CLIFFORD KONOLD AND ALEXANDER POLLATSEK
influenced by various factors. Furthermore, we cannot see how comparing the two
groups is meaningful unless we have (a) an implicit model that gender may have a
real genetic effect on height that is represented by the difference between the
average for men and the average for women, and (b) a notion that other factors have
influences on height that we will regard as random error when focusing on the
influences of gender on height.11 Thus, we claim that the concept of an average as
approximating a signal, or true value, comes more clearly into focus when we are
considering the influence of a particular variable on something (in this case, gender
on height). Such a comparison scheme provides a conceptual lever for thinking
about signal (gender influences) and noise (other influences). We return to this point
later.