• Skip to primary navigation
  • Skip to main content
  • Skip to primary sidebar
The Analysis Factor

The Analysis Factor

Statistical Consulting, Resources, and Statistics Workshops for Researchers

  • Home
  • Our Programs
    • Membership
    • Online Workshops
    • Free Webinars
    • Consulting Services
  • About
    • Our Team
    • Our Core Values
    • Our Privacy Policy
    • Employment
    • Collaborate with Us
  • Statistical Resources
  • Contact
  • Blog
  • Login

Anatomy of a Normal Probability Plot

by Karen Grace-Martin 3 Comments

A normal probability plot is extremely useful for testing normality assumptions.  It’s more precise than a histogram, which can’t pick up subtle deviations, and doesn’t suffer from too much or too little power, as do tests of normality.

There are two versions of normal probability plots: Q-Q and P-P.  I’ll start with the Q-Q.  The Q-Q plot plots every observed value against a standard normal distribution with the same number of points.  We have 111 observations in this data set, and you can see a histogram of the distribution on the right, and the corresponding Q-Q plot on the left.

image001

Across the bottom are the observed data values, sorted lowest to highest.  You can see that just like on the histogram, the values range from about -2.2 to 2.2.  (Note, these are standardized residuals, so they already have a mean of 0 and a standard deviation of 1.  If they didn’t, the plot would standardize them before plotting).

On the Y axis are the values that you would have gotten if they came from a standard normal distribution with the same number of data points–111.

This concept is a little strange if you’re not used to it, so think through it a bit.  Remember back to your intro stats class, when you learned these rules about a standard normal distribution:

  • The mean is 0 and the standard deviation is 1
  • 34% of points are between the mean and one standard deviation below the mean.
  • Another 12.5% are between one and 2 standard deviations below the mean.
  • The final 2.5% are above 2 standard deviations below the mean.
  • Because the distribution is symmetric, these same percentages apply above the mean.

So in a distribution of 111 points with a mean of 0 and standard deviation of 1, we know that the 56th point must have a value of 0—in a normal distribution the median equals the mean.

Only 2.5% of the 111 points—about 3 of them—should have values at or below -2.  So the 3rd value should be right around -2.

Likewise, there is an exact value that the 17th, the 42nd, each of the 111 ordered values would have under a standard normal distribution.  We don’t have easy ways to remember them, but those are the values plotted on the graph.

So if each value is exactly where it should be, if the distribution is perfectly normal, every single point would fall right on the line.  The further a point is from the line, the further it is from where a normal distribution would put it.

The following Q-Q plot shows a positively skewed distribution.  You can see that it’s not matching a normal distribution well at all—there are too many low values and too few high values.

image002

I find it helpful to always plot a histogram along with the Q-Q plot, to aid interpretation.  As you do more of these, you’ll get better at reading them without the histogram.

Q-Q vs. P-P

The Q-Q is plotting the quantiles—the actual values of X against the theoretical values of X under the normal distribution.  That’s what I’ve described above.

A P-P plot, one the other hand, plots the corresponding areas under the curve (cumulative distribution function) for those values.

In both, the points fall right on the line when normality has been met.  For the most part, the normal P-P plot is better at finding deviations from normality in the center of the distribution, and the normal Q-Q plot is better at finding deviations in the tails.  Q-Q plots tend to be preferred in research situations.

Both Q-Q and P-P plots can be used for distributions other than normal.

tn_assum_lmLearn more about each of the assumptions of linear models–regression and ANOVA–so they make sense–in our new On Demand workshop: Assumptions of Linear Models.

Four Critical Steps in Building Linear Regression Models
While you’re worrying about which predictors to enter, you might be missing issues that have a big impact your analysis. This training will help you achieve more accurate results and a less-frustrating model building experience.

Tagged With: normality, P-P plot, Q-Q plot

Related Posts

  • Checking Assumptions in ANOVA and Linear Regression Models: The Distribution of Dependent Variables
  • Why ANOVA is Really a Linear Regression, Despite the Difference in Notation
  • ANCOVA Assumptions: When Slopes are Unequal
  • How Simple Should a Model Be? The Case of Insignificant Controls, Interactions, and Covariance Structures

Reader Interactions

Comments

  1. Lauren Shead says

    March 21, 2020 at 9:28 pm

    I dont understand what “doesn’t suffer from too much or too little power” is- could you explain this more?

    Reply
    • Karen Grace-Martin says

      March 24, 2020 at 9:14 am

      Hi Lauren, have you seen this article? It goes into the Why of power calculations more including why you odn’t want too much: https://www.theanalysisfactor.com/5-reasons-sample-size-calculations/

      Reply
  2. Hasan Yahya says

    February 27, 2016 at 12:01 pm

    For convenient a reader in understanding this article, both the term Q-Q plot and P-P plot need to be cleared formerly. Thank a lot.

    Reply

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Please note that, due to the large number of comments submitted, any questions on problems related to a personal study/project will not be answered. We suggest joining Statistically Speaking, where you have access to a private forum and more resources 24/7.

Primary Sidebar

This Month’s Statistically Speaking Live Training

  • Member Training: Introduction to SPSS Software Tutorial

Upcoming Free Webinars

Poisson and Negative Binomial Regression Models for Count Data

Upcoming Workshops

  • Analyzing Count Data: Poisson, Negative Binomial, and Other Essential Models (Jul 2022)
  • Introduction to Generalized Linear Mixed Models (Jul 2022)

Copyright © 2008–2022 The Analysis Factor, LLC. All rights reserved.
877-272-8096   Contact Us

The Analysis Factor uses cookies to ensure that we give you the best experience of our website. If you continue we assume that you consent to receive cookies on all websites from The Analysis Factor.
Continue Privacy Policy
Privacy & Cookies Policy

Privacy Overview

This website uses cookies to improve your experience while you navigate through the website. Out of these, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may affect your browsing experience.
Necessary
Always Enabled
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Non-necessary
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
SAVE & ACCEPT