Comparion of mobile-based and paper-based tests for secondary school students

The present study aimed at comparing the psychometric properties of the tests and student performances according to gender by applying three different achievement tests in mobile (online) environment and by traditional method. This study adopted descriptive research design. The research group consists of 100 students (Female=50, Male=50) studying in an Anatolian high school in Gaziantep province. Computer science, biology and foreign language (German) 20-item multiple-choice achievement tests were applied to the research group in mobile and paper-based form. Socrative (student) was chosen as a mobile application because of its advantages such as being free of charge and easy and quick installation. Exploratory Factor Analysis (AFA) was applied to reveal evidence of construct validity of the tests. The Kaiser-Meyer Olkin (KMO) value was calculated to test the adequacy of the sample size before the analysis was performed, Bartlett test was performed to meet the assumption of multivariate normality, and central tendency measures and skewness and kurtosis coefficients were calculated for the normality of the distribution. As the data were categorical (1-0), exploratory factor analysis was performed on the tetrachoric correlation matrix. Cronbach Alpha internal consistency coefficient, KR20 internal consistency coefficient, Spearman-Brown coefficient, and Guttman two-half reliability coefficients were found. No significant difference was found between the reliability coefficients between the forms. The t-test was used to compare the scores on the related measurements; independent t-test and non-parametric Man-Whitney-U test were used for gender comparisons. For the decision whether the distribution deviates from the normal distribution or not; normality test, histogram graph, skewness and kurtosis values, mean, mode and median values were examined. Pearson product-moments multiplication correlation coefficient was calculated because the scores of the participants' relationship between the scores obtained in Computer Science and foreign language tests on mobile and paper-pencil applications showed normal distribution pattern. Since the distribution of the data on the relationship between the scores obtained in the mobile and paper-based applications of the biology achievement test deviated from normal, Spearman's rank-order correlation coefficient was calculated. As a result of the analysis conducted for the reliability relationship between the forms, the correlation numbers of all three achievement tests were found to be high. There was no significant difference between the achievement scores in the Computer Science, Biology and Foreign Language (German) achievement test among the participants in terms of gender in mobile-supported achievement tests. While there was no difference between the achievement scores for the Computer Science achievement test applied in paper-based form, there was a difference in favour of girls in the Biology and Foreign Language (German) achievement test.