[Rasch] Test Translation & DI

Purya Baghaei puryabaghaei at gmail.com
Sat Jan 17 02:27:12 EST 2009

To check measurement invariance in multiple versions of a test either
scale-level analyses or item-level analyses are used (Zumbo, 2003). In
scale-level analyses confirmatory factor analysis (CFA) and
coefficient alpha are used and in tem-level analyses DIF is employed.
In scale-level analyses using CFA the measurement model specified in
one group is expected to be reproduced in the other with the same
magnitude of loadings and error variances. If this happens one can
conclude that the same construct is present in both tests and it is
measured equivalently in both versions. This is strict invariance.
However, if the same model is reproduced but with different loadings
and error variances, it is said that the same construct is present in
both tests but it is not measured equivalently (weak invariance).
Equal coefficient alpha is also another measure of invariance. Note
that Zumbo (2003) with a simulation study demonstrated that DIF does
not manifest itself in scale-level analyses even when 42% of the items
exhibited DIF. So item-level analyses are essential.
I think within Rasch measurement, principal component analysis of
residuals should be relevant too. Comparing the dimensionality of the
versions and the magnitude of the unexplained variances in the two
groups and the clustering of the items, whether they change or not and
whether their loadings on the 1st contrast change considerably or not
should be indicative signs.

On 1/15/09, Anthony James <luckyantonio2003 at yahoo.com> wrote:
> Dear all,
> I want to study DIF in the context of test translation. That is, whether
> items exhibits different levels of difficulty in the original version and
> the translated version of a test. How should one go about doing this in
> Winsteps?
> Two different linguistic groups have taken the original version of a test
> and its translation.
> Do I need some bilingual test-takers to take both versions to establish a
> link to run DIF?
> What else can be done, within classical test theory and Rasch measurement,
> to investigate translation equivalence and the validity of the translated
> version of a test?
> Cheers
> Anthony

More information about the Rasch mailing list