Results
In this section we report the results of Rasch analyses used to investigate four aspects of
validity described by Messick (1995) as well as carelessness and guessing. We also
examine interpretability as defined by the Medical Outcomes Trust Scientific Advisory
Committee (1995) and report on a qualitative investigation of test functioning with 22
participants.
Content aspect of construct validity
The content aspect of construct validity consists of three parts: content relevance,
representativeness, and technical quality (Messick, 1995). Content relevance, which
concerns the relevance of the test items to the targeted construct, was described in
subsection 2 (The source of target vocabulary) and subsection 3 (Test creation).
a Representativeness. Representativeness concerns the degree to which the test is sensitive
to variations in the measured construct (Borsboom et al., 2004). Representativeness can
be operationalized in terms of whether:
1. a sufficient number of items is included on the test;
2. the empirical item hierarchy shows sufficient spread; and
3. gaps exist in the empirical item hierarchy