Item Fit Statistics (Outfit and Infit)

Outfit = mean of standardized squared residuals (sensitive to outliers). Infit = information-weighted mean-square (sensitive to systematic misfit).

Usage

item_fit(object)

Arguments

object

A fitted model from rasch_fit(), twopl_fit(), or threepl_fit().

Value

A data frame with columns: item, obs_prop, exp_prop, outfit, infit.

Details

Both should be near 1.0: Values greater than 1.3 means more noise than the model expects Values less than 0.7 means less noise than expected (items too predictable)