BRPM.DIF: BRPM.DIF
In changxiulee/BayesianRasch: Bayesian Rasch Modeling

This function assess the differential item functioning in your data Note: 1) When your data has only two categories, the polytomous Rasch model will fall back to the dichotomous Rasch model. Therefore, both choice of model is equivalent 2) To ensure the PCM model is identifiable, ability is assumed to be normally distributed. (theta ~ N(0, 1)) 3) To ensure the RSM model is identifiable, ability is assumed to be normally distributed. (theta ~ N(0, 1)) and My intention is to make the andrich threshold follow the sum-to-zero constraint. However, I could not find a way to do it in JAGS. Therefore, the first andrich threshold is assumed to be 0 in JAGS. Then, the andrich threshold and beta is transformed to allow the andrich threshold to follow the sum-to-zero constraint. 4) I follow Winsteps anchor theta method. The first analysis is to get the mean theta for each group. The mean is then used to anchor the theta in the next analysis

1 2	BRPM.DIF(data, item, DIF.var, n.chains, model = "BRSM", g = c(1, 2), ROPE = c(-0.5, 0.5))

`data`	A data frame or matrix
`item`	Item to be included in your model
`DIF.var`	grouping variable. Group should be in numeric
`n.chains`	Number of chains to be included in MCMC
`model`	Choose between 'BRSM' (Bayesian Rating Scale Model) and 'BPCM' (Bayesian Partial Credit Model).
`g`	group to be include in analysis
`ROPE`	region of practical equivalence. DIF for each item won't be equal to 0. But not all non-zero DIF is practical significant. Set a region that has practically insignificant DIF

Reference Linacre, J. M., & Wright, B. D. (2000). Winsteps. URL: http://www. winsteps. com/index.html. Soares, T. M., Gonçalves, F. B., & Gamerman, D. (2009). An integrated Bayesian model for DIF analysis. Journal of Educational and Behavioral Statistics, 34(3), 348-377.

changxiulee/BayesianRasch documentation built on Nov. 18, 2019, 6:54 a.m.