Interpreting MNSQ and T fit indices

[Georg II]

Dear colleagues,
How should weighted and unweighted MNSQ and T fit indices be interpreted in ConQuest?
Are the conventional .70-1.30 for MNSQ and +-2 for T applicable?
How should the confidence intervals (CI) be used? Do they help in deciding on the fit of individual items or just help have a feel of the accuracy of the MNSQ indices?

Georg
Cheers

[Eveline Gebhardt]

Hi Georg

Yes, the (-)1.96 t-values are applicable and used to evaluate the fit of the item. The CI is the range between t-values -1.96 and 1.96. We generally use the t-values (unless the sample size is so large that almost all fit statistics are significantly different from 1) in conjunction with ICC plots and item analysis tables to decide on the fit of individual items.

The weighted MNSQ is the slope of the item characteritic curve. Less than one is steeper than expected, more than 1 is flatter than expected, 1 is expected (the average slope when using a one parameter model).

Eveline