3/4

¾ GENERAL MATHEMATICS
Core: Univariate Statistics

Name:____________________________
 Chapter 1 Types of Data
UNIVARIATE DATA (One variable only)

Types of Variables (CLASSIFYING DATA)

1|P a g e
3. The level of water usage of 200 homes was rated in a survey as low, medium or high and the size of
the houses as small, standard or large. The results of the survey are displayed in the table below.

Water Usage House Size

Large Standard Small
Low 2 10 7
Medium 19 60 13
High 19 40 30
The variables, level of water usage and size of house, as recorded in this survey are:

A. Both Nominal categorical variables

B. Both Ordinal variables
C. Nominal and Ordinal variables respectively
D. Continuous and Ordinal variables respectively
E. Ordinal and Nominal variables respectively

2|P a g e
 Chapter 1B Displaying & Describing distributions of

categorical variables
Univariate Analysis

Representing CATEGORICAL DATA

Bar Chart
Example

3|P a g e
Bar Charts:
- For categorical data
- Length (or height) of each bar represents a measure of frequency
- Width of bars must be equal
- Bars do not touch one another

Segmented Bar Charts:

- For categorical data
- A single bar is used, all other bars that could be used are stacked atop one another
- Length of segment relates to frequency
o Usually percentage frequency
- Height of bar always equates to 100% or total frequency
- Often asked to compare two segmented bar charts (technically bivariate data)

Example

4|P a g e
Representing and Analysing NUMERICAL DATA

The main objective when analysing univariate data of numerical data is to determine

 Type of distribution (shape and outliers (if any))

 Central tendency (centre)
 Spread

Frequency Tables
These are used to display data
- Can use class intervals (groups) depending on the type of data
o Continuous data is most likely to use class intervals
- Usually draw histograms or bar charts from this information

Example:
Calculate summary statistics on your calculator using the following two tables
a) b)
Siblings Frequency
0 6 Weight (kg) Frequency Midpoint
1 4 600 -< 650 6
2 10 650 -< 700 4
3 48 700 -< 750 10
4 22 800 -< 850 22
5 10 850 -< 900 10
*Must use midpoint in this case

5|P a g e
Histograms:
- For numerical data only (usually continuous but discrete can be given the right circumstances)
- Height of bars gives a frequency measure
- Bars must touch (unless a frequency of zero for an interval exists)
- If discrete, bars lie above numerical values
- If continuous, bars lie between lower and upper values of interval

* For Hstart use a value lower than your lowest interval.

* For Hstep this should be the size of your interval

* Using the ‘Trace’ function allows you to determine the frequency of each bar.

* For grouped data, use the Midpoint in list 1 and change the frequency to list 2. Still use the same
Hstep as the interval defines in your question.

6|P a g e
Analysing Histograms
Always look for: 1. Centre (median)
2. Spread (range)
3. Shape
Shape of Histograms

- Must ALWAYS mention outliers if there are any

Note: Usually not required to construct a histogram, however, 2016 Exam had a construction
question.

* Use a ruler to help investigate histograms

7|P a g e
Stem Plots (Stem and Leaf):
- Summarises small groups of numerical data
- Back to back uses categorical and numerical data (technically bivariate data)
- Turn a stem plot on its side and it will look like a histogram
- All values must be ordered
o Smallest value in interval goes closest to stem
- MUST have a legend

Regular Stem plot Split Stems

 The superscript 5 is not usually indicated

 Can break stems into other intervals not just 5’s

Back to Back Stem Plot

Dot Plots:
- Graphical display of discrete or categorical data sets
- Single dot represents each data point
- Shows shape of distribution
- Must have an even scale and dots placed above appropriate values

8|P a g e
Examples:

9|P a g e
10 | P a g e
11 | P a g e
12 | P a g e
Should be 19

13 | P a g e
LOGARITHMIC SCALES
- Used to enable us to view data with extreme values in one sample set.
- Only use log10 values (or powers of 10)

Converting to and from a Log Value

Using the lifespan of various animals. Complete the table below.

Animal Life Span (in days) Log (lifespan)

Rabbit 3.5623
Guinea Pig 1460
Horse 4.0394
Carp 73 000
Queen Ant 3.0394
Tortoise 69 350

Convert from a log

Convert to a log

=Log of the number

Number Log Scale

=10 to the power of log scale

= 10log scale

14 | P a g e
Examples:

15 | P a g e
16 | P a g e
17 | P a g e
Summary

Comments: When commenting usually you are required to quote a value. Below is a list of the most
suitable. Others can be ok but these are most suitable.

Display Centre Spread

Dot Plot Median Range
Stem Plot Median IQR/Range
Histogram Median IQR/Range
Box Plot Median IQR
Normal Distribution Mean SD

18 | P a g e
Statistical Measures of Numerical Data -Measures of Centre:
Mode:
- Most frequently occurring value
- Modal Interval: interval with highest frequency
- Not usually a great measure of centre

Median:
- Middle value of data
- 50th percentile
- Median is not affected by extreme values (outliers) or by skewed data
𝑛+1
- To find the median position use , where n is the number of values in a data set
2
- If n is odd, median is an actual value
- If n is even, median is between two values

Mean:
- Average value of the data set
- Is affected by outliers and only a good measure of centre when the data is symmetrical
- For grouped data, you must find the midpoint of the interval first (as per earlier in notes)
Σ𝑥
- Formula by hand: 𝑥̅ = Σ𝑓

19 | P a g e
Examples:

Question 4

Mr Higgins has a biology class of 24 students that averaged 65%. Another student new to the class took the
test and the new average is 64%. What did the new student score?

