Šepec, Tanja.
(2007).
Psychometric evaluation of three aptitude tests from DAT battery in selection procedure.
Diploma Thesis. Filozofski fakultet u Zagrebu, Department of Psychology.
[mentor Ljubotina, Damir].
Abstract
Interpretation of test results is most accurate when made with as many relevant characteristics of aimed population as possible. DAT battery is well known and commonly used in I/O psychology. Using Verbal, Numeric and Abstract reasoning tests, the goal of this research was to see if there are differences between the standard psychometric validation of tests based on Croatian high school population and the one gathered from young job candidates with college degree that go through selection procedure in one of Croatian large company. This also included making norms that best fit this group regarding gender and college (social and humanistic vs. nature science and technological) orientation. During the standard selection procedure within company, the tests were administrated to 541 participants. Result distribution is similar to normal, with tendency to upper values. Sensitivity of the tests is satisfactory, with the average difficulty index ranged from .52 to .61. Reliability of Numeric and Abstract reasoning tests is high - �± =.86, while lower on Verbal reasoning test (�± = .66). Factor analysis points to general factor in each test, but dichotomy of variables makes that conclusion hard, especially with Numeric and Abstract reasoning tests. The existence of second significant factor should be explored by additional item analysis. Results show decreased scores obtained by this group in comparation to the ones of high school students, with exception of Verbal reasoning test, where the difference has not been significant. In this research, men, as well as candidates with natural science and technological college orientation obtained higher scores on Numeric and Abstract reasoning tests. Significant interaction between gender and college orientation has not been found. Findings of this research should be treated as contribution to easier and more accurate evaluation of young, highly educated candidates in this specific company. They do not allow generalization to all assessments of verbal, numerical and abstract abilities, neither to whole population of highly educated people.
Actions (login required)
 |
View Item |
Statistics
Statistics