A Consumer's Guide to LOGIST and BILOG

Abstract

Since its release in 1976, Wingersky, Barton, and Lord's (1982) LOGIST has been the most widely used computer program for estimating the parameters of the three-parameter logistic item response model. An al ternative program, Mislevy and Bock's (1983) BILOG, has recently become available. This paper compares the approaches taken by the two programs and offers some guidelines for choosing between the two pro grams for particular applications. Index terms: Bayesian estimation, BILOG, IRT estimation procedures, LOGIST, marginal maximum likelihood, maximum like lihood, three-parameter logistic model estimation pro cedures.

Empirical Estimation Errors in Item Response Theory as a Function of Test Properties*

“…In the present instance, LOGIST does not typically produce JML estimates of item and person parameters; BILOG does not typically produce MML estimates of item parameters and Bayesian estimates of abilities (Mislevy & Stocking, 1987). Rather, the procedures produce approximations that are considered to be 'good enough'.…”

“…A small investigation is made of the more recently developed marginal maximum likelihood approach incorporated in the computer program BILOG (Mis levy & Bock, 1983). These two programs incorporate very different approaches to the problem of obtaining estimates of item and Empirical Estimation Errors 5 person parameters (Mislevy & Stocking, 1987), but suffer some of the same deficiencies.…”

Untitled

“…This type of numerical integration is commonly used in calculus for finding the area under curved functions that are not easily integrated using integration rules from calculus (Larson & Edwards, 2014). The use of numerical integration is not completely foreign to psychometrics; the item-response theory software package BILOG uses numerical integration (Mislevy & Stocking, 1989). Thus, rather than implementing the formulaic integral shown above, Miwa et al's procedure implements numerical integration to estimate the volume of a multivariate-normal distribution.…”

Missing data as a validity threat for medical and healthcare education research: problems and solutions

“…Coding nonresponse observations as incorrect is palatable for free-response items, but less so for multiple choice items. [15] …”