For a study on the effect of social media on exercise habits, you assign individuals in groups to either check their social media accounts daily or weekly (good luck on subject compliance, but let us pretend). After 2 weeks, you ask them to record how often they exercised. Explain why it would be better to use a between subjects design versus a within subjects design. Give three reasons, and explain how those apply to this example.

Question

What will be an ideal response?

Answer 1

Ans: Between subjects designs are the default when one thinks of an experiment. The main advantage is that the effect of a particular treatment can be isolated because participants are being exposed to only one condition, and all other (controllable) aspects of the experiment are held constant. In this case, you would have to control other variables (such as how often the participants exercised before the study and how often they actually checked their accounts beforehand as well). To do this, it might be better to match your groups and then randomly assign them to conditions. That way, you can make sure the only difference between your groups is social media usage. You also lower the likelihood that participants will guess the hypothesis because they are exposed to only one condition. In this case, participants might assume that the researchers think social media affects something, but unless you tell them ahead of time they will not know you are studying exercise. If you had the same participants go through both conditions, they would get an idea because you already asked. Finally, there will not be any carryover effects. If your participants started off with only checking social media once a week, and they got bored and started exercising, at the end of 2 weeks, a habit might have formed. Then, once they are allowed to check daily, the participants might have the carryover effect of exercising regularly. So, the researcher would be unable to determine the effect of social media usage on exercise. While the result would be healthy for the participants, the lack of results would be damaging to the researcher.