Depressive symptoms and hopelessness are central to suicidal states and various psychiatric conditions. The Beck Depression Inventory-II and Beck Hopelessness Scale are widely used measures in clinical and research contexts to evaluate the severity of these conditions. Although both scales demonstrate strong psychometric properties for assessing between-person differences, their sensitivity to within-person variation across repeated measures has not been evaluated in psychiatric populations. Generalizability theory provides a framework to partition variance from multitime-point assessments into meaningful sources (i.e., person, time, item, and their interactions), enabling the estimation of two complementary reliability coefficients that we evaluated in the present study: reliability of individual differences in change (Rc), with fixed items and time, and reliability of within-person variation averaged across items, with time nested within persons (Rcn). Up to four waves of data across 1 year were collected from 157 treatment-seeking veterans at risk of suicide enrolled in a suicide safety planning randomized clinical trial. Variance decomposition analysis was conducted via crossed and nested multilevel models. Individual differences in change accounted for 15% of the variance for both measures (crossed design, N = 87), and within-person variation accounted for 17%-19% (nested design, N = 157). Rc and Rcn coefficients exceeded .80. These findings indicate that the Beck Depression Inventory-II and Beck Hopelessness Scale reliably capture individual differences in change and within-person variation of moderate magnitude in depressive and hopelessness symptoms. Future work should examine their sensitivity to change and variation across different contexts, timescales, and populations. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Glatt et al. (Mon,) studied this question.