P-P plot Vs Q-Q plot

[Karen Wood]

I wish to test for normality using SPSS. (for scores on ability tests). Which is best, the Normal P-P (probability) Plot with Expected Cumulative Probability Vs Observed Cumulative Probability OR the Q-Q Plot (quantile) of Expected Normal Vs Observed value.
I believe that differences in the middle of the distribution are more apparent with P-P Plots and the tails Q-Q plots. But this information doen't make it any easier to know which one to use.
Any suggestions?

Karen

[Art Kendall]

Use an Aristotelean approach, slip between the horns of the dilemma. Use both.

However, in a vast array of statistical procedures, normality of variables is seldom as important as normality of residuals. Both of these kinds of plots are among the useful tools in examining residuals.

Art Kendall
Social Research Consultants

Karen Wood wrote:

I wish to test for normality using SPSS. (for scores on ability tests). Which is best, the Normal P-P (probability) Plot with Expected Cumulative Probability Vs Observed Cumulative Probability OR the Q-Q Plot (quantile) of Expected Normal Vs Observed value.
I believe that differences in the middle of the distribution are more apparent with P-P Plots and the tails Q-Q plots. But this information doen't make it any easier to know which one to use.
Any suggestions?

Karen

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

[ajayohri]

Why not use Socrates approach instead of Aristotle approach?



Because Socrates never published.


Ohri

Graduate Student
University of Tennessee, Knoxville.


Go Vols!

Websites-
http://decisionstats.com
http://dudeofdata.com
http://prayers2go.com

Linkedin- www.linkedin.com/in/ajayohri
Facebook-www.facebook.com/ajayohri
Twitter-www.twitter.com/dudeofdata

Quote for the Day-
Samuel Goldwyn  - "I'm willing to admit that I may not always be
right, but I am never wrong." -
http://www.brainyquote.com/quotes/authors/s/samuel_goldwyn.html


On Thu, Dec 3, 2009 at 7:21 AM, Art Kendall <[hidden email]> wrote:

> Use an Aristotelean approach, slip between the horns of the dilemma. Use
> both.
>
> However, in a vast array of statistical procedures, normality of variables
> is seldom as important as normality of residuals. Both of these kinds of
> plots are among the useful tools in examining residuals.
>
> Art Kendall
> Social Research Consultants
>
> Karen Wood wrote:
>
> I wish to test for normality using SPSS. (for scores on ability tests).
> Which is best, the Normal P-P (probability) Plot with Expected Cumulative
> Probability Vs Observed Cumulative Probability OR the Q-Q Plot (quantile) of
> Expected Normal Vs Observed value.
> I believe that differences in the middle of the distribution are more
> apparent with P-P Plots and the tails Q-Q plots. But this information doen't
> make it any easier to know which one to use.
> Any suggestions?
>
> Karen
>
>
> ===================== To manage your subscription to SPSSX-L, send a message
> to [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command SIGNOFF SPSSX-L For a list of
> commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[SR Millis-3]

Karen,
The Q-Q plot is better than the P-P plot when assessing the goodness of fit in the tail of the distributions: the values of the P-P plot will tend towards 100%, so bad fit in the tail is not always apparent.

The normal quantile plot is more sensitive to deviances from normality in the tails of the distribution, whereas the normal probability plot is more sensitive to deviances near the mean of the distribution.

~~~~~~~~~~~
Scott R Millis, PhD, ABPP (CN,CL,RP), CStat, CSci
Professor & Director of Research
Dept of Physical Medicine & Rehabilitation
Dept of Emergency Medicine
Wayne State University School of Medicine
261 Mack Blvd
Detroit, MI 48201
Email:  [hidden email]
Email:  [hidden email]
Tel: [phone]
Fax: [phone]


--- On Thu, 12/3/09, Karen Wood <[hidden email]> wrote:

> From: Karen Wood <[hidden email]>
> Subject: P-P plot Vs Q-Q plot
> To: [hidden email]
> Date: Thursday, December 3, 2009, 2:43 AM
>
>
>
>
> I wish to test for normality using SPSS. (for scores on
> ability tests).
> Which is best, the Normal P-P (probability) Plot
> with Expected
> Cumulative Probability Vs Observed Cumulative Probability
> OR the Q-Q
> Plot (quantile) of Expected Normal Vs Observed value.
>
> I believe that differences in the middle of the
> distribution are more
> apparent with P-P Plots and the tails Q-Q plots. But this
> information
> doen't make it any easier to know which one to use.
>
> Any suggestions?
>
>
> Karen
>
>
>
>
>
>
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Steve Simon, P.Mean Consulting]

Art Kendall wrote:

> Use an Aristotelean approach, slip between the horns of the dilemma. Use
> both.
>
> However, in a vast array of statistical procedures, normality of
> variables is seldom as important as normality of residuals. Both of
> these kinds of plots are among the useful tools in examining residuals.
>
> Karen Wood wrote:
>>
>> I wish to test for normality using SPSS. (for scores on ability
>> tests). Which is best, the Normal *P-P (probability) Plot *with
>> Expected Cumulative Probability Vs Observed Cumulative Probability OR
>> the *Q-Q Plot (quantile) *of Expected Normal Vs Observed value.
>> I believe that differences in the middle of the distribution are more
>> apparent with P-P Plots and the tails Q-Q plots. But this information
>> doen't make it any easier to know which one to use.
>> Any suggestions?

If you use both, you have to decide which to trust when they disagree
and which to report in your publication, even when they don't disagree.
There is a strong preference in the research community for using q-q
plots rather than p-p plots, so I would encourage you to do the same,
unless you like swimming against the tide.

Also, from a practical perspective, one of your most important tasks
will be identification of outliers and whether outliers appear more
frequently on the low end versus the high end of the distribution. Thus,
anything that emphasizes the middle of the distribution rather than the
extremes is likely to miss this.
--
Steve Simon, Standard Disclaimer
The Monthly Mean is celebrating its first anniversary.
Find out more about the newsletter that dares
to call itself "average" at www.pmean.com/news

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD