Tema 1 a

Market Research: An Introduction

Definition

Market research involves systematically designing, collecting, analyzing,

and presenting data to support marketing decision-making. It is crucial for
a company's strategic planning and helps identify market opportunities,
problems, and consumer behaviors.

Purpose of Market Research

1. Understand Problems:

o Example: Analyzing a drop in product sales or assessing the

best discount strategy.

2. Plan for the Future:

o Example: Exploring the potential for new products or

diversification into other product categories.

3. Support Strategic Decision-Making:

o Packaging decisions, quality control, and satisfaction surveys

all rely on market research insights.

Objectives of Market Research

Market research aims to achieve the following six goals:

1. Qualitative Analysis: Understand consumer motivations (e.g.,

importance of brand image or price).

2. Quantitative Analysis: Quantify consumer preferences (e.g.,

surveys to measure the effect of price and nutritional content).

3. Product Analysis: Design products that align with consumer

preferences.

4. Advertising Analysis: Identify target audiences and craft

appropriate messaging.

5. Distribution & Sales Analysis: Evaluate sales channels and

outlets (e.g., organic products in short channels).

6. Competitive Analysis: Study competitors to adapt strategies or

identify new opportunities.
Phases of Market Research

1. Research Design:

o Define the problem and type of research needed.

o Formulate hypotheses and variables.

2. Information Gathering:

o Use primary (new) and secondary (existing) data sources.

3. Analysis:

o Validate collected data.

o Apply statistical techniques for interpretation.

4. Result Presentation:

o Summarize findings and provide actionable conclusions.

Types of Market Research

1. Exploratory:

o Collects basic information using secondary data and

qualitative methods like focus groups.

o Example: Understanding consumer preferences.

2. Descriptive:

o Describes current situations, such as packaging preferences

among different consumer groups.

3. Explanatory/Causal:

o Measures cause-and-effect relationships through experiments.

o Example: Studying how price reductions impact sales.

4. Predictive:

o Projects future outcomes based on explanatory research.

Scales of Measurement

In market research, data is measured using four scales, progressing from

basic identification to complex analysis:
 1. Nominal Scale:

o Numbers are used for identification, with no inherent order

(e.g., gender: male = 1, female = 2).

2. Ordinal Scale:

o Data is ranked in order but without equal intervals (e.g.,

income levels: high, medium, low).

3. Interval Scale:

o Measures have equal intervals, but no true zero point (e.g.,

temperature in Celsius).

4. Ratio Scale:

o Includes a true zero, allowing proportional comparisons (e.g.,

sales figures, weight).

Variables, Indicators, and Indices

 Variable: A measurable element that varies, like age or income.

 Indicator: Used to measure abstract concepts (e.g., fuel poverty as

a ratio of energy expenses to income).

 Index: Aggregates multiple indicators for a comprehensive

assessment (e.g., sustainability index).

Special Scales for Measuring Attitudes

Attitude scales are used to assess preferences, opinions, or perceptions.

Examples include:

1. Comparative Scales:

o Paired comparisons: Compare two objects based on a criterion.

o Constant sum: Allocate points to different options to reflect

importance.

2. Non-Comparative Scales:

o Likert Scale: Rates agreement with statements on a scale

(e.g., 1 = strongly disagree to 5 = strongly agree).

o Semantic Differential: Rates an object on bipolar adjectives

(e.g., "cheap–expensive").
 o Stapel Scale: Rates intensity using a single descriptor (e.g., +3
to -3).

Measurement Accuracy

 Systematic Error: Consistent bias caused by faulty instruments or

procedures.

 Random Error: Unpredictable variations due to random factors like

mood or sample differences.

 Characteristics of a Good Measurement:

o Validity: Measures what it’s supposed to measure.

o Reliability: Produces consistent results.

o Sensitivity: Detects significant differences between data

points.

Applications in Business

1. Exploration:

o Assess market size and preferences for new products.

2. Implementation:

o Select appropriate sales channels (e.g., supermarkets vs.

specialized stores).

