Perceived Control of Academic Events in Turkish and English Adolescents: Comparing Classical Test Theory and Item Response Theory to Detect Item Biases Across Cultures

Ertubey, Candan. 1999. Perceived Control of Academic Events in Turkish and English Adolescents: Comparing Classical Test Theory and Item Response Theory to Detect Item Biases Across Cultures. Doctoral thesis, Goldsmiths, University of London [Thesis]

[img]
Preview
Text (Perceived Control of Academic Events in Turkish and English Adolescents: Comparing Classical Test Theory and Item Response Theory to Detect Item Biases Across Cultures)
thesis_eturbey_1998.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (15MB) | Preview

Abstract or Description

Perceived control was investigated in secondary school adolescents from secular Turkish and British schools. Five published questionnaires were used. Perceived control was measured mainly with the Control beliefs, Means-Ends beliefs, and Agency Scale (CAMI - Skinner, Baltes, Chapman, 1988) As supplementary measures, Nowicki and Strickland's (1973) Internal-External Locus of Control for Children Scale (CNSIE) and Palenzuela's (1988) Multiple Academic Specific Locus of Control Scale (MASLOC) were also used. Two other questionnaires, measuring social antecedents, were included in the study. These were the Religiosity in Youth Scale (Rothbough & Jessor, 1976) and the F-Scale (Adorno, 1950; Christie, 1991).

In order to detect item biases within and between cultures, two procedures were followed: a) The questionnaires were translated from English to Turkish, and then both versions were completed by a bilingual Turkish sample. Translation fidelity was tested using three different approaches: Generalizability theory (Cronbach et al., 1972; Shavelson & Webb, 1991), Classical Item Analysis (Nunnally, 1978) and Item Response Theory (Lord, 1980; Hambleton et al., 1990). The results are discussed in terms of consistency and the cost of the procedure. b) The Turkish and English versions of the questionnaires were completed by unilingual Turkish and English samples in their own language. The purpose here was to detect similar (etic) and different (emic) functioning of the items across cultures using Classical and Modem item analysis. The questionnaires were found to be functioning similarly in the two cultures. But CNSIE was found to be a poor scale in both cultures and was not used in the next stage.

In the final part of the thesis perceived control was investigated as a function of culture, religion, religiosity, authoritarianism, age and gender. The perceived control scale showed significant cultural differences on 3 out of 10 subscales of CAM!. Agency beliefs and MASLOC showed significant effects of religion, religiosity, authoritarianism and gender. Generally, the Turkish sample was more Internal, more religious, and authoritarian. The gender differences were more inconsistent in terms of the direction of the relationship, with boys more internal on attribute while girls were more internal on unknown factors and effort.

Item Type:

Thesis (Doctoral)

Identification Number (DOI):

https://doi.org/10.25602/GOLD.00028904

Keywords:

perceived control, secondary school, adolescents, Turkey, Britain, f culture, religion, religiosity, authoritarianism, age, gender

Date:

1999

Item ID:

28904

Date Deposited:

30 Jun 2020 10:26

Last Modified:

30 Jun 2020 10:26

URI:

http://research.gold.ac.uk/id/eprint/28904

View statistics for this item...

Edit Record Edit Record (login required)