Another example of this
type of problem is a taste-test study that compared consumer preference for
Coca-Cola versus Pepsi-Cola. In this study, individuals were asked to taste the
colas in two different glasses and identify the one they preferred. The participants
were not told which cola was in each glass but the glasses were marked with a Q and an M so that the researcher could record the responses. However,
the glass containing Coca-Cola was always marked with the letter Q, and the
Pepsi-Cola glass was always marked with the letter M. Although the results
indicated that people prefer Pepsi, an alternative explanation is that people
prefer the glass labeled with the letter M (Huck & Sandler, 1979). In this
study, the identifying letter was allowed to vary systematically with the brand
of cola, so the letters M and Q became a confounding variable. To avoid confounding
variables and ensure the internal validity of a research study, it is
necessary that there are no systematic differences in the general environment
from one treatment condition to another. Whenever a difference exists, there
is an alternative explanation for the results and the internal validity of the
study is threatened.