3. Consolidation:

o Monitor customer satisfaction and evolving preferences over a

product's lifecycle.
Tema 1b

Market Research: Sources of Information

Basic Concepts

 Data collection is a critical process in market research. The goal is to

collect just enough data to meet the research objectives without
unnecessary costs or complexity.

 Types of Information Sources:

1. Secondary Data (Pre-existing):

 Internal: Data from within the company (e.g., sales

records, customer databases).

 External: Data from outside sources (e.g., government

statistics, publications).

2. Primary Data (Newly collected for a specific purpose):

 Methods include qualitative research, quantitative

research, observation, experiments, and surveys.

Advantages and Disadvantages

Secondary Data:

 Advantages:

o Cost-effective and time-saving.

 Disadvantages:

o Not always tailored to the specific research question,

potentially requiring further primary data collection.

Primary Data:

 Advantages:

o Directly relevant and customized to the research question.

 Disadvantages:

o More costly and time-intensive.

Assessment of Data Sources

To ensure data quality and relevance, evaluate these criteria:

1. Research Objective:

o Check if the data aligns with your research goals.

o Analyze how variables were defined and operationalized.

 Example: How "small businesses" or "unemployed

individuals" were categorized.

2. Methodological Quality:

o When and how was the data collected?

o Assess time lag between data collection and availability (e.g.,

census data collected every 10 years may become outdated).

3. Source Institution:

o Consider the institution's methodological rigor and

independence.

4. Accessibility:

o Determine the ease of access and whether the data is free or

requires payment.

o Check for data disaggregation (e.g., by location,

demographics).
Tema 2a

Market Research: Observation Method

Definition

The observation method in market research involves watching and

recording the actions or behaviors of people, often without their
knowledge. It is used to directly record behavioral patterns, with little or no
interaction between the researcher and the observed subjects.

Advantages of Observation

1. No Manipulation: Observers do not interfere with the subject or

variables.

2. Faster and Cost-Effective: Observations can be made in real-time

using scanners or other technologies.

3. Accurate Insights: Captures actual behavior, not just reported or

claimed behavior.

4. Ethical Superiority in Sensitive Areas: Useful in observing

socially disapproved behaviors.

Disadvantages of Observation

1. Ethical and Legal Concerns: Covert observations can raise ethical

issues.

2. Limited Scope: Cannot capture motives or intentions.

3. External Influences: Lack of control over outside variables.

4. Potential Bias: Observer distraction or subjectivity can affect data

accuracy.

5. Reactivity: Subjects may alter behavior if they realize they are

being observed.

Types of Observation

1. Degree of Structure:

o Structured: Focuses on specific objectives (e.g., monitoring

consumer behavior near particular products).
 o Unstructured: Gathers broader information for later analysis
(e.g., tracking movement through a store).

2. Timing:

o Direct Observation: Records events as they happen (e.g.,

watching customer reactions in a shop).

o Indirect Observation: Uses pre-recorded material or system-

generated data (e.g., analyzing loyalty card data).

3. Awareness of Subjects:

o Covert Observation: Subjects are unaware they are being

observed (e.g., hidden cameras).

o Uncovered Observation: Subjects know they are being

monitored (e.g., during focus groups).

4. Location:

o Natural Setting: Observations occur in the subject's usual

environment (e.g., supermarkets).

o Artificial Setting: Observations occur in a controlled

environment (e.g., experimental labs).

5. Recording Mechanism:

o Personal Observation: Data is recorded manually by

observers.

o Mechanical Observation: Technology records behavior (e.g.,

eye-tracking, galvanometers).

Applications of Observation

1. Merchandising Studies:

o Track customer movements and purchases to optimize store

layouts or promotions.

o Example: Use of scanner data to evaluate the impact of a new

discount strategy.

2. Advertising Analysis:

o Analyze content of advertisements (e.g., message, visuals,

audience targeting).

3. Mystery Shopping:
 o Researchers pose as customers to evaluate service quality
(e.g., at restaurants or banks).

4. Eye-Tracking Studies:

o Use technology to monitor what customers focus on during an

interaction or purchase.

