[Link](z, TRUE) gives the cumulative
probability up to the z-test statistic, and
subtracting from 1 gives the right tail area.

Left-Tailed Test =[Link](z, TRUE) This function calculates the cumulative

probability up to the z-test statistic for a one-
tailed test. If the test is left-tailed, use the z-test
statistic directly. No absolute value is needed.

2c. Right-Tailed Hypothesis Test for a Proportion

Let’s walk through a practical example of a right-tailed hypothesis test for a proportion.

EXAMPLE A university wants to determine if the new online class format has significantly increased
student preference for online classes, compared to a historical preference rate of 55%. After implementing
the new online class format, the university conducted a survey of 100 students to see if the preference rate
has increased.

You need to conduct a right-tailed hypothesis test to determine if the new online class format has
significantly increased student preference for online classes, compared to the historical preference rate of
55%. This test will help you evaluate whether the proportion of students who prefer online classes is
greater than 55% after the implementation of the new format.

In performing this hypothesis test, you will complete the following steps.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 106
 Step 1: State the Hypotheses

H₀: (student online preference rate for online classes is equal to the historical preference rate
of the hypothesized proportion, )
H₁: (student online preference rate is not equal to the historical online preference rate of
)

Step 2: Gather the Data

In Excel, you have a sample of 100 student responses from a survey that indicates whether the student
prefers an online or in-person format class. Using the data in Excel, you find:

(the sample proportion of students who prefer the online class format)

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Perform a Statistical Test

For hypothesis tests for proportions, you use a test statistic from a standard normal distribution, a z-test
statistic.

where is the hypothesized proportion in the null hypothesis, which is 0.55, and is

the sample proportion, which is 0.65 in this example. The test statistic tells you how many standard
deviations the sample proportion is from the hypothesized proportion of 0.55.

A test statistic of 2 means that the sample proportion is 2 standard deviations above the hypothesized
population proportion.

You use the standard normal distribution to find the p-value for the right-tailed hypothesis test. The graph
below shows the standard normal distribution for a right-tailed hypothesis test. The orange dashed line at
shows the observed test statistic. The orange shaded area represents the p-value, indicating the
probability of observing the test statistic, or something more extreme, if the null hypothesis is true.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 107
 In a right-tailed test, you look for extreme values to the right end of the distribution, because you want to
see if the sample proportion is significantly greater than the hypothesized mean.

You will now use Excel to find the p-value for this right-tailed hypothesis test, using the data in the
student_preferences.xlsx file.

For guidance, the Excel formulas are shown in the screenshot below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 108
 Perform the following.

1. In cell C2, enter hypothesized proportion. In cell D2, enter 0.55.

2. In cell C3, enter sample size. In cell D3, enter 100.

3. In cell C4, enter sample proportion (p-hat). In cell D4, enter the following formula:

=COUNTIF(A2:A101,”Online”)/COUNTA(A2:A101)

This formula calculates the proportion (or percentage) of cells in the range A2:A101 that contain the value
“Online.”

4. Notice that the denominator for this sample proportion is using the Excel function COUNTA() and not
COUNT(). COUNT() is used to count numeric values. COUNTA() is used to count any data type in an Excel
cell. Since the data type in the Preference column is text, COUNTA() is the appropriate function to use to
count the number of students that prefer the online format.

5. In cell C6, enter standard error. In cell D6, enter the following formula:

=SQRT((D2*(1-D2))/D3)

6. In cell C7, enter z test statistic. In cell D7, enter the following formula:

=(D4-D2)/D6

7. In cell C8, enter p-value (right-tailed test). In cell D8, enter the following formula:

=[Link](D7,TRUE)

Your Excel spreadsheet should contain these values for all of the computations.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 109
 Step 5: Make a Decision

Since the p-value (level of significance), you reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.0222) is less than the significance level (0.05), you reject the null hypothesis. This
means there is sufficient evidence to conclude that the proportion of students who prefer online classes is
significantly greater than the historical preference rate of 55%.

 TRY IT

A tech company wants to determine if a new customer service chatbot has significantly increased customer
satisfaction compared to the previous quarter. The company surveyed 150 customers after implementing
the new chatbot to see if the satisfaction rate had increased. The previous quarter’s satisfaction rate was
50% (that is, 50% of customers reported being satisfied after a customer service encounter).

The Excel sheet named customer_satisfaction_chatbot.xlsx contains survey data from 150 customers
regarding their satisfaction with a new customer service chatbot. There are two columns in the data.
Customer_ID is a unique identifier for each customer. Satisfaction is the satisfaction status of each
customer with two possible values:

Satisfied: Indicates customer was satisfied with the customer service provided by the new chatbot.
Not Satisfied: Indicates customer was not satisfied with the customer service provided by the new
chatbot.

Using the data in the customer_satisfaction_chatbot.xlsx Excel file, conduct a hypothesis test to determine
if the new chatbot has significantly increased customer satisfaction compared to the previous quarter’s
satisfaction rate of 50%. Use a significance level of 0.05. Interpret the results of the hypothesis test.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 110
 The Excel worksheets with the values of the test statistic and p-value are provided below.

Solution:

Null and alternative hypotheses:

H₀:

H₁:

Using Excel, you find the p-value for this test to be 0.0072.

Interpretation: Since the p-value (0.0072) is less than the significance level (0.05), you reject the null
hypothesis. Based on the sample of customers, there is sufficient statistical evidence to conclude that the
new customer service chatbot has increased customer satisfaction compared to the previous quarter’s
satisfaction rate of 50%.

 WATCH

Follow along with this video on conducting a right-tailed hypothesis test to determine chatbot impact on
customer satisfaction.

2d. Left-Tailed Hypothesis Test for a Proportion

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 111
 Let’s walk through a practical example of a left-tailed hypothesis test for a proportion.

EXAMPLE You are working as a data analyst for BrightPay Credit, a credit card company. Historically,
80% of customers pay their credit card bills on time. Recently, there have been concerns that the proportion
of on-time payments might have decreased. To investigate this claim, you take a sample of 100 customers
and record whether each customer paid their bill on time. You need to conduct a left-tailed hypothesis test
to determine if the company’s concern is valid.

In performing this hypothesis test, you will complete the following steps.

Step 1: State the Hypotheses

H₀: (proportion of on-time payments is equal to the historical rate of the

hypothesized proportion, )
H₁: (proportion of on-time payments is less than the historical rate of )

Step 2: Gather the Data

In Excel, you have a sample of 100 customers and an indicator variable that denotes whether the customer
has paid their bill on time or not for the last billing cycle. Using the data in Excel, you find:

(the sample proportion of customers who have paid their bill on time)

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Perform a Statistical Test

For hypothesis tests for proportions, you use a test statistic from a standard normal distribution, a z-test
statistic.

where is the hypothesized proportion in the null hypothesis, which is 0.80, and is

the sample proportion, which is 0.70 in this example. The test statistic tells you how many standard
deviations the sample proportion is from the hypothesized proportion of 0.80.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 112
 A test statistic of -2.5 means that the sample proportion is 2.5 standard deviations below the hypothesized
population proportion.

You use the standard normal distribution to find the p-value for the left-tailed hypothesis test. The graph
below shows the standard normal distribution for a left-tailed hypothesis test. The orange dashed line at
shows the observed test statistic. The orange shaded area represents the p-value, indicating the
probability of observing the test statistic, or something more extreme, if the null hypothesis is true.

In a left-tailed test, you look for extreme values to the left end of the distribution, because you want to see if
the sample proportion is significantly less than the hypothesized mean.

You will now use Excel to find the p-value for this left-tailed hypothesis test, using the data in the
on_time_payments.xlsx file. A value of 1 in the OnTime_Payment column indicates that a customer paid on
time. A value of 0 indicates that a customer paid late.

For guidance, the Excel formulas are shown in the screenshot below.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 113
 Perform the following.

