[Rasch] Missing Data in RUMM 2010

Eric Wong mcwong at cuhk.edu.hk
Mon Feb 4 20:33:43 EST 2008


Dear all,

How RUMM handle missing item data?

In addition, RUMM uses ANOVA to assess DIF. However, the method itself is 
sample size dependent (i.e. tend to detect statistically significant if the 
sample size is large); I wonder if there is any guideline in the sample 
size, and what is the size of DIF to be real significance?

Many thanks.

Eric




More information about the Rasch mailing list