Observation Techniques

1. Personal Observation:

o Example: A researcher observes which products in a store

attract attention.

2. Mechanical Observation:

o Examples:

 Eye-Tracking: Tracks gaze and focus of customers.

 Galvanometer: Measures emotional responses.

 Counters: Tracks the number of vehicles or people.

3. Trace Analysis:

o Studies physical or digital traces left behind by consumers

(e.g., website cookies, litter analysis).

Key Advantages of Mechanical Observation

 Detailed and permanent records of behavior.

 Objective data collection with no observer bias.

Example Studies

1. Eye-Tracking Technology:

o Measures pupil behavior and viewing time.

o Useful in website design or store displays.

2. Content Analysis:

o Studies how brands communicate their message and target

demographics.
 o Example: Mercedes using campaigns to attract younger
audiences.

Tema 2b

What is an experiment?

An experiment changes something (called the independent variable) to

see how it affects another thing (called the dependent variable). For
example:

 Independent Variable: Giving a 50% discount.

 Dependent Variable: How many items are sold.

How are experiments set up?

 Groups: Usually, we divide people or places into groups:

o Experimental Group: Gets the change (a discount or new

product).

o Control Group: Doesn’t get the change (this is for

comparison).

 Randomization: To make it fair, people or places are randomly

assigned to these groups. This reduces bias.

Types of Experiments

1. Pre-Experimental:

o Not very rigorous. Example: Measure sales after offering a

discount without comparing to another group.

2. Quasi-Experimental:

o Some control but not fully random. Example: Measure sales

before and after a campaign but don’t randomize which stores
get the campaign.

3. Fully Experimental:

o Best type. Groups are random, and we compare results

between them. Example: Compare sales between two groups
of stores, one with a discount and one without.

Real-Life Applications

Experiments can be used to:

 Test product designs or packaging.

  Evaluate how effective an advertisement is.

 See if discounts increase sales.

 Check if a website or app design affects how people use it.

Tema 2c

What Are Qualitative Techniques?

 These techniques help researchers understand "why" people do

things (like why they choose one brand over another).

 They’re often used before quantitative research to explore ideas

or afterwards to interpret results.

Why Use Qualitative Techniques?

 Advantages:

o They deal with things you can’t measure easily (like emotions
or opinions).

o They help discover hidden motivations and non-rational

behaviors (like impulse buying).

 Disadvantages:

o Results can’t always be generalized to the whole population.

o They require skilled researchers to interpret data accurately.

When to Use Them?

 Exploratory Research: To identify important issues and create

better surveys

 Sensitive Topics: When people might feel uncomfortable answering

directly (like about alcohol or spending habits).

Main Qualitative Techniques

1. Expert Interviews:

o Talk to specialists in a field for their insights.

o Useful when no data is available elsewhere.

2. Delphi Technique:
 o Collect opinions from a group of experts anonymously.

o Repeat this process until they agree on key points.

3. Focus Groups:

o Small groups of people discuss a topic.

o A moderator guides the discussion to collect diverse opinions.

4. In-Depth Interviews:

o One-on-one conversations to explore personal motivations and

attitudes deeply.

5. Projective Techniques:

o Indirect methods to uncover hidden feelings:

 Word Association: Say the first word that comes to

mind.

 Sentence Completion: Fill in the blank (e.g., "Buying a

car is...").

 Picture Tests: Look at a picture and tell a story about it.

Tema 3a

What is a Survey?

A survey is a set of questions given to people or organizations to collect

their responses. It’s used to understand opinions, behaviors, or
preferences.

Types of Surveys

1. Structured Surveys:

o Use clear, fixed questions (mostly closed-ended).

o Example: "Do you like this product? (Yes/No)"

o Best for collecting standardized data.

2. Unstructured Surveys:

o Use open-ended questions that let people explain their

answers.

o Example: "What do you think about this product?"

o Used more in exploratory research.

Key Steps in a Survey

1. Create a Questionnaire:

o Questions should be clear

o Choose between open-ended or closed-ended formats.

o Organize questions logically to ensure flow.