1. In cell C2, enter hypothesized proportion. In cell D2, enter 0.80.

2. In cell C3, enter sample size. In cell D3, enter 100.

3. In cell C4, enter sample proportion (p-hat). In cell D4, enter the following formula:

=COUNTIF(A2:A101, 1)/COUNT(A2:A101)

This formula calculates the proportion (or percentage) of customers with on-time payments.

4. In cell C6, enter standard error. In cell D6, enter the following formula:

=SQRT((D2*(1-D2))/D3)

5. In cell C7, enter z test statistic. In cell D7, enter the following formula:

=(D4-D2)/D6

6. In cell C8, enter p-value (left-tailed test). In cell D8, enter the following formula:

=[Link](D7,TRUE)

Your Excel spreadsheet should contain these values for all of the computations.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 114
 Step 5: Make a Decision

Since the p-value (level of significance), you reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.0062) is less than the significance level (0.05), you reject the null hypothesis. This
means there is evidence to conclude that the current proportion of on-time payments is significantly less
than the historical rate of 80%.

 TRY IT

A financial firm wants to determine if a new investment strategy has significantly decreased the proportion
of investments that meet or exceed the target return compared to the previous quarter. In the past, the
proportion of investments that meet or exceed a target return has been 50%. The firm collected a sample of
200 investments after implementing the new strategy to see if the proportion had decreased from the
historical rate of 50%.

The Excel sheet named investment_returns.xlsx contains a random sample of 200 investments regarding
their performance relative to a target return. There are two columns in the data. Investment_ID is a unique
identifier for each investment. Investment_Return is the performance status of each investment, with two
possible values:

Above Target: Investment met or exceeded the target return.

Below Target: Investment did not meet the target return.

Using the data in the investment_return.xlsx Excel file, conduct a hypothesis test to determine if the
proportion of investments meeting or exceeding the target return is less than 50%. Use a significance level
of 0.05. Interpret the results of the hypothesis test.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 115
 Solution:

Null and alternative hypotheses:

H₀:

H₁:

Using Excel, you find the p-value for this test to be 0.2398. The Excel worksheets with the values of the test
statistic and p-value are provided below.

Interpretation: Since the p-value (0.2398) is greater than the significance level (0.05), you fail to reject the
null hypothesis. Based on the sample of investments, there is not enough evidence to conclude that the
new investment strategy has significantly decreased the proportion of investments meeting or exceeding
the historical target return of 50%.

 WATCH

Check out this video on conducting a left-tailed hypothesis test to analyze investment strategies.

 SUMMARY

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 116
 In this lesson, you were provided with a comprehensive guide on performing and interpreting
hypothesis tests for proportions. The focus was on three types of tests: two-tailed, right-tailed, and
left-tailed tests for a proportion. You were provided an outline for performing a hypothesis test for a
population proportion, such as defining the null and alternative hypotheses, gathering data, choosing a
significance level, performing the statistical test, making a decision, and interpreting the results of the
test. Practical examples were provided for each test type of test, including a two-tailed test for
customer conversion rates, a right-tailed test related to the modality of university class offerings, and a
left-tailed test on payment data for a credit card company. The tutorial also included instructions for
using Excel’s [Link]() function to calculate the p-values for each of these tests and provided
hands-on exercises to reinforce performing each of the three types of hypothesis tests for proportions.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 117
 Two-Sample Hypothesis Testing for Means: Two-
Tailed Test
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to perform and interpret three different types of hypothesis tests for
comparing two population means. Specifically, this lesson will cover:
1. Introduction to Two-Sample Hypothesis Tests for Means
1a. Types of Two-Sample Hypothesis Tests for Means
1b. Steps in a Two-Sample Hypothesis Test for Means
2. Equal Variance Assumption in Two-Sample Hypothesis Tests for Means
3. Two-Sample Hypothesis Testing for Means: Two-Tailed
3a. Applications of Two-Sample Hypothesis Tests for Means: Two-Tailed Tests

1. Introduction to Two-Sample Hypothesis Tests

for Means
Transitioning from one-sample to two-sample hypothesis testing involves expanding your focus from analyzing
a single group to comparing two distinct groups. In one-sample hypothesis testing, you assess whether the
mean of a single sample significantly differs from a known or hypothesized population mean, denoted as
This approach is useful when you have a specific benchmark or standard to compare against. However, in many
real-world scenarios, you are interested in comparing the means of two independent groups to determine if
there is a significant difference between them. This is where two-sample hypothesis testing comes into play. By
comparing the means of two samples, you can draw conclusions about the differences between two
populations. An example of this could be comparing the effectiveness of a marketing campaign for two different
sales regions of the country.

1a. Types of Two-Sample Hypothesis Tests for Means

Just as there are three types of tests for one-sample hypothesis tests, there are similar tests for two-sample
hypothesis tests, including two-tailed, right-tailed, and left-tailed tests.

In the table below, we'll look at each of these types of tests and provide an example of each.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 118
 Type of Test Hypothesis Example

Two-Tailed Comparing the average monthly expenses of the marketing and

development departments to see if there is any difference.

Right-Tailed Comparing the average monthly sales of Store A to Store B to

see if Store A has higher sales.

Left-Tailed Comparing the average monthly expenses of marketing strategy

A to see if it is less expensive compared to marketing strategy B.

 TRY IT

For each scenario state whether the test is a two-tailed, right-tailed, or left-tailed two-sample hypothesis test
and state the hypotheses for the test.

1. A company wants to evaluate whether a new training program for customer service
representatives has led to a decrease in the average response time to customer inquiries compared +
to the previous year.

This is a left-tailed test, and the hypotheses are:

(mean response time after the training program is the same as the mean response time
before the training program)

(mean response time after the training program is less than the mean response time before
the training program)

2. A company wants to compare the average number of tasks completed per day by employees in
two different departments, Sales and Customer Support, to see if there is any difference in +
productivity.

This is a two-tailed test, and the hypotheses are:

(mean number of tasks completed per day by employees in Sales and Customer Support
are equal)

(mean number of tasks completed per day by employees in Sales and Customer Support
are different, not equal)

3. A retail company is comparing the average revenue generated by two different marketing
campaigns, Campaign A and Campaign B, to determine if Campaign A generates more revenue +
than Campaign B.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 119
 This is a right-tailed test, and the hypotheses are:

(mean revenue generated by Campaign A is equal to the mean revenue generated by

Campaign B)

(mean revenue generated by Campaign A is greater than the mean revenue generated by
Campaign B)

1b. Steps in a Two-Sample Hypothesis Test for Means

Below are the steps you can use to perform a two-sample hypothesis test for means.

1. State the Hypotheses: Clearly define the null and alternative hypotheses.
2. Gather the Data: Gather data in a way that is designed to test the hypotheses.
3. Choose the Significance Level: Decide on the α level (for example, 0.05 or 0.10).
4. Check for Equal Variances: Before performing the t-test, check if the variances of the two groups are equal.
Use an F-test to check if the variances of the two groups are equal.
5. Perform a Statistical Test: Use an appropriate statistical test to analyze the data. A two-sample t-test will be
used.
6. Make a Decision: Based on the test results, decide whether to reject or fail to reject the null hypothesis.
This decision is guided by a p-value, which indicates the probability of observing the data if the null
hypothesis is true.
7. Interpret the Results: Explain the results of the hypothesis test in the context of the business problem.

Step 4 is now required for a two-sample hypothesis test for a mean. The next section discusses the importance
and reason for this added step and how to conduct this test in Excel.

2. Equal Variance Assumption in Two-Sample

Hypothesis Tests for Means
When comparing the means of two groups using a t-test, it’s important to check if the variances (a measure of
how spread out the data is) of the two groups are equal. This is known as the equal variance assumption.

A t-test will be used to perform the two-sample hypothesis test for two means. The standard two-sample t-test
assumes that the variances of the two groups are equal. If this assumption is not true, the test results might be
inaccurate. This could lead to wrong conclusions about whether the means of the two groups are different.

If the variances are equal, you use the pooled t-test. This test combines the variances of both groups into one,
making the calculation simpler. If the variances are not equal, you use Welch’s t-test. This test adjusts for the
difference in variances, providing more reliable results when the spreads of the data are different.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 120
 Let’s look at how you would determine if you had equal variances or not among two groups of interest.

