Whenever three variables all change together systematically, it is impossible
to reach a simple, clear conclusion about the relationship between any two
of them. Thus, whenever a confounding variable exists, internal validity is
threatened. One more look at the music and problem-solving study should illustrate
this point. This time, suppose that one group of individuals is given a
problem-solving task in a room with background music at nine o’clock in the
morning. A second group is given the problem-solving task in a room with no
music at four o’clock in the afternoon. Finally, suppose that the results show
much better performance for the first group than for the second group. Note
that this study involves three variables and that all three variables change together
systematically: As the background changes from music to no-music, the
time of day also changes from morning to afternoon, and the students’ performance
changes from good to bad.
The researcher would like to explain the results by saying that there is a
relationship between background music and performance:
Changing the background from music to no-music caused a decrease in
performance from good to bad.
However, the time of day (morning or afternoon) is an extraneous variable
that has turned into a confounding variable and provides an alternative
explanation for the results. Specifically, it is possible to explain the results by
saying that there is a relationship between time of day and performance:
Changing the time of day from morning to afternoon caused a decrease
in performance from good to bad.