The uses and restrictions of commonality analysis are described. Commonality analysis has been increasingly used as a method to examine the relative importance of independent variables, through the partitioning of variance among the variables of the regression equation into unique and common components. The effects of all other independent variables are partialed out, and the effect is attributed to the variable of interest. Problems are encountered, using this method. One limitation involves the possibility of negative commonalities, which can occur when one variable negates or suppresses another. A second problem is the proliferation of commonality components as more independent variables are added, and the resulting difficulty of making substantive interpretations. The third difficulty is that commonality analysis does not help to differentiate between situations in which high commonalities are the result of using duplicative variables and situations in which high commonalities reflect causal relationships. It is concluded that commonality analysis should be limited to research which has as its objectives either description or prediction. (MH)