2. Pilot Test:

o Test the survey on a small group first.

o Check if questions are understood and identify potential

issues.

3. Avoid Errors:

o Sampling Errors: Using a sample that isn’t representative of

the population.

o Interviewer Errors: Misinterpreting or influencing responses.

o Tabulation Errors: Mistakes when entering or analyzing data.

Types of Surveys Based on How They’re Done

1. Personal Surveys:

o Conducted face-to-face.

o Best for long or complex questionnaires.

o High response rates but expensive and time-consuming.

2. Telephone Surveys:

o Conducted over the phone.

o Faster and cheaper but limited to shorter questions.

3. Mail Surveys:

o Sent through postal mail.

o Low cost but very low response rates.

4. Online Surveys:

o Respondents fill out a form on a website.

o Cheap and fast, but sample representation can be tricky.

What Are Panels?

 A panel is a group of people or businesses surveyed regularly over

time to track changes in behavior or preferences.

 Types:

1. Retailer Panels: Focus on sales, prices, and market share.

2. Consumer Panels: Focus on consumer behavior (e.g.,

household spending).

Advantages of Panels

 Provide consistent and high-quality data.

 Allow tracking of trends and changes over time.

 Help analyze specific groups (e.g., frequent buyers).

Tema 3b

What is Sampling?

Imagine you want to find out what type of snacks students in your school
like. Asking every single student would take too much time and effort, so
you pick a smaller group to represent everyone. This smaller group is
called a sample, and the process of selecting this group is called
sampling.

Key Concepts in Sampling

1. Population: This is the whole group you are interested in (e.g., all
the students in your school).

2. Sample: A smaller group taken from the population (e.g., 50

students chosen from your school).

3. Sampling Frame: A list of everyone in the population that you can

choose from (e.g., the class attendance sheet).

4. Sampling Unit: The individual parts of the population that you

study (e.g., each student in the school).
Why Use Sampling?

1. Saves Time and Money: It’s quicker and cheaper than talking to
everyone.

2. Gives Good Enough Information: If done well, sampling can give

results that are very close to asking everyone.

3. Necessary in Some Cases: If testing damages the sample (e.g.,

testing a light bulb until it breaks), you can’t test everything.

Types of Sampling Methods

There are two main ways to pick a sample:

1. Non-Probability Sampling (Not Random)

 Convenience Sampling: Picking the easiest people to talk to (e.g.,

students in your class).

 Purposive Sampling: Choosing people you think are important for

your study (e.g., only snack lovers).

 Snowball Sampling: Asking someone in the sample to recommend

others (e.g., asking friends of friends).

 Quota Sampling: Dividing people into groups (e.g., age or gender)

and picking a set number from each.

2. Probability Sampling (Random)

 Simple Random Sampling: Everyone has an equal chance of

being chosen (e.g., drawing names from a hat).

 Systematic Sampling: Picking every nth person (e.g., every 5th

student on the list).

 Stratified Sampling: Dividing the population into groups (e.g., by

grade) and choosing randomly from each group.

 Cluster Sampling: Dividing the population into clusters (e.g.,

classes) and choosing a whole cluster randomly.

How to Do Sampling Step by Step

1. Identify the Population: Decide who you want to study.

2. Create a Sampling Frame: Make a list of all possible participants.

3. Choose a Sampling Method: Pick one of the methods above.

 4. Decide Sample Size: Figure out how many people you need to
study.

5. Collect Data: Ask the sample your questions (e.g., by surveys or

interviews).

Tema 3c

What is a Statistical Distribution?

A statistical distribution shows how often different values happen in a

dataset. Think of it like a chart that shows how many students are in
different height ranges in your school.

Examples:

 Normal Distribution: Looks like a bell curve. It’s used when the
dataset is large (e.g., 30+ students) and shows most values near the
average.

 Student’s t-Distribution: Used for smaller datasets (less than 30).

It’s similar to the normal distribution but with fatter tails.

