ABSTRACT
The discussions in this and in the next chapter assume the fixed-effects model, in which the two or more levels of the independent variable that are being compared are all of the possible variations of the independent variable (e.g., female and male), or have been specifically chosen by the researcher to represent only those variations to which the results are to be generalized. For example, if ethnicity were the independent variable and there were, say, a white group and two specifically chosen nonwhite groups, the fixed-effects model is operative and the results should not be generalized to any nonwhite group that was not represented in the research. Methods for dependent groups are discussed in the last section of this chapter.
