Abstract:
Researchers often import and adopt surveys from one cultural setting to
another in order to collect comparative data or to simplify the laborious process of
instrument development. Even when the instrument has been proven to have high
reliability in the original setting, the reliability may prove to be much weaker in the
new setting, especially when Western instruments are imported into non-Western
countries. In this chapter, we discuss the problems of importing an instrument from
one culture to another and associated methodological challenges. More importantly,
we present a detailed account of using structural equation modeling (SEM) and
MPlus software to validate a survey instrument imported to Ghana. The students’
Views of Mathematics (VOM) instrument is based on earlier Western research and
was further developed in Finland, where it had been validated to have high reliability.
First, we used confi rmatory factor analysis to test whether the seven factors identifi ed
in Finland were identifi able in Ghana. As the original factor structure was found not
to fi t the Ghanaian data, we continued with an explorative approach to identify the
Ghanaian factor structure, resulting in a four-factor structure. For cross- validation
purposes, the sample was randomly split into two, one-half of the sample assigned
as the calibration sample and the other half as the validation sample. Measurement
invariance was established at the confi gural, metric and structural levels between the
calibration and validation sample. We further discuss the measurement artifacts and
cultural differences as possible causes for the observed differences in the factor
structures between the Ghanaian and the Finnish sample .
