
Empirical Investigation ◾ 153
to eliminate the eects of other variables so that only the eects of the
independent variables are reected in the values of the dependent variable.
at is, we would like to eliminate the experimental error. Realistically,
complete elimination is rarely possible. Instead, we try to design the
experiment so that the eects of irrelevant variables are distributed
equally across all the experimental conditions, rather than allowing them
to inate articially (or bias) the results of a particular condition. In fact,
statisticians like, whenever possible, to measure the extent of the variabil-
ity under “norma ...