FitResponseStyles
FitResponseStyles is a repository containing modeling code using the R package TAM to fit Divide-by-Total modeling approaches accounting for response styles. Response styles, such as extreme, acquiescence, or mid response styles, are a potential source of bias when inferences are drawn from rating scale measurement.
Usage
To use the modeling code, download and unzip the folder and open the RStudio Project file. Using devtools::load_all(), you load the contents of the project.
In the data folder, you find a simulated data set with the structure of the German Big Five questionnaire (NEO-FFI; 5 dimensions, 12 items per dimensions, some reversed coded items). With data("dat") and data("which_rev"), you can load the data and a vector containing the item numbers of reversed coded items in the generated data.
You can find a main R file "analysis/FitModels.R" that fits the single modeling files in the folder "analysis/fit_models/" using the R package TAM (Robitzsch, Kiefer, & Wu, 2017) and saves them in the folder "analysis/fitted/".
Please adapt the modeling code to match your own data (number of items, number of dimensions, etc.).
The models for which code is provided here are:
Models based on the response style literature (variants of Divide-by-total approaches)
- Partial Credit Model (ignoring response styles, e.g., Masters 1982)
- Random Threshold Model (Wang et al., 2006)
- Generalized Random Threshold Model (adapted from Wang & Wu, 2011)
- Multidimensional NRM with estimated scoring weights for response styles (Bolt & Johnson, 2009)
- Multidimensional PCM with category preferences parameters (Bolt et al., 2014)
- Multidimensional PCM with fixed scoring weights for response styles (Bolt & Newton, 2011; Wetzel & Carstensen, 2017)
- Generalized Multidimensional PCM with fixed scoring weights and estimated discrimination parameters (Falk & Cai, 2016)
Modeling extensions proposed:
- A model lifting equality constraints on scoring weights of ARS
- A model putting equality restrictions on discrimination parameters based on dummy-coded item attributes
- A model putting equality restrictions on discrimination parameters randomly
see also:
Henninger, M. & Meiser, T. (2019). Different approaches to modeling response styles in Divide-by-Total IRT models: A model integration with extensions. Manuscript in preparation.
References:
Bolt, D. M., & Johnson, T. R. (2009). Addressing score bias and differential item functioning due to individual differences in response style. Applied Psychological Measurement, 33, 335–352. https://doi.org/10.1177/0146621608329891
Bolt, D. M., Lu, Y., & Kim, J.-S. (2014). Measurement and control of response styles using anchoring vignettes: A model-based approach. Psychological Methods, 19, 528–541. https://doi.org/10.1037/met0000016
Bolt, D. M., & Newton, J. R. (2011). Multiscale measurement of extreme response style. Educational and Psychological Measurement, 71(5), 814–833. https://doi.org/10.1177/0013164410388411
Falk, C. F., & Cai, L. (2016). A flexible full-information approach to the modeling of response styles. Psychological Methods, 21, 328–347. https://doi.org/10.1037/met0000059
Masters, G. N. (1982). A Rasch model for partial credit scoring. Psychometrika, 47, 149–174. https://doi.org/10.1007/BF02296272
Robitzsch, A., Kiefer, T., & Wu, M. (2017). TAM: Test analysis modules. R package version 2.8-21. https://CRAN.R-project.org/package=TAM
Wang, W.-C., Wilson, M., & Shih, C.-L. (2006). Modeling randomness in judging rating scales with a random-effects rating scale model. Journal of Educational Measurement, 43, 335–353. https://doi.org/10.1111/j.1745-3984.2006.00020.x
Wang, W.-C., & Wu, S.-L. (2011). The random-effect generalized rating scale model. Journal of Educational Measurement, 48, 441–456. https://doi.org/10.1111/j.1745-3984.2011.00154.x
Wetzel, E., & Carstensen, C. H. (2017). Multidimensional modeling of traits and response styles. European Journal of Psychological Assessment, 33, 352–364. https://doi.org/10.1027/1015-5759/a000291