Unit 1 – Exploring One-Variable Data
Notes 7 – Comparing Distributions
Shape
• When asked to describe a distribution with quantitative data, we have four items we want to comment
on:
o
o
o
o
• We do NOT describe categorical distributions with the above – why?
o We cannot find the __________________ of a categorical variable
o ______________ doesn’t matter because we can put the data in any order we want
Common Shapes
Once the distribution is graphed, the first thing we identify is the shape of the graph. Here are some examples of
common distribution shapes you will encounter.
Unimodal Bimodal Uniform
Symmetric Left Skewed Right Skewed
Graph Shapes
• We often use two or more of the shape descriptions above to describe what a graph looks like
Distribution Histogram Stemplot Boxplot
Symmetric
Uniform
Symmetric
Bimodal
Unimodal
Left Skewed
Unimodal
Right
Skewed
Multimodal
None of the
main shapes
above
Outliers
• We can comment on any _______________________
• Use the 1.5*IQR rule to mathematically determine outliers
• Use the 2*standard deviation rule to mathematically determine outliers
Center
• When deciding what measure of center to use, it is best to consider the shape and if the distribution has
extreme values (which may or may not be outliers)
• The mean is ___________________ to extreme values, so it is best used with symmetric, non-skewed
distributions
• The median is ___________________ to extreme values, so it is best used with skewed distributions
In a skewed distribution (which doesn’t always have outliers), the median is a good indication of center because
the mean gets pulled towards the tail of the distribution.
Unimodal, Symmetric Left Skewed Right Skewed
Spread
You can describe the spread in a few different ways:
• If you have data and use the mean as the measure of center, find the ______________________
• If you have data and use the median as the measure of center, find the __________
• If you do not have data, report the _____________ from the graph
Example: Describe the distribution below by commenting on the shape, center, spread, and potential outliers.
Word Length for Taylor Swift’s Song, “I Can Do It with a Broken Heart”
𝑥̅ s min Q1 med Q3 max
3.5 1.94 1 2 3 4 10
Comparing Distributions
Do students in AP classes spend more time on homework than students in non-AP classes? A random sample of
21 "AP Students" (students enrolled in 1 or more AP classes), and 20 "non-AP students", and they were asked
how long they spent on homework last night. The data is shown in the histograms below.
Time Spent on Homework Last Night for AP Students Time Spent on Homework Last Night for non-AP Students
Summary Statistics: Summary Statistics:
n 𝑥̅ s min Q1 med Q3 max n 𝑥̅ s min Q1 med Q3 max
21 88.905 25.894 15 82.5 95 106.5 120 20 47.3 17.357 5 39 51 60 70
Task: Compare the two distributions.
When we compare two distributions, we want to comment on shape, outliers, center, and spread, but we want to
make sure we are using comparison words and descriptions.
For Example:
• “The distribution of Group A is slightly skewed right, while Group B is roughly symmetric.”
• “Group A typically scores higher than Group B.”
• “The mean of Group A’s distribution is greater than the mean of Group B.”
• “Group B shows more variability than Group A, as seen in the larger IQR.”
• “There is an outlier in Group A that Group B does not have.”
• “The range of Group A is about the same as the range of Group B.”
Do not do… What you should do…
The shape of the AP student homework time is left
Both distributions appear left skewed, while the AP
skewed with an outlier, and the shape of non-AP
student homework group appears to have a low visual
student homework time is left skewed with no
outlier, the non-AP homework group does not.
outliers.
The mean amount of time AP students spend on
The mean amount of time AP students spend on
homework is 88.905 minutes, which is higher than
homework is 88.905 minutes. The mean amount of
the mean amount of time non-AP students spend on
time non-AP students spend on homework is 47.3
homework, which is 47.3 minutes.
minutes.