Abstract
I consider two tests of receptive vocabulary administered, respectively, in the first and third wave of starting cohort 3 of the National Educational Panel Study (NEPS). The first test was administered in 2011 in a sample of children visiting a Kindergarten in that year; a second test took place two years later. I compare test results of 518 children who have participated in both tests. I first use 38 items which are identical
in both tests. Based on sum scores, one sees a remarkable increase in the
receptive vocabulary. I then show that the data are not compatible with
a joint Rasch model which assumes time-invariant item parameters. As
an alternative basis for longitudinal comparisons I suggest that a sufficient
condition for two tests measuring the same kind of competence is that they
consist of identical items.
Users
Please
log in to take part in the discussion (add own reviews or comments).