EXAMPLE TechWave Innovations is a rapidly growing tech company specializing in developing cutting-
edge software solutions. The company has two main departments: Marketing and Development. Recently,
the CFO (Chief Financial Officer) has noticed fluctuations in the monthly expenses of these departments
and wants to ensure that resources are being allocated efficiently.

The CFO has asked you to compare the average monthly expenses of the Marketing and Development
departments to determine if there is a significant difference between them. The results of this hypothesis
test will help in making informed budget allocation decisions for the upcoming fiscal year.

You realize that you will need to perform a two-sample t-test for means. If the variability in expenses is
similar for both departments, you would use the t-test for equal variances. If one department has much
more variability in expenses than the other, you would use the t-test for unequal variances.

By checking for equal variances, you ensure that you are using the correct method for your data, leading to
more accurate and trustworthy results. This step is crucial for making informed business decisions based on
your analysis.

You test the following in Excel.

(variance in monthly expenses for the Marketing department is equal to the variance in
monthly expenses for the Development department)

(variance in monthly expenses for the Marketing department is not equal to the variance in
monthly expenses for the Development department)

To perform the equality of variances test in Excel using the monthly_expense_comparison.xlsx file, perform
the following:

1. Select F-Test Two Sample for Variances in the XL Miner Analysis Toolpak. The dialog box opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 121
 2. Place your cursor in the Variable 1 Range: box and select the data in column A.

3. Place your cursor in the Variable 2 Range: box and select the data in column B.

4. Place your cursor in the Output Range: box and select cell D3 in the Excel worksheet. Select OK.

You will obtain the following output in your Excel worksheet.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 122
 The value circled in column E11 is the probability that the F-statistic is less than or equal to the observed
value under the null hypothesis for a one-tailed test. That is, it is the p-value for a one-tailed test.

To obtain the two-tailed p-value, you need to multiply the one-tail p-value by 2. So is
the p-value for the hypothesis test for the equality of variances. With this p-value, using a level of
significance (α) of 0.05, you fail to reject the null hypothesis.

This result means there is not enough evidence to conclude that the variances of the two groups are
different. In other words, the variances are considered equal for the purposes of selecting the appropriate t-
test.

When you perform the two-sample t-test for the means, you will select the t-test that assumes equal
variances.

3. Two-Sample Hypothesis Testing for Means:

Two-Tailed
In the upcoming sections, you will explore how to perform a two-tailed hypothesis test for detecting differences
in means. You will discover the purpose of the test, how to perform the test, and how to interpret the results.
Examples and hands-on exercises will illustrate how to set up and interpret a two-tailed test for testing the
difference between two means, helping you understand when and how to use them. By working through these
scenarios, you will gain a deeper insight into hypothesis testing for testing differences among the means from
two populations and their applications in real-world data analysis.

3a. Applications of Two-Sample Hypothesis Tests for Means: Two-Tailed Tests

Let's walk through a practical example of performing a two-sample hypothesis test for means for a two-tailed
test.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 123
 EXAMPLE Continuing with the TechWave Innovations example from the previous section, conduct the
two-sample t-test, now that you know you can assume equal variances among the Marketing and
Development departments' expenses.

Step 1: State the Hypotheses

H₀: (no difference in the average monthly expenses between the Marketing and Development
departments)
H₁: (there is a difference in the average monthly expenses between the Marketing and
Development departments)

Step 2: Gather the Data

The business data analytics team collects monthly expense data for the past year from both departments,
resulting in 50 observations for each department.

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Check for Equal Variances

You have already performed the equality of variances hypothesis test in the last section and concluded that
the variances for the two groups of expenses among the Marketing and Development departments are
considered to be equal.

Step 5: Perform a Statistical Test

You will now use the XL Miner Analysis Toolpak to find the p-value for this two-sample hypothesis test (two-
tailed) using the data in the monthly_expense_comparison.xlsx Excel file.

1. Select t-Test: Two-Sample Assuming Equal Variance in the XL Miner Analysis Toolpak. The dialog box
opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 124
 2. Place your cursor in the Variable 1 Range: box and select the data in column A.

3. Place your cursor in the Variable 2 Range: box and select the data in column B.

4. Enter 0 in the Hypothesized Mean Difference: box.

5. Place your cursor in the Output Range: box and select cell D16 in the Excel worksheet. Select OK.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 125
 You will obtain the following output in your Excel worksheet.

In the output, find the row labeled P(T<=t) two-tail. This row contains the two-tailed p-value. In this case, the
p-value is 2.21968E-06. This number is in scientific notation. 2.21968 is the coefficient and E-06 means the
coefficient needs to be multiplied by To convert 2.21968E-06 to standard form, multiply
which is a very small p-value!

Step 6: Make a Decision

Since the p-value (level of significance), you reject the null hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 126
 Step 7: Interpret the Results

Since the p-value is less than the significance level (0.05), you reject the null hypothesis. There is a
statistically significant difference between the average monthly expenses of the Marketing and
Development departments at TechWave Innovations.

The significant difference in expenses suggests that the two departments manage their budgets differently.
This information can be crucial for the CFO and management team when making decisions about budget
allocations, cost control measures, and financial planning.

The company might investigate further to understand why the expenses differ and whether any
adjustments are needed to optimize spending and resource allocation.

 TRY IT

For the scenario provided, construct a two-sample t-test to address the business problem. Use a level of
significance of 0.05. Ensure your solution contains:

1. A statement of the hypotheses.

2. A test for equality of variances.
3. An interpretation of what the result of the hypothesis test means in the context of the problem.

A leading consumer goods company, GreenWave Inc., has recently launched two new eco-friendly
detergent products: EcoClean Detergent and FreshWave Detergent. Both products are marketed as
environmentally friendly and effective cleaning solutions, but they target slightly different customer
segments. EcoClean is promoted as a budget-friendly option, while FreshWave is positioned as a premium
product with additional fragrance options.

The marketing team at GreenWave Inc. wants to understand how potential customers engage with the
product pages for EcoClean and FreshWave on their website. Specifically, they are interested in comparing
the average time spent on each product’s webpage to determine if there are any significant differences in
user engagement.

You have been tasked with performing a two-sample t-test to determine if there is a significant difference in
the average time spent on each product’s webpage.

Solution:

Before you conduct the hypothesis test, you need to determine whether the variability in the time spent on
the website for each product is similar or not. Conduct an equality of variances test in Excel and interpret
the results. Use a level of significance of 0.05. The Excel file detergent_time_website.xlsx contains the
data.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 127
 The hypotheses are:

H₀: (mean amount of time spent on EcoClean product page is the same as mean amount of
time spend on FreshWave product page for all visitors)
H₁: (mean amount of time spent on EcoClean product page is not the same as mean amount of
time spent on FreshWave product page for all visitors)

You need to perform the equality of variances test to determine which t-test you will select for the two-
sample hypothesis test.

The hypotheses are:

is population variance for EcoClean and measures how much the time spent on the EcoClean product
page varies among all visitors.

is the population variance for FreshWave and measures how much the time spent on the FreshWave
product page varies among all visitors.

Use the F-Test Two-Sample for Variances option in the XL Miner Analysis Toolpak. Your options should look
like:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 128
 From Excel, you obtain the following output. In cell E13, you can enter =2*E10 to obtain the 2-sided p-value
for this test. With this p-value, using a level of significance (α) of 0.05, you fail to reject the null hypothesis.

This result means there is not enough evidence to conclude that the variances of the time spent on the
EcoClean and FreshWave product pages are different. In other words, the variances are considered equal
for the purposes of selecting the appropriate two-sample t-test.

For the two-sample hypothesis test, you will select the t-Test: Two-Sample Assuming Equal Variances option
in XL Miner.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 129
 You obtain the following output from Excel. In the output, find the row labeled P(T<=t) two-tail. This row
contains the two-tailed p-value. The p-value for this two-sample hypothesis test is 0.0157.

