out_infit function

Item Outfit and Infit Statistics

To avoid bias observed item responses are compared to expected responses under the conditional distribution of responses given the total score. This leads to standardized residuals which can be summarized to outfit and infit statistics in the usual way.

out_infit(
  object,
  se = TRUE,
  p.adj = c("BH", "holm", "hochberg", "hommel", "bonferroni", "BY", "none")
)

Arguments

• object: An object of class "Rm", a fitted Rasch model or partial credit model using the functions RM or PCM in package eRm, or an object of class "pcmodel", a fitted partial credit model using the function pcmodel in package psychotools.
• se: If TRUE the standard errors will be included.
• p.adj: Correction method for multiple testing. The methods are "BH","holm", "hochberg", "hommel", "bonferroni", "BY", "none". See p.adjust.

Returns

an object of class outfit containing: - outfit: outfit statistics

• outfit.se: standard errors of outfit statistics

• out.pvalue: p values of outfit statistics

• out.pvalue.adj: adjusted p values of outfit statistics if selected

• infit: infit statistics

• infit.se: standard errors of infit statistics

• in.pvalue: p values of infit statistics

• in.pvalue.adj: adjusted p values of infit statistics if selected

• padj: adjustment method

Details

The fit statistics and their standard errors are calculated as described in Christensen et al. P values are are based on the normal distribution of the standardized fit statistics.

Examples

rm.mod <- RM(amts[,4:13])
out_infit(rm.mod)

References

Christensen, K. B. , Kreiner, S. & Mesbah, M. (Eds.) Rasch Models in Health. Iste and Wiley (2013), pp. 86 - 90.

Kreiner, S. & Christensen, K. B. (2011) Exact evaluation of Bias in Rasch model residuals. Advances in Mathematics Research, 12, 19-40.

Author(s)

Marianne Mueller