Question 5

Mr Grogan has two Systems classes. One has 21 students averaged 70% on a test and the other class of 14
students averaged 60%. What is the overall average of his two classes?

20 | P a g e
Statistical Measures of Numerical Data - Measures of Centre:
Range:
𝑅𝑎𝑛𝑔𝑒 = max 𝑣𝑎𝑙𝑢𝑒 − min 𝑣𝑎𝑙𝑢𝑒

- For grouped data it is the start of the smallest interval and the end of the largest interval
- Range is heavily affected by outliers
- Always include outliers

Interquartile Range (IQR):

- Middle of the 50% of the data
- Is not affected by outliers
- Must have data in order
- Lower quartile, Q1
o Median of first half of the data
o 25th percentile
- Upper quartile, Q3
o Median of upper half of the data
o 75th percentile

𝐼𝑄𝑅 = 𝑄3 − 𝑄1

Example:

Five Figure Summary:

- Min, Q1, Median, Q3, Max
- Min and Max remain even if found to be outliers

Outliers:
- Unusually large or small values
- They lie outside upper and lower fences

𝐿𝑜𝑤𝑒𝑟 𝐹𝑒𝑛𝑐𝑒 = 𝑄1 − (1.5 × 𝐼𝑄𝑅) 𝑈𝑝𝑝𝑒𝑟 𝐹𝑒𝑛𝑐𝑒 = 𝑄3 + (1.5 × 𝐼𝑄𝑅)

- When asked to look for outliers always make a final statement that supports your findings

21 | P a g e
Boxplots:
- Middle 50% of all data (IQR) is represented as a box
- Outliers need to be clearly marked
- A scaled number line must be used (when constructing)\
- When there are outliers present, the next highest/lowest value is marked as the new max/min
on the plot
o Not the upper or lower fence (unless this co-incides with the next value)

Relating a Box Plot to shape

Example

22 | P a g e
Drawing a Box Plot on your calculator.

After entering your data

Then click the graph icon (always check the show outliers box before continuing)

Use Analysis – trace to find the relevant values on the boxplot

23 | P a g e
Examples:

24 | P a g e
25 | P a g e
26 | P a g e
Standard Deviation (sx):
The standard deviation provides how a sample of values are spread about the mean.

- A small SD suggests data varies closely around the mean (less spread)
- A large SD suggests data varies widely about the mean (more spread)
- Different data sets can have the same mean but different SD’s
- Different data sets can have the same SD but different means.
- SD can be affected by outliers or skewedness
o Only use mean and SD for symmetrical (normally distributed )data

* It is always good to draw a little diagram next to a question to help visualise

Example
The average height of a team of basketballers is 196cm with a standard deviation of 12cm. A new member of
the team is exactly 196cm. What effect would his height have on the mean and standard deviation?

A. The mean and standard deviation would be the same

B. The mean would stay the same but the standard deviation would increase
C. The mean would stay the same but the standard deviation would decrease
D. Mean and standard deviation would both decrease
E. Mean and standard deviation would both increase

NORMAL DISTRIBUTION

27 | P a g e
28 | P a g e
Examples:

29 | P a g e
30 | P a g e
31 | P a g e
Standard Scores (Z):

𝒂𝒄𝒕𝒖𝒂𝒍 𝒔𝒄𝒐𝒓𝒆 − 𝒎𝒆𝒂𝒏

𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 𝒔𝒄𝒐𝒓𝒆 =
𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 𝒅𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏

𝒙−𝒙
𝒛=
𝒔

Example
Two students compare their scores for a test for two different subjects.
One student scored 73 out of 100 for an English test.
The other student scored 65 out of 100 for a Mathematics test.
Which student performed better?

32 | P a g e
Example:
A class test (out of 50) has a mean mark of 34 and a standard deviation of [Link]’s standardised test
mark was -1.5. What was Joe’s actual mark?

To interpret a z-score:

- Negative z-score of (-1.8)

o Is 1.8 standard deviations below the mean
- Positive z-score of (2.4)
o Is 2.3 standard deviations above the mean

* Always be careful of the context of the question when saying if someone has done better than
someone else.

33 | P a g e
Examples:

1.

Fruit content (%) Supermarket pie Bakery Pie

mean 24.8 32.4

standard deviation 3.9 3.6

a. What does the standard deviation indicate about the fruit content of supermarket pies compared with
bakery pies?

The label on a packaged pie bought at a local supermarket claimed the pie contained at least 30% fruit.

b. i). Calculate the standardised fruit content (z score) relative to our sample of supermarket pies. Write
your answer correct to two decimal places.

ii). What does the z score tell us about the fruit content of this particular pie relative to our sample of
supermarket pies?

c. One of the packaged pies was tested and given a z score of – 0.5. What was the actual fruit content (%) of
this particular pie? Write your answer correct to the nearest whole number.

The fruit content (%) of supermarket pies is approximately normally distributed.

d. Approximately what percentage of supermarket pies would be expected to have a fruit content less than
20.9% ?

e. Approximately what percentage of supermarket pies would be expected to have a fruit content between
20.9% and 28.7% ?

f. Calculate the actual percentage of supermarket pies, found in Table 1, with a fruit content between
20.9% and 28.7% to the nearest whole number.

34 | P a g e
35 | P a g e