Since the p-value (0.0157) is less than the significance level (0.05), you reject the null hypothesis. This result
means there is enough evidence to conclude that there is a statistically significant difference in the average
time spent on the EcoClean and FreshWave product pages for all customers.

 WATCH

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 130
 Follow along with this video on conducting a two-sample t-test to analyze the average time spent on
webpages.

 SUMMARY

In this lesson, you shifted from thinking about one-sample to two-sample hypothesis testing for means.
The tutorial highlighted the importance of comparing two population means to determine whether
significant differences existed between them. An outline provided the process of conducting a two-
sample hypothesis test, including stating hypotheses, gathering data, choosing a significance level,
checking for equal variances, performing the statistical test, making a decision, and interpreting the
results of the test. Understanding the equal variance assumption is crucial for accurate results. A
discussion was provided on how to check for equal variances and the appropriate two-tailed
hypothesis test to select based on the variance equality. A practical example was provided for a two-
tailed test for monthly expenses of the Marketing and Development departments for a company. The
example covered all the steps required for a two-sample two-tailed hypothesis test and demonstrated
how the company could use the results to make informed decisions about budget allocations and
identify areas for resource planning.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 131
 Two-Sample Hypothesis Testing for Means: Right-
Tailed Test
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to perform and interpret a right-tailed test for a two-sample hypothesis
test. Specifically, this lesson will cover:
1. Introduction to Two-Sample Hypothesis Tests for Means: Right-Tailed Tests
1a. Applications of Two-Sample Hypothesis Tests for Means: Right-Tailed Tests

1. Introduction to Two-Sample Hypothesis Tests

for Means: Right-Tailed Tests
Two-sample hypothesis testing for means in right-tailed tests is a statistical method used to determine if the
mean of one group is significantly greater than the mean of another group. In this type of test, the null
hypothesis states that the means of the two groups are equal, while the alternative hypothesis suggests that the
mean of the first group is greater than the mean of the second group. This approach is particularly useful when
you want to test for improvements or increases in a specific metric.

For example, suppose you want to analyze the effectiveness of a new marketing strategy. You could compare
the average sales figures from two different periods: one before the implementation of the new strategy and
one after. The right-tailed test would help determine if the new strategy has significantly increased sales. The
null hypothesis would be that the average sales before and after the strategy are the same, and the alternative
hypothesis being that the average sales after the strategy are higher.

1a. Applications of Two-Sample Hypothesis Tests for Means: Right-Tailed Tests

Let us walk through a practical example of performing a two-sample hypothesis test for means for a right-tailed
test.

EXAMPLE BrightFuture Electronics, a company specializing in consumer electronics, recently launched

a new marketing strategy aimed at boosting sales. To evaluate the effectiveness of this strategy, the
marketing team decided to compare the average monthly sales figures from two different periods: three
months before the strategy was implemented and three months after.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 132
 Before implementing the new marketing strategy, the company recorded the monthly sales data for a
sample of 50 stores. After the strategy was launched and completed, they collected sales data for the same
50 stores. The marketing team hypothesizes that the new marketing strategy has led to an increase in
sales. To test this hypothesis, they decide to use a two-sample right-tailed t-test.

You will perform the two-sample right-tailed t-test to determine if the mean sales after the implementation of
the new marketing strategy are significantly higher than the mean sales before the strategy.

Step 1: State the Hypotheses

H₀: (mean sales after the implementation of the new marketing strategy are equal to the
mean sales before the implementation)
H₁: (mean sales after the implementation of the new marketing strategy are greater than
the mean sales before the implementation)

Step 2: Gather the Data

Sales data was randomly sampled from 50 stores for three months before and three months after
implementing the new marketing strategy. The data is in the Excel file named
sales_before_after_marketing_strategy.xlsx.

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Check for Equal Variances

You test the following in Excel.

H₀: (variance in the sales data before and after the implementation of the new marketing
strategy are equal)
H₁: (variances of the sales data before and after the implementation of the new
marketing strategy are not equal)

To perform the equality of variances test in Excel using the sales_before_after_marketing_strategy.xlsx file,
perform the following.

1. Select F-Test Two Sample for Variances in the XL Miner Analysis Toolpak. The dialog box opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 133
 2. Place your cursor in the Variable 1 Range: box and select the data in column B.

3. Place your cursor in the Variable 2 Range: box and select the data in column C.

4. Place your cursor in the Output Range: box, select cell D2 in the Excel worksheet, and select OK.

From Excel, you obtain the following output. In cell E13, you can enter =2*E10 to obtain the 2-sided p-value
for this test. This p-value (0.0005) is less than the level of significance of 0.05, so you can reject the null

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 134
 hypothesis and conclude that the variances among store sales before and after the implementation of the
new marketing strategy are not equal.

For the two-sample hypothesis test, you will select the t-Test: Two-Sample Assuming Unequal Variances
option in XL Miner.

Step 5: Perform a Statistical Test

Using the data in the sales_before_after_marketing_strategy.xlsx Excel file, you will now use the XL Miner
Analysis Toolpak to find the p-value for this two-sample hypothesis test (right-tailed).

1. Select t-Test: Two-Sample Assuming Unequal Variance in the XL Miner Analysis Toolpak. The dialog box
opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 135
 2. Place your cursor in the Variable 1 Range: box and select the data in column C.

3. Place your cursor in the Variable 2 Range: box and select the data in column B.

4. Enter 0 in the Hypothesized Mean Difference: box.

5. Place your cursor in the Output Range: box, select cell D16 in the Excel worksheet, and select OK.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 136
 Notice in the dialog box above that the column in the Variable 1 Range specifies the data related to the
sales after the marketing strategy was implemented, After_Sales. The options you select for the Variable 1
and Variable 2 boxes in the t-test dialog box are crucial, because they directly determine the direction of
your hypothesis test using the directional math operator in your alternative hypothesis.

The alternative hypothesis is:

H₁:

To set this alternative hypothesis up correctly in the t-test dialog box, perform the following:

Variable 1 Range: Select the range for After_Sales (cells C2:C51).

Variable 2 Range: Select the range for Before_Sales (cells B2:B51).

By correctly setting the Variable 1 and Variable 2 ranges, you ensure that the test is aligned with your
alternative hypothesis and that the results will accurately reflect whether the sales after the marketing
implementation are significantly greater than the sales before (Wright, n.d.).

The key here is to make sure that whatever group is on the left-hand side of the mathematical operator in
the alternative hypothesis is the group that is specified in the Variable 1 Range box.

The Hypothesized Mean Difference: option is usually set to 0. This option directly relates to the null
hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 137
 The null hypothesis is:

H₀:

It means we start by assuming there is no difference between the sales before and after the marketing
implementation. Another way to specify this is to say that the difference between the means of the two
groups is zero. Subtract from both sides of the equation to obtain H₀: which we
test against the alternative hypothesis that the sales after the marketing implementation are greater than
the sales before.

You will obtain the following output in your Excel worksheet. If the test statistic (row labeled t Stat) is
positive, then the row labeled P(T<=T) one-tail is the right-tailed p-value. The p-value for this two-sample
hypothesis test is 0.0259.

Below is a table to help you calculate the p-value for a right-tailed test based on the sign of the test statistic
using the P(T <= t) one-tail output from the XL Miner Analysis ToolPak (Wright, n.d.):

Test Statistic P(T<=t) one tail Right-Tailed p-value Calculation

Positive p p

Negative p 1-p

Positive Test Statistic: If the test statistic is positive, the P(T <= t) one-tail value is already the p-value for
the right-tailed test.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 138
 Negative Test Statistic: If the test statistic is negative, you need to subtract the P(T <= t) one-tail value
from 1 to get the right-tailed p-value.

Step 6: Make a Decision

Since the p-value (level of significance), you reject the null hypothesis.

Step 7: Interpret the Results

Since the p-value (0.0259) is less than the significance level (0.05), you reject the null hypothesis. This
means that there is sufficient evidence to conclude that the sales after the marketing implementation are
significantly greater than the sales before the implementation.

