SPIRIT Macro

Item Response Theory in SPSS

About SPIRIT

The SPIRIT macro allows researchers to conduct one-parameter item response theory (IRT) analyses through the typical SPSS syntax interface or a custom point-and-click dialogue box. It is available and easily implemented in SPSS version 21 and above. Since a one-parameter IRT model is a specific case of a generalized linear mixed model (GLMM), the macro utilizes the GENLINMIXED function of SPSS.

One of the main advantages of SPIRIT is its flexibility in terms of model specification. Person and item covariates, multidimensional models, and multi-group designs can be specified to allow users to move beyond basic measurement models into the realm of explanatory IRT models. Further, SPIRIT allows the item responses to be dichotomous or ordinal, which is a feature not included in some other general purpose GLMM software packages (such as the lme4 package for R) that are often utilized when analyzing item response data.

In addition to its flexibility, SPIRIT is intuitive and easy to use. The dialogue box allows the user to specify the model through a point-and-click interface and the syntax version requires a minimal amount of code. Concurrently, the user may request a series of output options that include estimated trait levels for each individual, item and person fit statistics, and item plots. These options are specific to IRT and, thus, not usually provided by GLMM software. 

Download

Version 1 of SPIRIT can be downloaded using the link above. The .zip file contains the syntax and point-and-click versions of the macro for SPSS Versions 23 and below and SPSS Version 24. In addition, the User Guide is provided, as well as the dataset used in the example analyses at the end of the User Guide. Please make sure to view the README file prior to using SPIRIT. Without following the directions in the file, the macro will not work properly. Do not hesitate to email me (rockwood.19@osu.edu) or Jack DiTrapani (ditrapani.4@osu.edu) for any questions regarding the macro, or to report any potential bugs. 

Cite

DiTrapani, J., Rockwood, N. J., & Jeon, M. (accepted). IRT analysis using the SPIRIT macro in SPSS. Applied Psychological Measurement [Computer program exchange]

A screenshot of the SPIRIT dialogue box.