PI(s): Lead – Jennifer Bogner, PhD (Ohio Regional TBI Model System, Columbus, OH); Site – Gale Whiteneck, PhD
Dates: 2013 - 2017
Contact(s): Gale Whiteneck firstname.lastname@example.org
This module will establish estimates of test-retest reliability for each of the items and instruments that comprise the Form II interview. For any measures found to have problematic reliability, recommendations will be made to the TBIMS Data Committee to investigate means to improve reliability. A consecutive or random sample of 200 cases will be asked (after completing their Form II interview) to complete a second Form II interview within the next 14 to 28 days, to be completed by the same interviewer; interviews requiring proxies will be excluded. After examining the distributions of each variable to assure assumptions are being met, the reliability of continuous measures will be estimated using intraclass correlation coefficients based on a two-way random effects model (both absolute agreement and consistency will be measured); the reliability of categorical variables will be evaluated using Cohen’s kappa (weighted kappa or ICC with ordinal variables).