I like the idea of providing the actual p value and leaving it to the reader
to decide but I don't think many readers will, without any justification
being provided, give much weight to a result with a probability greater than
.05. And an author should never try the scam (which I have seen) of
presenting the results like, "Group A scored higher than Group B,
t(24)=2.31, p=.06" without pointing out to the reader that the alpha
standard for significance was not met (even though they didn't actually use
the word "significantly" in the presentation). In fact, with such a result,
you would need to provide some additional justification for putting any
stock in those results.

In addition to providing actual p values and letting the reader decide, the
presentation of confidence intervals can give readers more help in deciding
how much weight to give an experimental result. It provides the range in
which the population mean is likely to be found. 

I am also always curious with z or t results with p-values less than .05 but
greater than .025. I remember one time that I was looking closely at the
results of a research article and I determined that, without mentioning it
or really having any justification for it, the author had used a one-tailed
probability, most likely because the two-tailed would not have been
significant. If you have the time, such checks are now easy in many cases
with the probability functions of many distributions calculated in Excel.

I also agree that nonsignificant results give us a great opportunity to
address the topic of power analysis. The higher the power, the more we might
make from nonsignificant results being actually indicative of no effect in
the population.

The suggestions of the Task Force on Statistical Inference of the APA Board
of Scientific Affairs on this and many other statistical issues is available
at: http://www.apa.org/journals/amp/amp548594.html. It was originally
published in the American Psychologist in August 1999, Vol. 54, No. 8,
594-604. The task force also repeatedly referenced the seminal work of
Cohen: 

Cohen, J. (1994). The earth is round (p < .05). American Psychologist, 49,
997-1003. 

Rick

Dr. Richard L. Froman
Psychology Department
John Brown University
Siloam Springs, AR 72761
e-mail: [EMAIL PROTECTED]
phone and voice mail: (479)524-7295
http://www.jbu.edu/sbs/rfroman.html

-----Original Message-----
From: Martin J. Bourgeois [mailto:MartyB@;uwyo.edu]
Sent: Monday, November 11, 2002 10:38 AM
To: Teaching in the Psychological Sciences
Subject: RE: Marginally Significant?


I like Bob Abelson's advice about this: report the alpha level (e..g,
.05) along with the observed p value (.056) and the effect size, and let
the journal editor and reader decide what to make of it. Abelson has a
great explanation of the topic in Chapter 4 of his great book,
Statistics As Principled Argument. He also discusses John Tukey's
suggestion that we call results between p=.05 and p=.15 as leaning
toward significance, and results from p=.15 to p=.25 as hinting toward
significance. Thinking in this way avoids what Abelson calls categoritis
regarding statistical tests and gets students to think about what
statistics do.

Marty Bourgeois
University of Wyoming

-----Original Message-----
From: Rob Flint [mailto:flintr@;mail.strose.edu] 
Sent: Monday, November 11, 2002 5:40 AM
To: Teaching in the Psychological Sciences
Subject: Marginally Significant?

One of my students doing her senior thesis ran her stats and got results
of
.056 and .08 for two different ANOVAs. In the past I have seen published
studies indicating that these are "marginally significant." How do you
deal
with results of this nature? More importantly, do you have any citations
(journals or books) that discuss the value of including/discussing
results
that seem to "approach significance"?

Thanks,

Rob Flint
-------------------------------------------------------------
Robert W. Flint, Jr., Ph.D.
The College of Saint Rose
Department of Psychology
432 Western Avenue
Albany, NY  12203-1490

Office: 518-458-5379
Lab: 518-454-2102
Fax: 518-458-5446

Behavioral Neuroscience Homepage:
http://academic.strose.edu/academic/flintr/
Department of Psychology Homepage:
http://academic.strose.edu/academic/psychology/index.htm


---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to
[EMAIL PROTECTED]

---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]

---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]

Reply via email to