Replication Data for: Idiographic Item Response Theory: Modeling Person-Specific Differential Item Functioning in Intensive Longitudinal Data
Year: 2026
Type: dataset
Abstract: This dataset contains the simulation code, empirical analysis scripts, and replication materials for "Idiographic Item Response Theory: Modeling Person-Specific Differential Item Functioning in Intensive Longitudinal Data" (Gilbert, Kim, Himmelsbach, Ulitzsch, & Zhang). The study proposes a Bayesian two-parameter logistic explanatory item response model that allows item location parameters to vary by individual respondent—idiographic differential item functioning (DIF)—using a spike-and-slab likelihood mixture to identify items with meaningful person-specific heterogeneity, a method applicable to intensive longitudinal, experience-sampling, and ecological momentary assessment data. The materials include: (1) Stan/R code to generate and fit the model across a fully-crossed simulation design (360 conditions × 100 replications, varying sample size, number of occasions, number of DIF items, and DIF magnitude) evaluating convergence, parameter recovery, false positive rates, and power; and (2) analysis code for an empirical application to repeated-measures depression-symptom data (9 dichotomous DSM-5 items in a psychotherapy RCT).
Source: Harvard Dataverse
DOI: 10.7910/dvn/g68c5y
Author: Gilbert, Josh
Institution: Harvard University
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7213331259