Mixed methods investigated the cognitive processes reflected in retrospective pretest (thentest) discrepancy scores [i.e., recalibration response shift (RS)].
People with human immunodeficiency virus/acquired immune disease syndrome (HIV/AIDS) (n = 521) were interviewed at baseline and 6 months using the Quality of Life (QOL) Appraisal Profile, the Rand-36, General Health thentest, and recall items. Open-ended appraisal questions were coded, and factor analyses reduced the data. Ipsative (based on the then-minus-pretest) and normative (based on regression residuals) discrepancy scores were compared. Hypothesis testing related to recall bias and relationships among appraisal parameters and ipsative discrepancies, after covariate adjustment.
Coded frame of reference themes were distinct from experience sampling, standards of comparison, and combinatory algorithm. There was convergence between the ipsative and normative discrepancy scores (r = 0.30), but the former were associated with more appraisal changes and goal-related appraisals than the latter. Thentest effect sizes (ES) were larger than standard change scores, even controlling for recall bias. Multivariate models including appraisal parameters explained 9% more variance over the standard (unadjusted for RS) model.
Ipsative and normative discrepancy scores measure distinct constructs, represent different configurations of appraisal change, and are not invalidated or explained by recall bias. The thentest does not imply recalibration alone but rather a host of health- and self-care-related concerns.