p values - OncoLink FAQ: "P Value" in Statistical Analysis

OncoLink FAQ: "P Value" in Statistical Analysis

Last Revision Date: Friday, 13-Apr-2001 16:32:20 EDT

Copyright ? 1994-2001, The Trustees of the University of Pennsylvania

This is a response to a question about the meaning of the term "p value" in

statistical analysis.

Question:

While researching my diagnosis I have come across the term "p value" quite

often. What does this mean?

Thank you,

MH

Todd Doyle, MD, OncoLink Editorial Assistant, responds:

Dear OncoLink readers:

Thank you very much for your interest and questions regarding the meaning of

the term "p value".

The "p value" is a statistical term found in almost all scientific papers in

which two or more outcomes are compared. By definition, a p value is the

probability that a given outcome, or one more extreme than that outcome

could have occurred by chance alone.

Research papers often state that one treatment was better than another and

that the difference in outcome for example was statistically significant

with a p value <0.05. This means that there was a less than 5% probability

of the observed difference occurring between the outcomes of the two

different groups of patients who got the different treatments. It is very

unlikely that some fluke caused the difference in survival and alternatively

the difference is likely secondary to a better treatment, all other things

being equal.

The actual significance level is somewhat arbitrarily chosen, but the lower

the value, the less likely it is that the difference between the two

outcomes occurred by chance alone. Most investigators agree that p>0.05 is

not significant, p<0.05 is significant, and p<0.01 is highly significant.

Some experts would consider p<0.10 to be marginally significant and would

leave it up to the reader to decide whether the information is sufficient to

conclude that the difference is not simply due to chance alone.

One important thing to keep in mind is that just because the difference

between two treatments is not statistically significant (p>0.05), does not

mean that the two treatments are the same. Rather, there may be insufficient

evidence to be certain the differences are "real" and not due to chance. One

reason that this sometimes happens is that the number of patients enrolled

in the particular study was not large enough to show the difference.

Furthermore, a statistically significant difference does not mean that the

difference is clinically significant. A comparison between toxicity of

treatment X = 10.1% and treatment Y = 10.2% may be statistically

significant, but may also be so small as to have little practical, impact.

For a more detailed description of the statistics regarding p values and

tests of significance the reader is referred to any number of medical

statistics texts including the following:

1. Kuzma JW. Basic Statistics for the Health Sciences, 3rd Edition, Mayfield

Publishing Company, 1998.

GENERAL DISCLAIMER

OncoLink is designed for educational purposes only and is not engaged in

rendering medical advice or professional services. The information provided

through OncoLink should not be used for diagnosing or treating a health

problem or a disease. It is not a substitute for professional care. If you

have or suspect you may have a health problem, you should consult your

health care provider.

ONCOLINK PRIVACY STATEMENT

--

Posted from [38.26.235.37] by way of oe64.pav1.hotmail.com [64.4.30.199]

via Mailgate.ORG Server - http://www.***.com/