The significant increase in sales after the implementation of the new marketing strategy validates its
effectiveness. This suggests that the marketing strategy has positively impacted consumer behavior and
sales performance. The marketing team at BrightFutures can be confident in their approach and consider it
a successful initiative worth continuing or expanding to all stores.

Demonstrating a successful marketing strategy can strengthen BrightFuture’s position in the market,
potentially attracting more customers and increasing market share. Effective marketing can enhance the
brand image, making it more appealing to consumers and differentiating it from competitors.
Now that you have explored the practical implications of BrightFuture Electronics' marketing strategy using a
two-sample hypothesis testing approach, it is time for you to apply your knowledge by performing a two-sample
hypothesis test using another real-world scenario!

 TRY IT

For the scenario provided construct a two-sample t-test to address the business problem. Use a level of
significance of 0.05. Ensure your solution contains:

1. A statement of the hypotheses.

2. A test for equality of variances.
3. An interpretation of what the result of the hypothesis test means in the context of the problem.

TrendyTech is a popular retail chain specializing in the latest tech gadgets and accessories. The company
has two flagship stores: Store A, located in a bustling downtown area, and Store B, situated in a trendy
suburban shopping center. Both stores have been performing well, but the management team wants to
determine if Store A’s location gives it a significant sales advantage over Store B.

The management team has tasked you, the business data analyst, to compare the average monthly sales of
Store A and Store B to see if Store A has significantly higher sales. This analysis will help in making strategic
decisions about future store locations and marketing investments.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 139
 The business data analytics team collects monthly sales data for the past year from both stores, resulting in
50 observations for each store. This data is in the Excel file named store_sales_comparison.xlsx.

Solution:

The hypotheses are:

H₀: (mean monthly sales of Store A are equal to the mean monthly sales of Store B)
H₁: (mean monthly sales of Store A are greater than or equal to the mean monthly sales of Store
B)

You need to perform the equality of variances test to determine which t-test you will select for the two-
sample hypothesis test.

The hypotheses are:

(variance in monthly sales of Store A is equal to the variance in monthly sales of Store B)

(variance in monthly sales of Store A is not equal to the variance in monthly sales of Store
B)

Use the F-Test Two-Sample for Variances option in the XL Miner Analysis Toolpak. Your options should look
like:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 140
 From Excel, you obtain the following output. In cell E13, you can enter =2*E10 to obtain the 2-sided p-value
for this test. This p-value is less than 0.0001 so you
can reject the null hypothesis and conclude that the variances among the monthly stores of both stores are
not equal.

For the two-sample hypothesis test, you will select the t-Test: Two-Sample Assuming Unequal Variances
option in XL Miner.

You obtain the following output from Excel. If the test statistic (row labeled t Stat) is positive, then the row
labeled P(T<=T) one-tail is the right-tailed p-value. The p-value for this two-sample hypothesis test is 0.1413.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 141
 Since the p-value (0.1413) is greater than the significance level (0.05), you fail to reject the null hypothesis.
This means there is not enough evidence to conclude that the mean monthly sales of Store A are greater
than the mean monthly sales of Store B.

 WATCH

Check out this video on conducting a two-sample t-test to analyze average sales.

 THINK ABOUT IT

In the previous Try It exercise, you failed to reject the null hypothesis and concluded that there was not
enough evidence to conclude that the mean monthly sales of Store A are greater than the mean monthly
sales of Store B.

What are the practical implications for TrendyTech based on the results of this hypothesis test? +

Location Strategy: The lack of significant difference in sales between Store A and Store B suggests
that the downtown location does not provide a substantial sales advantage over the suburban location.
This insight can guide future decisions on store placements, indicating that suburban areas might be
just as viable as downtown locations.

Marketing Investments: Since there is no evidence to support that Store A is performing better than
Store B in terms of sales, TrendyTech might consider distributing marketing resources more evenly
between the two locations. Alternatively, they could focus on other factors that might drive sales, such
as in-store promotions, customer service enhancements, or product variety.

 SUMMARY

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 142
 In this lesson, you learned how to perform and interpret a two-sample right-tailed hypothesis test to
compare two means. The tutorial covered the fundamental concepts of two-sample hypothesis tests,
focusing on right-tailed tests, and their applications. Through a practical example from an electronics
company, you were guided step-by-step to evaluate the effectiveness of a new marketing strategy by
comparing sales data from two different periods: before and after a marketing implementation.
Specifically, you stated the hypotheses, gathered the data, checked for equal variances, performed the
statistical test using Excel, and interpreted the results of the two-sample right-tailed hypothesis tests for
means. The implications of how the results from the hypothesis could be used at the electronics
company were provided.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

REFERENCES

Wright, D. (n.d.). Tail of the test: Interpreting Excel data analysis t-test output. [Link]/tail-of-
the-test-interpreting-excel-data-analysis-t-test-output/.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 143
 Two-Sample Hypothesis Testing for Means: Left-
Tailed Test
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to perform and interpret a left-tailed test for a two-sample hypothesis
test. Specifically, this lesson will cover:
1. Introduction to Two-Sample Hypothesis Tests for Means: Left-Tailed Tests
1a. Applications of Two-Sample Hypothesis Tests for Means: Left-Tailed Tests

1. Introduction to Two-Sample Hypothesis Tests

for Means: Left-Tailed Tests
Two-sample hypothesis testing for means in left-tailed tests is a statistical method used to determine if the
mean of one group is significantly less than the mean of another group. In this type of test, the null hypothesis
states that the means of the two groups are equal, while the alternative hypothesis suggests that the mean of
the first group is less than the mean of the second group. This approach is particularly useful when you want to
test for reductions or decreases in a specific metric.

For example, suppose you want to analyze the impact of a cost-cutting measure. You could compare the
average expenses from two different periods: one before the implementation of the measure and one after. The
left-tailed test would help determine if the new measure has significantly reduced expenses. The null
hypothesis would be that the average expenses before and after the measure are the same, and the alternative
hypothesis being that the average expenses after the cost-cutting measure are lower.

1a. Applications of Two-Sample Hypothesis Tests for Means: Left-Tailed Tests

Let's walk through a practical example of performing a two-sample hypothesis test for means for a left-tailed
test.

EXAMPLE QuantumLeap IT is a mid-sized technology firm specializing in software development and IT

consulting services. The company has been experiencing steady growth over the past few years, but recent
economic challenges have prompted the management team to implement cost-cutting measures to
maintain profitability and ensure long-term sustainability.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 144
 To evaluate the effectiveness of these cost-cutting measures, the finance department at QuantumLeap IT
has decided to analyze the monthly expenses before and after the implementation of the measures. The
goal is to determine whether the new measures have significantly reduced the company’s expenses.

The finance team collected monthly expense data for a sample of 50 expenses from several departments
within the company. The data includes expenses for six months before the cost-cutting measures were
implemented and six months after.

As the data analyst for QuantumLeap IT, you will perform a two-sample left-tailed hypothesis test to
compare the average expenses before and after the cost-cutting measures.

Step 1: State the Hypotheses

H₀: (mean expenses before and after the cost-cutting measures are the same)
H₁: (mean expenses after the cost-cutting measures are lower than the average
expenses before)

Step 2: Gather the Data

Expense data was randomly sampled for 50 different expenses from several different departments across
the company before and after the cost-cutting measures were put in place. The data is in the Excel file
named expenses_before_after_cost_cutting.xlsx.

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Check for Equal Variances

You test the following in Excel:

H₀: (variances in the expense data before and after the cost-cutting measure are equal)

H₁: (variances in the expense data before and after the cost-cutting measure are not
equal)

To perform the equality of variances test in Excel using the expenses_before_after_cost_cutting.xlsx file,
perform the following:

1. Select F-Test Two Sample for Variances in the XL Miner Analysis Toolpak. The dialog box opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 145
 2. Place your cursor in the Variable 1 Range: box and select the data in column A.

3. Place your cursor in the Variable 2 Range: box and select the data in column B.

4. Place your cursor in the Output Range: box, select cell D2 in the Excel worksheet, and select OK.

From Excel, you obtain the following output. In cell E13, you can enter =2*E10 to obtain the 2-sided p-value
for this test. This p-value (0.2866) is greater than the level of significance of 0.05, so you fail to reject the
null hypothesis.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 146
 This result means there is not enough evidence to conclude that the variances of the expenses before and
after the cost-cutting implementation are different. In other words, the variances are considered equal for
the purposes of selecting the appropriate two-sample t-test.

For the two-sample hypothesis test, you will select the t-Test: Two-Sample Assuming Equal Variances option
in XL Miner.

Step 5: Perform a Statistical Test

You will now use the XL Miner Analysis Toolpak to find the p-value for this two-sample hypothesis test (left-
tailed) using the data in the expenses_before_after_cost_cutting.xlsx Excel file.

1. Select t-Test: Two-Sample Assuming Equal Variance in the XL Miner Analysis Toolpak. The dialog box
opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 147
 2. Place your cursor in the Variable 1 Range: box and select the data in column B.

3. Place your cursor in the Variable 2 Range: box and select the data in column A.

4. Enter 0 in the Hypothesized Mean Difference: box.

5. Place your cursor in the Output Range: box, select cell D16 in the Excel worksheet, and select OK.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 148
 Notice in the dialog box above that the column in the Variable 1 Range specifies the data related to the
sales after the marketing strategy was implemented, After_Expenses. Once again, the options you select
for the Variable 1 and Variable 2 boxes in the t-test dialog box are crucial, because they directly determine
the direction of your hypothesis test using the directional math operator in your alternative hypothesis.

The alternative hypothesis is:

H₁:

To set this alternative hypothesis up correctly in the t-test dialog box:

Variable 1 Range: Select the range for After_Expenses (cells B2:B51).

Variable 2 Range: Select the range for Before_Expenses (cells A2:A51).

By correctly setting the Variable 1 and Variable 2 ranges, you ensure that the test is aligned with your
alternative hypothesis and that the results will accurately reflect whether the expenses after the cost-cutting
measures are implemented are significantly less than the expenses before (Wright, n.d.).

Just like with the right-tailed test in the previous tutorial, the key is to make sure that whatever group is on
the left-hand side of the mathematical operator in the alternative hypothesis is the group that is specified in
the Variable 1 Range box.

The Hypothesized Mean Difference: option is usually set to 0. This option directly relates to the null

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 149
 hypothesis.

The null hypothesis is:

H₀:

It means we start by assuming there is no difference between the expenses before and after the cost-
cutting measures. Another way to specify this is to say that the difference between the means of the two
groups is zero. Subtract from both sides of the equation to obtain H₀: which we
test against the alternative hypothesis that the expenses after the cost-cutting measures are less than the
expenses before.

You will obtain the following output in your Excel worksheet. If the test statistic (row labeled t Stat) is
negative, then the row labeled P(T<=T) one-tail is the left-tailed p-value. The p-value for this two-sample
hypothesis test is 0.1698.

Below is a table to help you calculate the p-value for a left-tailed test based on the sign of the test statistic
using the P(T <= t) one-tail output from the XL Miner Analysis ToolPak (Wright, n.d.):

Test Statistic P(T<=t) one tail Left-Tailed p-value Calculation

Positive p 1-p

Negative p p

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 150
 Positive Test Statistic: If the test statistic is positive, you need to subtract the P(T <= t) one-tail value
from 1 to get the left-tailed p-value.
Negative Test Statistic: If the test statistic is negative, the P(T <= t) one-tail value is already the p-value
for the left-tailed test.

Step 6: Make a Decision

Since the p-value (level of significance), you fail to reject the null hypothesis.

Step 7: Interpret the Results

Since the p-value (0.1698) is greater than the significance level (0.05), you fail to reject the null hypothesis.
This means there is not enough statistical evidence to conclude that the cost-cutting measures have
significantly reduced the company’s expenses.

In practical terms, for QuantumLeap IT, this result suggests that the cost-cutting measures implemented by
the management team have not led to a statistically significant reduction in monthly expenses. The
company may need to reassess these measures or consider additional strategies to achieve the desired
cost savings.

The finance team may need to adjust future budgets and forecasts, considering that the expected savings
from the cost-cutting measures were not realized. This could impact financial planning and resource
allocation.

The results can inform broader strategic decisions. For instance, QuantumLeap IT might consider
diversifying its services or entering new markets to boost revenue instead of focusing solely on cost
reduction.
Now that you have explored the practical implications of the QuantumLeap IT expense cost-cutting example
using a two-sample hypothesis testing approach, it is time for you to apply your knowledge by performing a
two-sample hypothesis test using another real-world scenario!

In the next Try It exercise, you will work through the steps of a left-tailed test using a real-world business data
analytics scenario.

 TRY IT

For the scenario provided, construct a two-sample t-test to address the business problem. Use a level of
significance of 0.05. Ensure your solution contains:

1. A statement of the hypotheses.

2. A test for equality of variances.
3. An interpretation of what the result of the hypothesis test means in the context of the problem.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 151
 Innovatech Solutions, a rapidly growing tech startup specializing in cloud-based software solutions, is facing
challenges in optimizing its marketing strategies. Innovatech Solutions is evaluating the effectiveness of two
different marketing channels—social media and email—to determine which one results in higher Customer
Lifetime Value (CLV). The company wants to test if the CLV from customers acquired through social media is
less than those acquired through email marketing.

As a business data analyst at Innovatech Solutions, your task is to analyze the CLV data for customers
acquired through social media and email channels and perform a two-sample t-test to determine if the
average CLV from customers acquired through social media is less than the average CLV from customers
acquired through email.

The Excel file CLV_marketing.xlsx contains a sample of 50 customers where each row represents the CLV
value of an individual customer acquired through either the social media or email marketing channel. The
data captures the Customer Lifetime Value (CLV) for each customer, rather than monthly data.

Solution:

The hypotheses are:

H₀: (average CLV for customers acquired through social media is equal to CLV for customers
acquired through email marketing)
H₁: (average CLV for customers acquired through social media is less than the CLV for
customers acquired through email marketing)

You need to perform the equality of variances test to determine which t-test you will select for the two-
sample hypothesis test.

The hypotheses are:

(variance in CLV for customers acquired through social media is equal to variance in CLV
for customers acquired through email marketing)

(variance in CLV for customers acquired through social media is not equal to variance in
CLV for customers acquired through email marketing)

Use the F-Test Two-Sample for Variances option in the XL Miner Analysis Toolpak. Your options should look
like:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 152
 From Excel, you obtain the following output. In cell E13, you can enter =2*E10 to obtain the 2-sided p-value
for this test. This p-value is less than 0.6475, so you fail to reject the null hypothesis and conclude that the
variances among the CLV values for both marketing channels are equal when selecting the appropriate
two-sample t-test.

For the two-sample hypothesis test, you will select the t-Test: Two-Sample Assuming Equal Variances option
in XL Miner.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 153
 You obtain the following output from Excel. If the test statistic (row labeled t Stat) is negative, then the row
labeled P(T<=T) one-tail is the left-tailed p-value. The p-value for this two-sample hypothesis test is less than
0.0001.

Below is a table to help you calculate the p-value for a left-tailed test based on the sign of the test statistic

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 154
 using the P(T <= t) one-tail output from the XL Miner Analysis ToolPak:

Test Statistic P(T<=t) one tail Left-Tailed p-value Calculation

Positive p 1-p

Negative p p

Positive Test Statistic: If the test statistic is positive, you need to subtract the P(T <= t) one-tail value
from 1 to get the left-tailed p-value.
Negative Test Statistic: If the test statistic is negative, the P(T <= t) one-tail value is already the p-value
for the left-tailed test.

Since the p-value (< 0.0001) is less than the significance level (0.05), you reject the null hypothesis. This
means there is evidence to conclude that the mean CLV for customers acquired from social media channels
is less than the mean CLV for customers acquired from email marketing campaigns.