 Chi-square Distribution: Used for analyzing categories (like "yes"

or "no" answers).

What is Statistical Inference?

Statistical inference is about making guesses about a whole group

(population) based on data from a smaller group (sample).

Methods of Statistical Inference:

1. Point Estimation: A single number that estimates something about

the population.

o Example: The average oil consumption is 10 liters per person

per year.

2. Interval Estimation: A range of numbers where the true value is

likely to be, with a confidence level.

o Example: We are 95% confident that oil consumption is

between 8.9 and 11.1 liters per person per year.

3. Hypothesis Testing: Checking if an assumption about the

population is true or not.
 o Example: Testing if the average oil consumption is more than
9.5 liters.

Why Do These Matter?

 Distributions help us understand the shape of the data and

calculate probabilities.

 Inference helps us make conclusions about a whole population

without checking everyone.

Tema 4b

Key terms explained:

1. Population: This is the entire group you’re studying (e.g., all

students in the school).

2. Sample: A smaller group chosen from the population to represent it.

3. Confidence level: This tells you how sure you are about your
results. For example, a 95% confidence level means you’re 95%
certain your sample reflects the population.

4. Margin of error (precision): This is how much your result might

differ from the real answer for the whole group. A smaller margin of
error means more accurate results.

5. Variability: This refers to how diverse the population is. If opinions

are very different, you need a bigger sample to capture all the
variation.

Why is sample size important?

If your sample is too small, your results might not reflect the whole
population. If it's too large, it might waste resources like time and money.
The goal is to find the right balance.

What affects sample size?

1. Confidence level: A higher confidence level (e.g., 99% vs. 95%)

needs a bigger sample.

2. Margin of error: The smaller the error you want, the larger the
sample.
 3. Population variability: If opinions in the population vary a lot,
you’ll need more people in your sample.

4. Population size: For small populations, the sample size can be

adjusted using a formula.

Example:

If you want to find out how many people like a specific type of music with
95% confidence and a 5% margin of error, you would:

1. Decide your confidence level (e.g., 95% → Z = 1.96).

2. Use the worst-case variability (50% liking, 50% not liking →

maximum diversity).

3. Use the formula to calculate the minimum sample size.

The formula for infinite populations:

n=Z2⋅π⋅(1−π)e2n = \frac{Z^2 \cdot \pi \cdot (1 - \pi)}{e^2}

Where:

 nn: Sample size

 ZZ: Z-value for confidence level (e.g., 1.96 for 95%)

 π\pi: Proportion expected (e.g., 0.5 for maximum variability)

 ee: Margin of error (e.g., 0.05 for 5%)

If you’re dealing with a smaller population, you adjust nn with a "correction

factor."

Summary:

Sample size estimation helps you confidently study a group without

surveying everyone. By setting your confidence level, margin of error, and
accounting for variability, you can calculate the right sample size. It’s
about being smart and precise while saving effort

Tema 5

1. What is a hypothesis?
A hypothesis is like a guess or claim about something. For example:

 "The new medicine reduces fever faster than the old one."

 "Students who study with music score the same as those who don't."

2. Why test a hypothesis?

We test a hypothesis to see if the guess is true or false using data. Think
of it like investigating whether a rumor is true.

3. Key ideas in hypothesis testing

1. Two Types of Hypotheses:

o Null Hypothesis (H₀): This is the "default" assumption. It's

like saying "Nothing unusual is happening." Example: "The
new medicine is no better than the old one."

o Alternative Hypothesis (H₁): This is what you hope to

prove. It challenges the null. Example: "The new medicine
works better than the old one."

2. Testing the Claim:

o You collect data and calculate a statistic (a number based on

the data).

o Compare your statistic to a critical value (a threshold that

decides if the result is surprising enough to reject H₀).

3. Decisions:

o If the data shows something unusual, reject H₀ and accept H₁.

o If not, keep H₀ (you didn’t find enough evidence to support

H₁).

4. Errors in Hypothesis Testing

There are two possible mistakes:

 Type I Error: You think something is true, but it’s not (delulu).
Example: Concluding the new medicine works when it doesn’t.

