Confounding

Present students with studies that could benefit from an additional control group. Have students describe control groups that should be added. Alternatively, you could present them with an experiment that used several levels of an independent variable, but leave out one or more of the groups. Have students generate both the missing group(s) and the rationale for these groups. Tp help students appreciate the value of appropriate comparison groups, you might refer them to this interesting claim about automobile headlights. (This short article, from Consumer Reports, spells out the problem with the advertiser's claim.) 

Alternatively, you could use some humor and focus on the issue of confounding by using an exercise proposed by Michael Durnam.


Back to Chapter 11 Main Menu

Back to Research Design Explained Home Page