A meta-analysis of the effects of divorce on children's adjustment

Abstract

The two purposes of the present investigation were to: (a) clarify the large body of confusing and contradictory research results concerning the effects of divorce on children, and (b) provide an empirically verifiable knowledge base from which to better generalize about the substantive issues of children's adjustment to marital dissolution. In order to fulfill these goals, a complex integration of the literature was required and was undertaken through the use of meta-analysis. Methodologically, manual and automated literature retrieval procedures produced a group of 3073 possibly relevant citations. Previewing techniques reduced this number of references to an initial pool of 987 studies. Further screening resulted in a final sample comprised of 88 discrete studies or 69 data sets. The investigations included for analysis yielded a total of 846 effect sizes. Twenty-three mean effects were generated from research using mixed-sex samples. Sixteen of these values achieved practical significance (e.g., sex of the subjects, achievement and aptitude, teacher support). There were also 23 mean effects computed for the male-only subject group and practical significance levels were noted on 16 occasions (e.g., IQ, socialization, masculinity). Furthermore, investigations that analyzed females as a separate group produced 25 mean values. Fourteen of these effects were assigned levels of practical significance (e.g., parent-child interaction, behavior problems). Eight hundred one correlations were also available for analysis. Seven of the eight mean correlations calculated for the mixed-sex group of subjects achieved levels of practical significance (e.g., masculinity, behavior problems, care and living arrangements), while two of the four mean correlations computed for the male-only subject group attained practical significance (e.g., sex of the subjects, relations with ex-spouse). Finally, three mean correlations calculated for the female-only sample also reached practical significance (parent-child relationships, sex of the subjects, relations with ex-spouse). The above statistics represent 1647 individual effects and correlations and over 30,000 subjects. Interpretation of these findings suggests that, overall, children of divorce are less well adjusted than children from intact, two-parent families.