 Type II Error: You miss something that is actually true.(blind)

Example: Failing to realize the new medicine works better.
5. One-Tailed vs. Two-Tailed Tests

 One-Tailed Test: You check for an effect in one specific direction.

Example: Is the new medicine better than the old one?

 Two-Tailed Test: You check for any effect, either direction. Example:
Is the new medicine better or worse than the old one?

6. Effect Size

 Effect size tells you how big the difference is.

o Even if a small difference is statistically significant, it might not

be practically useful. Example: "The new medicine reduces
fever by 0.01 degrees. Is that helpful?" no , fuck off

7. Real-Life Example

Suppose a government wants to check if a "Green Nudge" (encouraging

people by comparing their electricity use to others) helps people save
energy.

Steps:

1. Null Hypothesis (H₀): The nudge doesn’t affect energy use.

2. Alternative Hypothesis (H₁): The nudge reduces energy use.

3. Collect data: Track energy use of people with and without the
nudge.

4. Analyze: Use statistical tests to compare the two groups.

5. Decide: If the data shows a clear reduction in energy use, reject H₀

and conclude the nudge works.

Tema 6a
1. What is descriptive analysis?

Descriptive analysis is about summarizing and understanding data to

make it easier to interpret. It helps you answer basic questions about your
data, like:

 How are the values distributed?

 What’s the most common value?

 How much do the values vary?

It’s the first step before doing more advanced analysis.

2. Types of Variables

1. Nominal Variables:

o Categories with no order.

o Example: Favorite color (red, blue, green).

2. Ordinal Variables:

o Categories with an order, but the difference between

categories is not meaningful.

o Example: Education level (high school, bachelor's, master's).

3. Metric Variables:

o Numbers with meaningful differences between values.

o Example: Income, temperature.

3. Key Measures in Descriptive Analysis

1. Central Tendency (What is "typical"):

o Mean (average): Add up all values, then divide by the

number of values.

 Example: Average age in a group.

o Median: The middle value when numbers are arranged in

order.

 Example: If ages are 20, 22, 24, 26, 28 → Median is 24.

o Mode: The most frequent value in the data.

 Example: In shoe sizes [7, 8, 8, 9, 10], the mode is 8.

 2. Dispersion (How "spread out" the data is):

o Range: Difference between the highest and lowest value.

 Example: If incomes are $2,000 and $10,000, the range

is $8,000.

o Standard Deviation (σ): Shows how much the values differ

from the average.

 High σ = Data is spread out. Low σ = Data is clustered

near the mean.

4. Analyzing Different Variables

Depending on the variable type, we use different tools:

1. Nominal Variables:

o Use frequencies (counts of each category) and pie/bar

charts.

o Example: Favorite fruit – apples (40%), bananas (30%),

oranges (30%).

2. Ordinal Variables:

o Use frequencies, median, and bar charts.

o Example: Education levels – 50% high school, 30% bachelor's,

20% master's.

3. Metric Variables:

o Use all measures: mean, median, mode, standard deviation.

o Example: Heights – Mean = 170 cm, standard deviation = 5

cm.

6. Why is this important?

Descriptive analysis helps you:

 Summarize data quickly.

 Spot trends or patterns.

 Understand the variability in data before moving to advanced

analysis.
Tema 6b

1. What is bivariate analysis?

relationship between two variables. The goal is to see if one variable

is connected to the other.

For example:

 Question: Does a politician’s approval rating depend on the party

people vote for?

 Variables:

o Politician approval (e.g., yes or no)

o Party voted for (e.g., A, B, or C)

2. Types of Tests

To study relationships, there are two main types of tests:

1. Parametric Tests:

o Used when the data follows a normal distribution and meets

specific rules.

o Examples:

 t-test: Compares averages between two groups.

 Pearson Correlation: Measures how strong the linear

relationship is between two variables.\

 ANOVA (Analysis of Variance): Compares averages

between more than two groups.