 WATCH

Check out this video on how to construct a two-sample t-test to evaluate the impact of two marketing
channels on Customer Lifetime Value.

 SUMMARY

In this lesson, you learned how to perform and interpret a two-sample left-tailed hypothesis test to
compare two means. The tutorial covered the fundamental concepts of two-sample hypothesis tests,
focusing on left-tailed tests, and their applications. Through a practical example from an IT consulting
company, you were guided step-by-step to evaluate the effectiveness of cost-cutting measures by
comparing expense data from two different periods: before and after cost-cutting methods were put
into place. Specifically, you stated the hypotheses, gathered the data, checked for equal variances,
performed the statistical test using Excel, and interpreted the results of the two-sample left-tailed
hypothesis tests for means. The implications of how the results from the hypothesis could be used at
the IT consulting company were provided.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

REFERENCES

Wright, D. (n.d.). Tail of the test: Interpreting Excel data analysis t-test output. Retrieved November 12, 2024, from
[Link]/tail-of-the-test-interpreting-excel-data-analysis-t-test-output/.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 155
 ANOVA
by Sophia

 WHAT'S COVERED

In this lesson, you will learn how to test if differences exist among three or more population means by
applying an ANOVA analysis. Specifically, this lesson will cover:
1. Introduction to ANOVA
1a. Steps in ANOVA
1b. ANOVA Example

 BEFORE YOU START

This lesson builds on key concepts from an Introduction to Statistics course. Specifically, this tutorial
assumes familiarity with the foundational idea of ANOVA.

1. Introduction to ANOVA
Since you have just learned how to compare two means for two populations using the two-sample t-test, you
are in a great spot to extend your knowledge to ANOVA (Analysis of Variance). ANOVA allows you to compare
the means of three or more groups simultaneously, helping you determine if there are any statistically
significant differences among them.

Imagine you work for a company called SalesPro Inc., and you want to find out if different sales teams have
different average monthly sales. The company has three sales teams:

Team Alpha
Team Beta
Team Gamma

You want to see if there is a significant difference in the average monthly sales among these three teams.
ANOVA will tell you if the differences in average sales are statistically significant. This means it will show if at
least one team’s sales are different from the others in a meaningful way.

If ANOVA shows a significant difference, you can conclude that at least one sales team is performing differently.
This insight helps SalesPro Inc. understand which teams might need more training or resources and which
strategies are working best.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 156
 Now, since you just learned about two-sample t-tests, you might be thinking about comparing the sales of each
team on a pairwise basis. For example, you could compare:

Team Alpha vs. Team Beta

Team Alpha vs. Team Gamma
Team Beta vs. Team Gamma

While this approach can work, it has some limitations. Performing multiple t-tests increases the risk of Type I
errors (false positives), where you might incorrectly conclude that there is a significant difference when there is
not one. This is where ANOVA comes in handy. ANOVA allows you to compare all three teams simultaneously,
reducing the risk of errors and providing a more comprehensive analysis. In the next section, you will learn how
ANOVA works.

 TERM TO KNOW

Type I Error
An error that occurs when you incorrectly reject the null hypothesis, concluding that there is a
significant difference between groups when, in reality, there is not one.

1a. Steps in ANOVA

Conducting an ANOVA analysis involves several key steps. Below is a guide to help you understand the
process.

 STEP BY STEP

1. State the Hypotheses: Clearly define the null and alternative hypotheses.

The hypotheses in ANOVA are:

The null hypothesis states all group means are equal. In other words, there is no significant difference
among the means of any of the k groups.

The alternative hypothesis states that at least one group mean is different from the others. This means that
there is a significant difference among the means of the groups.

2. Gather the Data: Gather data in a way that is designed to test the hypotheses.

3. Choose the Significance Level: Decide on the α level (for example, 0.05 or 0.10).

4. Perform a Statistical Test: You will use an F-test statistic to perform an ANOVA.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 157
 The F-test statistic is a value that will be computed when the ANOVA test is run. Imagine you have four different
brands of cereal, and you want to know if people like one brand more than the others. You ask a bunch of
people to rate each brand. The F-statistic helps you figure out if the differences in ratings are just by chance or
if they are really different.

Below is how the F-test statistic works in the ANOVA test in the context of the cereal example.

Calculate the Variance: First, you look at how much the ratings vary within each brand (this is called “within-
group variance”) and how much the ratings vary between the different brands (this is called “between-group
variance”).

Compare the Variances: The F-statistic is the ratio of the between-group variance to the within-group variance.
If the F-statistic is large, it means the differences between the groups are bigger than the differences within
each group.

Large F-statistic: This suggests that the differences between the groups are significant, providing more
support to reject the null hypothesis (which states that there are no differences between the groups).
Small F-statistic: This indicates that the differences between the groups are not significant, leading you to
fail to reject the null hypothesis.

How ‘large’ does the F-statistic have to be for you to reject the null hypothesis? This is what the p-value tells
you.
Check the Significance with the p-value:

Small p-value (usually ≤ 0.05): Indicates that the F-statistic is large enough to suggest that the differences
among the groups are unlikely to be due to random chance. This means the differences are statistically
significant.
Large p-value (usually > 0.05): Indicates that the F-statistic is not large enough to rule out random chance as
the reason for the differences among the groups. This means the differences are not statistically significant.

Make a Decision:

Reject the null hypothesis: If the p-value ≤ α (level of significance), reject the null hypothesis.
Fail to reject the null hypothesis: If the p-value > α (level of significance), fail to reject the null hypothesis.

Interpret the Results: Explain the results of the hypothesis test in the context of the business problem.
For example, suppose the ANOVA test for the mean ratings among the four cereals found the F-statistic was
large and the corresponding p-value was small (less than 0.05), you would reject the null hypothesis. This result
means that there are significant differences in the ratings between the cereal brands.

The business conclusion would be that certain cereals are preferred over others. Some actionable insights that
could be taken include:

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 158
 Focus marketing efforts on the preferred brands to capitalize on their popularity.
Investigate what makes the preferred brands more appealing and apply those insights to improve or
innovate other products.
Adjust production and stocking levels to meet the demand for the more popular brand.

1b. ANOVA Example

Let’s look at how you would apply an ANOVA in a real-world business scenario.

EXAMPLE A social media company wants to analyze the engagement of three different types of posts:
Memes, Educational, and Promotional. They collected the number of likes each type of post has received
over the past year.

Your job is to determine if there are significant differences in the average number of likes between the
three types of social media posts: Memes, Educational, and Promotional.

In performing this ANOVA test, you will complete the following steps.

Step 1: State the Hypotheses

H₀: (no difference in the average number of likes between the three
types of posts)
H₁: at least one is different (at least one type of post has a different average number of likes
compared to the others)

Step 2: Gather the Data

The social_media_engagement.xlsx Excel file contains engagement data for the three different types of
social media posts: Memes, Educational, and Promotional. Each type of post has a sample of 50 rows, and
each row is the number of likes for the corresponding type of post.

Step 3: Choose a Significance Level

Use a level of significance,

Step 4: Perform a Statistical Test

You will complete this step by running an ANOVA test in Excel. Perform the following:

1. Select ANOVA: Single Factor. The dialog box opens.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 159
 2. Place your cursor in the Input Range: box and select the data in columns A-C.

3. Place your cursor in the Output Range: box, select cell E2 in the Excel worksheet, and select OK.

You will obtain the following output in Excel. The output contains two parts: SUMMARY and ANOVA. The
SUMMARY portion provides a snapshot of the basic descriptive statistics for each group. The ANOVA
portion contains the results of the ANOVA test, which helps you determine if there are significant
differences between the groups.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 160
 Let's focus on the ANOVA portion.

Below is a short description of each part of the ANOVA table.

1. Source of Variation:

Between Groups: This shows the variation due to differences between the group means (that is,
differences in likes between Memes, Educational, and Promotional posts).
Within Groups: This shows the variation within each group (that is, how much likes vary within Memes,
within Educational posts, and within Promotional posts).

