You’ve probably seen social experiments before in the popular media – think about Big Brother House, Survivor, and other ‘reality’ shows where the producers manipulate variables in the environment and see what results. One of my favorite social experiments on TV is the hidden camera show “What would you do?”

The show’s anchor John Quiñones, test the public on ethnic issues regarding sexism, racism, bigotry, and hate. In the following episode we see him ‘testing’ the public’s reaction to racial profiling by a security guard.

While these shows are interesting, enlightening, and entertaining, they do not follow the experimental method very well. Generally a researcher will control a setting and carefully manipulate one (or more) variables watching for any change in other variables. These shows occur in natural settings, public that are impossible to control. That is not to say that social research can’t be conducted in real-life settings (see experiments in natural settings), just that there are many more concerns and less control. These shows also lack control groups. As you will read in the Babbie textbook (Chapter 8), the only way to insure internal validity is to test a treatment or research group (in which a variable is manipulated) against a control group (in which nothing is changed). In these cases, the research design calls for both a pre-test and a post-test to rule out other factors (aka confounding or intervening variables) which might be causing a change in the observed variable.

Moreover, the pre-test/ post-test scenario itself may be enough to cause a change in some variables. Think about it this way, if you wanted to test whether a lesson was effective in teaching students, you might give them a test on Monday. Then you may divide them into groups and teach half of the class about the topic (treatment group), while not teaching the other half (control group). Then you would give them all a test on Friday. More-than-likely you would see an improvement in everyone’s test scores. Why? Because they all had a pre-test and learned a little bit about the test itself (aka the ‘testing effect’). As a result of this effect, many researchers will use a Solomon Four Group Design to help test for control for the effect of the pre-test. In this design, two groups receive the treatment (one with and one without pretest) and two groups serve as controls (one with and one without pretest).


I would like for you to now read Babbie Chapter 8, take your chapter quiz, then do Excersise 12 – Social Experiments.

%d bloggers like this: