The use of the two- logistic model in building and grading the multi-dimensional virtual personality scale for university students according to the theory of item response

Abstract

      The current study aims to use the two-parameters logistic model in grading the multi-dimensional virtual personality scale for university students according to the item of theory response, and to come out with a generalization of the results of the research and concentration on the appropriate scientific procedures and provide a correct scientific practical application for approval by researchers.The researcher followed the scientific method of steps and procedures in the process of building the scale, where the dimensions of the hypothetical personality were limited to the dimensions (psychological, social, cultural, political, religious, emotional, and recreational) . The scale was presented with its seven dimensions and items to the arbitrators specialized in educational and psychological sciences.  Agreement between the experts on the apparent validity of the items, and the adoption of equality in the number of items for each dimension of the arbitrators agreement.  After the experts' agreement, the scale became at a rate of (13) items for each of the seven dimensions and in the manner of the declarative statements with five alternatives to the answer, so that the scale consisted of (91) test items, and after the initial application procedures, the scale became valid for application, and the scale was applied to a sample amounting to (915) male and female students from the universities of Iraq. The statistical analysis of the scale items was done using the statistical programs (STATA) for each of the seven dimensions, and after extracting the difficulty coefficients for the five thresholds for each item of the scale, 16 items of all dimensions were excluded, the calculated Z value of  it was less than the tabular value, then the statistical analysis was repeated for the virtual personal dimensions with its items (75), where the difficulty factor for each dimension ranged, respectively, between (2,590 to 2,116), (2,256 to 1,677), (089,10 to 1,251), (0,762 to 2,05), (2,105 to 0,719), (2,163 to 0,821), (0,811 to 1,016).  While discrimination coefficients ranged respectively (0,362 to 0,739), (0,289 to 0,841), (0,494 to 0,968), (0,381 to 0,976) (0,185 to 0,628), (0,264 to 0,901), (0,399 to 0,720), while the standard error of the seven dimensions ranged, respectively, (50,23 to 51,59), (50,18 to 51,96), (50,17 to 50,96), (50,22 to 51,21), (50,11 to 51,19), (50.46 to 51.45), (50.26 to 51.14). 

    The estimation of the power corresponding to the grade, its standard error, its information function, and the probability was extracted where the standard error was at score (98) high and at score (371) high as well, and this is due to the extreme power at this level is very weak and high at the middle.  It is also noticed that the information function was low at the periphery and improved at the middle, and the maximum value of the information function reached (32,79909) at the scores (287). The information function also reflects the stability of the item, which indicates the test suitability for students and is able to distinguish between them.  As it turns out, the thresholds are gradient in difficulty, and are almost equal in distinction, and guessing here has no clear effect.  Thus, the researcher verified all the assumptions of the item response theory.  The validity and reliability of the virtual personality scale were also verified.