Abstract
Clinical studies have shown that rumination functions as a mediator between overgeneral memory—the tendency to retrieve autobiographical memories in a non-specific format—and depression. Recently, rumination has been dismantled into two distinct subcomponents: reflection, which is more adaptive, and brooding, which is more maladaptive. In the present study we examined the differential relationships of these two rumination subcomponents with autobiographical memory specificity and their mediational role for the relationship between reduced memory specificity and depression in a non-clinical sample. In addition, we investigated the usefulness of a “minimal instructions” version of the Autobiographical Memory Test (AMT) to measure memory specificity in non-clinical populations. Results indicated that the use of minimal instructions can increase the AMT's sensitivity to detect reduced autobiographical memory specificity in non-clinical individuals. Further it was found that brooding, and not reflection, is significantly associated with reduced autobiographical memory specificity and functions as a mediator between reduced memory specificity and depression.
Notes
1The Dutch words were respectively: vertrouwen, bang, prettig, boos, moed, droevig, gerust, brutaal, verrast, and lomp. The Dutch practice items were gras and brood.
2The exact meaning and use of each code was extensively explained to the participants.
3As there was no self-report code to indicate semantic associates, the respondent's own code could not be used as a guideline in these cases.
4Both testing sessions contained other questionnaires, which were unrelated to this study.
5Hierarchical multiple regression analyses testing the significance of “instruction group” (TIG vs MIG) as a moderator for the relationship between reduced memory specificity and depression/rumination led to the same conclusion.
6However, one should be careful when interpreting these correlations. The categories of responses on the AMT (specific, extended, categoric, semantic associates, omissions, and rest) are not independent of each other (i.e., the proportions of responses in these categories add up to 1). Consequently, adding the semantic associates to the categoric index makes the correlations with this “overgenerality index” more similar to the opposite correlations with specific memories (as proportion of specific memories and proportion of categoric memories plus semantic associates might be relatively complementary).