Graduation Year

2013

Document Type

Dissertation

Degree

Ph.D.

Degree Granting Department

Educational Measurement and Research

Major Professor

Robert F. Dedrick

Co-Major Professor

Yi-hsin Chen

Keywords

Cross-cultural Measurement, Faculty Development, Higher Education, Invariance Testing, Teaching beliefs, Translations and Adaptations of Research Instruments

Abstract

As a result of growing attention in cross-cultural research, existing measurement instruments developed in one language are being translated and adapted for use in other languages and cultural contexts. The benefits of having the same instrument across cultures can only be realized if the process of translation and adaptation of the measurement instruments produces measurement operations that function similarly across national and cultural boundaries. Producing invariant measurement instruments that assess educational and psychological constructs provide a way of testing the cross-cultural generality of theories that include these constructs.

The major purposes of the study were to translate and adapt the Teaching Perspectives Inventory (Pratt, 1992, 1990) from English to Bahasa Malaysia and compare the psychometric properties of the two versions. The TPI is an instrument developed by Pratt (1992) to ascertain the different conceptions that teachers in higher education have about teaching. The TPI has 45 items, which are divided into five subscales or perspectives referred to as Transmission, Apprenticeship, Developmental, Nurturing, and Social Reform. The first phase of this study translated and adapted the TPI from English into the Malay language of Malaysia or Bahasa Malaysia (BM) using multiple approaches as recommended by the International Test Commission. The approaches used to translate the TPI included forward and back translations, an expert panel review, a pilot study, and cognitive interviews. In the translation process, three initial translators, two back translators, and six expert panel members, including the researcher, came up with a pre-final version of the Malay TPI. During the translation process, two items were found to contain expressions that had no exact equivalent forms in Malay: "virtuoso performers" and "higher ideals." Overall, translating the TPI was a challenging task due to the relatively large number of items in the instrument (45) as well as the complexity and very abstract nature of the constructs. Many of the words and expressions that were brief and concise in the English version became longer and more verbose when translated in Malay. As a result, the translated TPI version appeared longer than the original version. Pilot testing with 25 native speakers of Malay who were faculty members from a number of public universities in Malaysia revealed nine items that needed modification. Cognitive interviewing with five participants from the pilot group revealed one item requiring a change by adding a borrowed word "novis" in brackets next to the Malay expressions, which refers to the original word novice. Due to the confusion with the words referring to `people' in many of the items, additional instructions were added at the beginning of the survey to ensure that the participants responded according to the original intention of the items, which focuses on learners in the faculty's specific classroom context instead of people in the society in general. Following changes to the TPI, this instrument was administered in phase two to a Malaysian sample of 561 faculty.

In the second phase, the study assessed the psychometric properties of the original English version of the TPI with 605 faculty in the U. S. and the translated TPI version of the TPI with the Malaysian sample. The overall internal consistency reliability of both the English (α=.88) and the Malay TPI (α=.93) appeared to be adequate. At the subscale level, the internal consistency reliabilities of all the scales were on the lower side considering the large number of items (9) for each subscale (range = .67 to .83 for the U. S. and .59 to .81 for Malaysia). It was found that three out of the five subscales of the U. S. and Malay TPI had similar alpha reliabilities (Apprenticeship, Nurturing, Social Reform). To assess the cross-cultural factorial validity and measurement invariance of the TPI, confirmatory factor analysis (CFA) was carried out for both the original and the Malay TPI. The sample size for the U. S. group was 605 and the Malay group was 561. The fit for both the U. S. and the Malay correlated five-factor models was less than adequate with the Malay model showing a much worse fit. Correlated errors were found between 64 item pairs in the U. S. model and 389 item pairs in the Malay model. The correlations between the five perspectives in the Malay sample were much higher than those in the U. S. sample suggesting that the perspectives had limited discriminant validity. For example, the correlations between the Nurturing and Developmental perspectives and Nurturing and Social Reform perspectives were 1.0. The inadequate fit of the five-factor correlated model in the Malaysian sample and the minimally acceptable fit in the U. S. sample led to the decision to carry out analyses and compare the groups one subscale at a time. Model modifications for each subscale of both samples were carried out to improve the fit by adding one or more parameters (i.e., correlated errors) for each subscale model to obtain acceptable baseline models. The results of the invariance testing for each subscale did not support the existence of measurement invariance. Overall, the results indicate that the Malay version of the TPI is not ready for use and additional translation and adaptation work is recommended. Future efforts could incorporate improvements in the translation process in the form of recruiting a larger number of certified translators who have in-depth knowledge of teaching in higher education as well as a deep knowledge of the philosophy and purposes behind the TPI. Additional cognitive interviews before and after pretesting and pilot testing of the pre-final version are recommended. Finally, adding a large sample of bilingual educators who would complete both the Malay and English versions of the TPI would provide important psychometric data on the equivalence of the TPI items.

Download

Share

COinS