ghasem keshavarz gerami; Mohammad Reza Flasafi Nejad; jalil Younesi; ali delavar; zahra Rezaei Ghahroodi
Abstract
Response time has always been a concern in ability measurement, sometimes implicitly as in time-limit tests, sometimes explicitly as in speeded tests, and sometimes as the main outcome variable. Therefore, this study seeks to investigate the application of response time model in measuring cognitive ability ...
Read More
Response time has always been a concern in ability measurement, sometimes implicitly as in time-limit tests, sometimes explicitly as in speeded tests, and sometimes as the main outcome variable. Therefore, this study seeks to investigate the application of response time model in measuring cognitive ability of the children's quantitative reasoning. To this end, a sample of 502 children from different areas of Tehran was recruited through Cluster sampling and evaluated with a quantitative reasoning test. To analyze the collected data, modeling response time with the hierarchical model was used. Item and person parameters were also estimated and interpreted. The modeling results revealed that in addition to the response correctness parameters, response time parameters including discrimination coefficient, difficulty coefficient and gamma also been estimated. Moreover, the estimated parameters of the response time model compared with the 2-pl IRT model in order to identify their main similarities and differences. Findings revealed that there are some differences in discrimination coefficients and Thetas of individuals, which indicates the power of the response time model. It is worth mentioning that the lowest differences were observed in the difficulty coefficients. Finally, based on the number of parameters, increasing the variance of theta scores, and research background support, the study and application of response time models in measuring children's cognitive abilities were suggested.
Shoeayb Qasemi; ali delavar; noorali farrokhi; Farzad Eskandari
Abstract
This study was conducted with the aim of examining the gender differential item functioning of the English language items of the master psychology series exam using the methods available in the context of cognitive diagnostic models. The research sample included 2,455 female applicants and 919 male applicants ...
Read More
This study was conducted with the aim of examining the gender differential item functioning of the English language items of the master psychology series exam using the methods available in the context of cognitive diagnostic models. The research sample included 2,455 female applicants and 919 male applicants who attended the Master of Psychology series exam in 1396. The G-DINA cognitive diagnostic model was fitted with data. DIF of the questions were examined with three methods: the Wald test, the likelihood ratio test and the revised likelihood ratio test. The results of these three methods had a moderate agreement on DIF detection. Based on the results 16 questions had DIF. Most of the questions had uniform DIF in favor of men. Questions that had non-uniform DIF, for the dominant applicants were in favor of men and for non-dominated applicants were in favor of women. It is recommended that additional studies be conducted to investigate the existence of the bias and the cause of the DIF.