2. Sum of Squares (SS):

Between Groups SS: Measures the variability due to the differences between the group means. In this
context, it represents how much the average number of likes differs between the three types of posts
(Memes, Educational, and Promotional).
Context: Imagine you have the average number of likes for each type of post. The Between Groups
SS tells us how much these averages differ from each other. For example, if Memes have an average
of 150 likes, Educational posts have an average of 120 likes, and Promotional posts have an average
of 170 likes, the Between Groups SS measures the variability among these averages.
Within Groups SS: Measures the variability within each group. In this context, it represents how much
the number of likes varies within each type of post.
Context: Within each type of post (Memes, Educational, and Promotional), individual posts will have
different numbers of likes. The Within Groups SS measures how much these individual likes differ
from the average likes for that type of post. For example, if the average number of likes for Memes is

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 161
 150, but individual Memes have likes ranging from 100 to 200, the Within Groups SS captures this
variability.

3. Degrees of Freedom (df):

Between Groups df: Number of groups minus one (for example, ).

Within Groups df: Total number of observations minus the number of groups (for example,
).

4. Mean Square (MS):

Between Groups MS: Calculated by dividing the Between Groups SS by the Between Groups df.
Within Groups MS: Calculated by dividing the Within Groups SS by the Within Groups df.

5. F-Statistic:

The F-statistic is the ratio of the Between Groups MS to the Within Groups MS. It tells us if the variability
between the group means is larger than the variability within the groups.

6. P-value:

The p-value helps determine if the F-statistic is significant. A small p-value (usually less than 0.05)
means that the differences between the groups are significant.

Step 5: Make a Decision

Since the p-value (level of significance), you can reject the null hypothesis.

Step 6: Interpret the Results

Since the p-value (0.0175) is less than the significance level (0.05), you reject the null hypothesis. This
means there is sufficient evidence to conclude that there are significant differences in the average number
of likes between the three types of social media posts (Memes, Educational, and Promotional).

 DID YOU KNOW

In the previous example, the ANOVA test told you that at least two of the post types have different average
numbers of likes. However, an ANOVA test does not specify which pairs of post types are different. To
determine exactly which post types differ from each other, you would need to perform additional tests
known as post-hoc tests.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 162
 Now, it is your turn to put your ANOVA skills to use. Put your knowledge into practice with a hands-on problem
to solidify your understanding and see how these concepts apply to real-world data.

 TRY IT

For the scenario provided, perform an ANOVA test to address the business problem. Use a level of
significance of 0.05. Ensure your solution contains:

1. A statement of the hypotheses.

2. An interpretation of what the result of the ANOVA test means in the context of the problem.

A streaming service wants to analyze the average watch time for three different genres of shows: Drama,
Comedy, and Documentary. They collected watch time data (in minutes) for each genre over a month. The
goal is to determine if there are significant differences in the average watch time between these genres.

The streaming_service_watch_time.xlsx Excel file contains watch times (minutes) for three different genres.
Each type of genre has a sample of 50 rows, and each row is the watch time for the corresponding genre.

Solution:

The hypotheses are:

H₀: (no differences in the average watch time between the three genres
of shows (Drama, Comedy, Documentary))
H₁: at least one is different (at least one genre has a different average watch time compared to the
others)

From Excel, you obtain the following output. You observe that the F-test statistic is quite large. A large F-
statistic indicates that the variability in watch times between the different genres is much greater than the
variability within each genre.

This large F-statistic suggests that the average watch times for at least two of the genres are significantly
different from each other, and this is supported by the p-value. You notice that the p-value is being reported
as 0. When the p-value is reported as 0 in the ANOVA table, it means that the p-value is extremely small,
often less than 0.0001. In practical terms, this indicates that the observed differences between the group
means are highly significant and not due to random chance.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 163
 Since the p-value (<0.0001) is less than the significance level (0.05), you reject the null hypothesis. This
means there is sufficient evidence to conclude that there are significant differences in the average watch
times between the three genres (Drama, Comedy, and Documentaries).

 WATCH

Check out this video on performing an ANOVA test to address the business problem.

 SUMMARY

In this lesson, you learned how to test differences among three or more population means using an
ANOVA analysis. An introduction to ANOVA was provided that explained the purpose and application,
such as comparing the sales team’s performance. The tutorial outlined the steps in ANOVA, including
stating hypotheses, gathering data, choosing a significance level, performing the statistical test, and
interpreting results. An ANOVA example demonstrated the process using a real-world scenario where
a social media company analyzed engagement for three types of posts: Memes, Educational, and
Promotional. This example helped illustrate how ANOVA can be used in a real-world business scenario
to make data-driven decisions about content strategy and engagement on social media platforms.

Source: THIS TUTORIAL WAS AUTHORED BY SOPHIA LEARNING. PLEASE SEE OUR TERMS OF USE.

 TERMS TO KNOW

Type I Error

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 164
 An error that occurs when you incorrectly reject the null hypothesis, concluding that there is a
significant difference between groups when, in reality, there is not one.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 165
 Terms to Know
Alternative Hypothesis
A statement in statistical hypothesis testing that proposes there is an effect or a difference in
the population.

Central Limit Theorem (CLT)

A property that states the sampling distribution of approaches a normal distribution as the
sample size becomes large, regardless of the original population’s distribution.

Cluster Sampling
A type of sampling method where entire clusters are selected and all elements within the
cluster are sampled.

Confidence Intervals
A range of values used to estimate an unknown parameter of a population.

Critical Value
A factor used to calculate the upper and lower bounds of a confidence interval.

Degrees of Freedom
Value that determines the shape of a t-distribution and is calculated as the sample size minus
one.

Hypothesis Testing
A statistical method used to make decisions about a population based on sample data.

Law of Large Numbers

A property that states as the size of a sample increases, the sample mean will become closer
to the population mean and the sample proportion will become closer to the population
proportion.

Left-Tailed Test
Type of hypothesis test used when the research question is to determine if the population
parameter is less than the null hypothesis value.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 166
 Margin of Error
A numerical value that expresses the range within which the true value of the population
parameter is expected to fall.

Null Hypothesis
A statement used in statistical hypothesis testing that proposes there is no effect or no
difference in a population.

P-value
A measure used in statistical hypothesis testing to help determine the significance of your
results.

Point Estimators
Statistics used to provide a single best guess or estimate of an unknown population
parameter based on sample data.

Right-Tailed Test
Type of hypothesis test used when the research question is to determine if the population
parameter is greater than the null hypothesis value.

Sampling Distribution
Distribution of a given statistic based on a random sample, showing how the statistic varies
from sample to sample.

Sampling Distribution of p̂
A distribution of all possible sample proportions of a given sample size from a population,
showing how the sample proportion varies from sample to sample.

Sampling Distribution of x̄
A distribution of all possible sample means of a given size from a population, showing how
the sample mean varies from sample to sample.

Sampling Error
The difference between the results obtained from a sample and the actual values in the
population from which the sample was drawn.

Standard Error

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 167
 Measure of the variability of a sample statistic (such as ) from the population, calculated as
the standard deviation of the sampling distribution of that statistic.

Strata
Subgroups or segments of a population that share similar characteristics.

Stratified Sampling
A type of sampling method that involves dividing a population into smaller groups, known as
strata, that share similar characteristics.

Systematic Sample
A type of sampling method where every kth element is selected.

Test Statistic
A standardized value calculated from sample data for a hypothesis test that measures how
much your sample data deviates from the null hypothesis.

Two-Tailed Test
Type of hypothesis test used when the research question is to determine if the population
parameter is different from the null hypothesis value.

Type I Error
An error that occurs when you incorrectly reject the null hypothesis, concluding that there is
a significant difference between groups when, in reality, there is not one.

t-distribution
A type of probability distribution that is symmetric, bell-shaped, and used when estimating
population parameters, particularly when the sample size is small and the population
standard deviation is unknown.

© 2025 SOPHIA Learning, LLC. SOPHIA is a registered trademark of SOPHIA Learning, LLC. Page 168