Friday, September 2nd, 2005

 

Continuing from last time on the scientific method as applied to Psychology.

 

Hypothesis testing—systematic observations to determine if observations are consistent with your initial observation…

 

One of the best ways to test hypotheses, when it is feasible, involves conducting an experiment.  Let’s take the example from the Shaffer reading regarding the Kim et al. experiment about prenatal ethanol exposure.  It is only when we use experiments that we are able to make cause-and-effect conclusions about our hypotheses.s

 

What was the hypothesis being tested in the Kim et al. study?  Ethanol during gestation impairs learning.

 

      Important terminology regarding features of an experiment

                  independent variable: what was manipulated? Presence of ethanol in diet during gestation.  3 levels of IV: ethanol in diet; pair-fed controls; normal fed controls

                  dependent variable: what was measured? Learning in test of spatial aptitude.                 

operational definition: how did they define the constructs under study? Morris Water Maze performance, reduction in latency to finding platform over repeated days.

                  ramdom assignment: how did subjects get put into groups? Randomly assigned pregnant dams to feeding groups, one adult selected from each litter in adulthood.

                  Blind: did the experimenter know what group the subject was in? Experimenters were blind…previous research showed that experimenters who thought they were testing smart rats recorded better learning than experimenters who thought they were testing dull rats (both groups of rats were the same).

                  confounding variables (may be associated with design of the independent variables, assignment to groups, and to the measures under study): how might incorrect conclusions have been made? How do you prevent that from happening?  How did trials 59 and 60 correct for potential confounds? 

Trial 59: raised the platform to eliminate the possibility that group differences weren’t due to learning, but may have been due to impairments in using the visual information from around the room (it is these cues that are important in the Morris Water Maze).  If the ethanol- treated rats showed the same performance as control rats, then it can be concluded that the treatment did not appear to affect their ability to process visual information.  If the ethanol-treated rats also showed slower latency to finding the visible platform, then maybe they have a problem with their vision, not their learning ability.

Trial 60: removed platform and tested number of times the rats crossed over where the platform used to be—a measure of persistence.  If the ethanol-treated group gave us sooner, then maybe “learning impairments” were reflective of changes in motivation level, rather than cognitive deficit.

Overall, the findings supported the hypothesis that ethanol treatment led to deficits in cognitive performance.

      control groups/placebo effects: what were the control groups?  Why?

      not all systematic studies are “experiments”

                  correlational approach---can detect relationships, cannot discern causality

When consistency is obtained the hypothesis becomes a theory and provides a coherent set of propositions which explain a class of phenomena. A theory is then a framework within which observations are explained and predictions are made. If observations, made more carefully, start to become inconsistent with a theory, then theory becomes modified.

 

But when does theory and hypothesis give way to fact?  Science is ostensibly about "search for truth", but we never really get there do we, since part of the scientific method leaves room for work to be falsified.  We can never prove a theory correct (logically impossible to provide positive proof)—we can only provide support for it.   Preponderance of evidence helps to provide strong support—converging lines of research all coming to similar conclusions…

 

We can't prove theories correct, because the very nature of the scientific method allows for falsifiability (give example of something non-falsifiable).

 

And failing to find support for a theory may, but does not necessarily disprove it either, because being wrong may mean that we didn't observe carefully enough…if one study does not support the hypothesis, it does not falsify the theory on which the hypothesis is based.  However, if many studies fail to support the hypotheses that derive from a theory, we can begin to question the theory.

 

So, in scientific research, you will rarely see the term PROOF—but we as lay people use it all the time, especially in advertising!