How To Draw Q-Q Plot

How
to draw a Q‐Q plot for checking the
validity of your assumed probability
distribution for the data?

Introduction
This document illustrates step by step procedure of how to produce a Q‐Q plot to check normality of
a given set of data. We are using data in “Travel Distance.xlsx” as an example. We assume that the
data can be modelled by a normal distribution with mean and variance specified by the data
analysts.
Procedure

Step 1. Sort data
We first need to sort the data in ascending order. To do this, highlight (use mouse to select) the
column of data, and then click “Sort” button on the tool bar. Refer to Figure 1 below:

Figure 1
On the popup window, choose “Continue with current selection” option, then click “Sort”.

Step 2. Summary statistics of data
Use the formula shown in the figure below to calculate some basic measurements of the data:

Step 3. Construct Histogram
Histogram is the first graphical view to guess what kind of probability distribution that the data
might follow. Each probability distribution has its unique PDF curve and the histogram of given data
should approximate the shape of the PDF curve of the probability distribution that it follows.
To construct the histogram, we need to first decide an appropriate set of bins. For the given data,
the minimum value is 72.57 and the maximum is 125.48. We can thus define the bins as the
following:

Now we can use “Histogram” function provided in “Data Analysis Tools” as shown in the screenshot
below. In order to access this tools, you have to install Data Analysis Toolpak first.

In the popup window, select “Histogram”:

Then specify the range of data and range of bins as shown in the screenshot below. You may play
around with different options and see what the results are.

A new worksheet will be added with a bar chart drawn:

Next, please fine‐tune the bar chart to be a histogram. There are two steps involved: first point your
cursor to any bar in the chart and right click your mouse to find “Format Data Series…”

The second step is to change “Gap Width” to be 0%. Then a nice histogram will be displayed.

The histogram seems to be close to bell‐shaped. We can therefore guess that the data might follow
normal distribution. However, please note that histogram can be misleading. The look of histogram
is largely dependent on the bins that you defined. It can only help us to narrow our search of
probability distribution to fit our data, but cannot be the sole way to determine the probability
distribution.
Further to it, normal distribution is not the only probability distribution that is bell‐shaped. Student
distribution, logistic distribution, etc have similar bell shapes. There are quite a number of methods
developed to test the normality of data. Q‐Q plot is one of them. In a Q‐Q plot, we compare the
quantiles of our empirical data with the theoretical data (ideal data that follows exact normal
distribution) and see how close they are.
The following steps explain how to construct a Q‐Q plot.

Step 4. Build a Column of Quantiles
There are several choices of quantiles to be used to construct Q‐Q plot. The simplest of them, given
a sample size of n, is k/n for k=1, 2, …, n. Other choices include the use of k/(n+1) and (k‐0.5)/n,
which, by minor adjustment in either the denominator or numerator, very often give better result
than k/n. In this document, we are going to use (k‐0.5)/n. You may try other choices.
The following screenshot shows part of quantile column developed. You can copy the formula to the
whole column. Note that $B$104 refers to the “Count” cell.

Step 5. Build a Column of Theoretical Data
The following screenshot shows the formula of theoretical data, where B102 is mean and B103 is
standard deviation derived from the given data.

Step 6. Q‐Q Plot
Select both columns of original data and theoretical data, then insert a scatter plot with them.

We are getting something like the following:
140
120
100
80
60
40
20
0
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00

As you can see, the data points almost form a straight line.
Step 7. Fine‐tuning Q‐Q Plot
There are several things we can do to fine‐tune the Q‐Q plot.
Option 1. Format axis to be a square plot
The points are a bit far away from the origin. Visually, it seems just occupying a corner of the graph.
We can somewhat “zoom in” by formatting the axis.
a. Click on the y‐axis.

b. Right click on the selected field and then choose “Format Axis”.

c. In the panel popup at the right of Excel worksheet,
change minimum to 70.

d. Repeat the same for the x‐axis.

We are going to see the following:
130
120
110
100
90
80
70
70.00 80.00 90.00 100.00 110.00 120.00 130.00

Option 2. Add a 45‐degree line on the scatter plot
Follow the screenshot below to add a line to the scatter plot. Please go to “INSERT” Ribbon and find
Shapes. Then add a line to the scatter.
130
120
110
100
90
80
70
70.00 80.00 90.00 100.00 110.00 120.00 130.00

If most of your data points scatter on the 45‐degree line like the chart above, you can conclude that
your assumed probability distribution is valid to model your sample data. In this application,
normality assumption on the data is tenable.

How To Draw Q-Q Plot

Uploaded by

Copyright:

Available Formats

How To Draw Q-Q Plot

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

How To Draw Q-Q Plot

Uploaded by

Copyright:

Available Formats

How

You might also like