At 09:52 AM 10/22/99 -0500, Paul Brandon wrote:
>>Rather than talk about reliability and validity (yes, both would be
>involved), I'd rather talk about the multiple determination of human
>behavior:
>in this case the stimulus control exerted by the subject matter under
>investigation competing with the reinforcement contingencies of the
>researcher and her staff.
>
>Barring a major change in the contingencies set by society, the only answer
>is replication, both direct and systematic.
There are other factors that could play an important role, such as
differences from lab to lab in small methodological and environmental
details ranging from time of day to environmental context cues.
One powerful factor that can lead to non replication are subject
differences. For example, if a cognitive study originally run at a school
with a relatively homogeneous group of students with average SATs of 1400,
is not replicated at a school with a much wider range of students and
average SATS of 950 it may not be at all surprising. These studies are
extremely important because they often require us to reject, qualify or
revise our existing (usually too simple theories) to include the influence
of factors like individual differences.
Dawn
Dawn G. Blasko Ph.D.
Associate Professor of Experimental Psychology
Penn State Erie, The Behrend College
Station Road
Erie, PA 16563-1501
Office phone: 814-898-6081
http://www.pserie.psu.edu/hss/psych/blasko.htm
