chapter  6
24 Pages

Generalizability Theory: Picking Up Where the Rasch IRT Model Leaves Off?

If Xpi is used to denote the observed score for any person (p) on any item (0, the universe score (a person's average score over the entire item universe) is defined as the expected value of a person's observed score across items or simply IIp == E Xpi' In a similar manner, the population mean for item

Similarly, the variance component for items is:

cr2 Xp; = E E (Xp; - 11)2 = crJ, + cr7 + crJ,i,e p ;

~ cr~ crT cr~ie 1.19 1.31 1.06 62 crx=-+-+--=--+--+--= . np ni npni 5 4 20

'" 2_~_ 1.19 - 82 ,-,po - 2 2 - -.