libirt

The libirt software is a C-based library that offers a variety of functions to estimate items and abilities based on the responses of individuals to questionnaires. While there are two command line programs included, the focus is on integrating the library into other applications.

The library includes several models to choose from, such as the One Parameter Logistic Model (1PLM), which estimates the location parameter of each item and allows the user to fix the scale parameter; the Two Parameter Logistic Model (2PLM), which estimates both the location and scale parameters of each item; and the Three Parameter Logistic Model (3PLM), which estimates the location, scale, and guessing parameters of each item. There are also models available for nominal and graded response options.

In addition to these parametric models, libirt offers nonparametric estimators such as kernel smoothing and penalized marginal maximum likelihood estimation. Users can choose between marginal maximum likelihood estimation and Bayes modal estimation for parametric models, and expected a posteriori or maximum likelihood for ability estimators (limited to univariate logistic models).

Overall, the Item Response Theory Library provides a comprehensive set of tools for estimating items and abilities based on questionnaire responses, making it a valuable resource for software developers and researchers alike.