Tipsters,
A true experiment requires both a manipulated independent variable and
random assignment to condition. It is not necessary to have any control
group. As mentioned in one of the posts, a control group, in the
example in question, wouldn't make sense. Control groups for specific
purposes not as a requirement of experimental design. It's important
students understand causal determination can be made without a control
group - as long as the other conditions are met. Just because a study
meets the basic requirements of a true experiment doesn't rule out many
threats to internal validity and questions as the external validity as
well.
In the example, sex of subject is a subject variable and cannot be
randomly assigned to condition. Therefore, it is not possible to
conclude that sex of subject causes anything. I agree with a previous
post about the limited ecological validity. I encourage students to be
very careful in generalizing the findings beyond setting as well. In
the example mentioned, helping a confederate in a parking lot may not be
a wise thing for women to do - at least these days. One of my students
did a similar study examining helping behavior and sex differences. In
this study, a person tripped in front of them and dropped their books.
The DV was helping the confederate or not helping (nominal) the
confederate. Participants were randomly assigned to sex of confederate
condition. The frequency of male and female helping and not helping
the confederate (male or female) were the same. Hope this helps,
Diana Kyle
Palomar College
909-744-1150 X5032
