The Problem of Carryover Effects (2 of 2)
Next section: Assumptions of within-subject
designs
Consider an experiment comparing the time it takes to read a list of color
names with the time it takes to name colors. If a within-subjects design
is used, then all subjects are tested in both conditions. Carryover effects
are certainly possible in this design. The second task performed may be
performed better because of some kind of practice effect or because subjects
have become primed to say color names. alternatively, the second task performed
may be performed worse because subjects have become tired or bored. However,
in either case, the carryover effects would likely be symmetric.
Counterbalancing
can be used to control for symmetric carryover effects. In this experiment,
this means simply that half of the subjects would be given the color-name-reading
task before the color-naming task and the other half of the subjects would
be given the color naming task first.
Next section: Assumptions of within-subject
designs