Examining the Results of Different Models Used in Multitrait-Multimethod Analysis


Unsal Ozberk E. B., DOĞAN N.

JOURNAL OF MEASUREMENT AND EVALUATION IN EDUCATION AND PSYCHOLOGY-EPOD, cilt.5, sa.2, ss.43-61, 2014 (ESCI) identifier

Özet

In this reseach, method related and trait related specifications are compared in multitrait-multimethod confirmatory factor analysis. Based on persons' problem solving and crticital thinking abilities, a model related to creative thinking abilities was introduced. The model datas collected with self-report, questionnaire and friend report methods in 500 higher education students. Reliability and validity was interpreted in a multitrait-multimethod matrix by measuring the different traits using different methods. Later, correlated method, uncorrelated method and "minus-1" method in multitrait-multimethod confirmatory factor analysis are compared to decide which method works better to evaluate the models. The reliability diagonal values confirmed to be highest values of the matrices while collecting validity evidence in creative thinking ability. The validity coefficients are also different from zero. In validity diagonal, heterotrait-heteromethod triangle and heterotrait-monomethod triangle values is lower than validity diagonal values. Total scores, which was calculated by three diffrent method on problem solving and creative thinking abilities set as a structural model on creative thinking ability. Fit indices of creative thinking model were found acceptable. In multitrait-multimethod matrices, best model to assess creative thinking ability is the "minus-1" model.