2. Non-Parametric Tests:

o Used when data does not follow a normal distribution or

doesn’t meet the rules for parametric tests.

o Examples:

 Mann-Whitney U Test: Compares averages for two

groups without assuming normality.

 Spearman Correlation: Measures the relationship

between two variables without assuming linearity.
  Chi-square Test: Examines the relationship between
two categorical variables (e.g., approval and party).

3. When to Use Which Test?

You choose the test based on:

 The type of variables:

o Nominal variables: Categories (e.g., party names).

o Ordinal variables: Ordered categories (e.g., education

levels).

o Metric variables: Numbers (e.g., income, age).

4. Example of Bivariate Analysis

Suppose you want to see if approval of a politician depends on the

party people vote for:

1. Null Hypothesis (H₀): There is no relationship between approval

and party.

2. Alternative Hypothesis (H₁): There is a relationship between

approval and party.

3. Test Used: Chi-square test (since these are nominal variables).

4. Process:

o Create a frequency table showing how many people approve

or disapprove by party.

o Use the chi-square test to see if the differences are statistically

significant.

5. Why is this important?

Bivariate analysis helps answer real-world questions, such as:

 Does advertising budget (metric) affect sales (metric)?

 Are people’s political preferences (nominal) related to their

education level (ordinal)?
Tema 7a

1. What is a t-test?(czy są różne czy przypadek)

 It’s a statistical tool that helps you decide if two groups are truly
different from each other or if the differences are just random.

 For example:

o Do customers who use a loyalty app spend more money than

those who don’t?

o Do men and women consume different amounts of a product?

2. Types of Comparisons

 One group against a known value: Did the average monthly

spending increase after an advertising campaign?

 Two groups against each other: Do men and women spend

differently on wine?

3. Steps to Perform a t-test

1. State your question clearly.

Example: Has the average spending increased after a campaign?
Hypothesis:

o Null Hypothesis (H₀): No change (e.g., spending stayed the

same).

o Alternative Hypothesis (H₁): Spending increased.

2. Choose the correct t-test:

o One-sample t-test for one group.

o Two-sample t-test to compare two groups.

3. Collect data: Measure the average (mean) and variability (standard

deviation).

4. Calculate a "t-value" using a formula that includes the mean,

standard deviation, and sample size.

4. Deciding the Result

  Compare the calculated t-value with a critical value (from a table
or software).

 If the t-value is bigger, it means the difference is significant, and you

can reject the null hypothesis.

5. Real-World Example

 Before a campaign: Average spending = $100.

 After a campaign: Average spending (sample) = $120.

 Use a t-test to check if the increase to $120 is real or just random.

Tema 7b

1. Chi-Square Test (powiązanie)

 Purpose: Checks if there is a relationship between two categorical

variables (e.g., age group and loyalty app usage).

 Steps:

1. Set a hypothesis:

 H₀ (null hypothesis): Variables are independent (no

relationship).

 H₁ (alternative hypothesis): Variables are related.

2. Collect data: Count how many times each category happens

(e.g., how many young vs. old people use the app).

3. Calculate the Chi-Square value (χ²):

 Compare observed counts (actual data) with expected

counts (what we expect if variables are independent).

 Formula:

OO: Observed values, EE: Expected values.

4. Compare χ² to a critical value:

 If χ² is larger, reject H₀ (there is a relationship).

2. ANOVA (Analysis wof Variance)

 Purpose: Compares averages of more than two groups to see if

they are different.

o Example: Do three advertising campaigns lead to different

sales?

 Steps:

1. Set a hypothesis:

 H₀: All groups have the same average.

 H₁: At least one group is different.

2. Collect data: Measure the variable for each group (e.g., sales
from each campaign).

3. Calculate F-value: (rozstrzał)

 Compares variation between groups to variation

within groups.

4. Compare F to a critical value:

 If F is larger, reject H₀ (there is a difference).

When to Use These Tests:

1. Chi-Square: For yes/no or category comparisons (e.g., does age

affect app use?).

2. ANOVA: For comparing averages across multiple groups (e.g., which

campaign is most effective?).