[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: RE: advice re: "intervening" variable
The classical citation in psychology is this one (also one of the most
cited articles, 7940 is the current ISI count...):
Baron, R. M., & Kenny, D. A. (1986). The Moderator-Mediator Variable
Distinction in Social Psychological Research: Conceptual, Strategic, and
Statistical Considerations. Journal of Personality and Social
Psychology, 51(6), 1173-1182.
I know that there are some publications on this in medicine and public
health, but I am not really in that field therefore I can only point you
to these to this abstract:
Helena Chmura Kraemer and Michaela Kiernan
"Moderators and Mediators: Comparing the Baron & Kenny and MacArthur
Abstract: This is our second talk in the series on mediation, and this
one expands the scope to address moderation as well, which is the
process by which one variable changes the effect of a second variable on
a third variable. Establishing moderators and mediators to elucidate
causal chains is essential for social science research, and the most
popular set of statistical tests used to do this were outlined by Baron
and Kenny in 1986. Recently, a new set of statistical testing methods
were proposed by a MacArthur network subgroup to extend the Baron and
Kenny approach and resolve its ambiguities by: (1) imposing screening
criteria that identify whether a variable is eligible for consideration
as a moderator or mediator; (2) imposing analytic criteria that
demonstrate whether an eligible variable functions as a moderator or
mediator; (3) not drawing causal inferences from observational data; and
(4) considering other relations among variables besides moderation and
mediation. To assist researchers considering moderator/mediator issues
in their choice of which approach is better suited to their research
needs, we will explain these various approaches.
Bios: Dr. Helena Chmura Kraemer is Professor of Biostatistics in
Psychiatry in the Department of Psychiatry and Behavioral Sciences of
Stanford University. Dr. Kraemer's research focuses on experimental and
observational research methods in the behavioral aspects of medicine.
Her work has mostly been in psychiatry and health psychology, but has
also touched on other fields of medicine, including cardiology,
epidemiology, pediatrics, and oncology. Her most recent work has
focused on effect sizes that convey clinical significance and on
moderators and mediators of treatment in randomized clinical trials.
> -----Original Message-----
> From: email@example.com
> [mailto:firstname.lastname@example.org] On Behalf Of
> Richard Goldstein
> Sent: Monday, March 05, 2007 6:57 AM
> To: statalist
> Subject: st: advice re: "intervening" variable
> Hi all:
> This is more a stat question that a Stata question.
> I have a system of 3 variables where one variable is in
> between, in time and physiology, the other two variables:
> blood pressure -> diameter of blood vessel -> heart rate
> That is, a change in blood pressure "causes" some change in
> the diameter of the blood vessel which in turn "causes"
> some change in heart rate (actually in "RRI" which is,
> basically, the inverse of heart rate).
> I have never come across this situation before, but I believe
> that several substantive disciplines do have such situations.
> Actually, I have found a fair amount of literature in
> psychology; a recent overview is MacKinnon, DP, AJ Fairchild,
> and MS Fritz (2007), "Mediation Analysis", _Annual Review of
> Psychology_, 58: 593-614
> I believe that other disciplines might also see such
> situations, e.g., economics and epidemiology, but have been
> unable to find any literature.
> So, is there literature on this in disciplines other than
> psychology? If so, any citations, esp. overviews, would be
> greatly appreciated (regardless of whether the cite is to
> ariticles or books).
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
* For searches and help try: