DOI: 10.36347/sjahss.2019.v07i09.006

Pages: 652-658

Recently developed testing theories and current technology make possible to develop a standard assessment and evaluation system for monitoring educational quality. In fact, this is achieved if the assessment system is supported by calibrated test item bank. However, its use is very limited in Haramaya University and assessment system has not been supported by test item bank due to the problems attributed to limited understanding of psychometric concepts for technical application of testing technology. This study assessed the practical application of 1-parameter item response model. The applications of this model determine the extent to which our instrument fits the standard Rasch model is measuring standards abilities. It is also used to calculate item difficulty and student’s ability parameters of common standards scale of measurements. This Study adopted introduction to psychology final examination consisting 40 multiple-choice test items. Responses for the tests were taken from 150 first year Haramaya university students’ of 2018/19 fresh entry students. The score result data were entered into Winstepe control data setup, as per the response category codes (zero for wrong and one for right) and then converted to a text file. The data pattern had 50 columns: columns 1-10 were for the ID; columns 11-50 were answers for 40 test items. The data were analyzed using the (Winstepe Rasch-Model John M. Linacre 2014) Computer Programs and all the item difficulties were linked on the logit scale along with the student ability. Thirty –four test items fitted the measurement model with probability p>0.03. ]. The item-trait interaction was not statistically significant at the 0.01 level, [Chi-square (df =1036 =1129.49, p =0.022]. indicating that dominant trait was measured and the test fits the standard item repose model, the item difficulties range from – 0.53. logits (SE= 0.30) to + 3.2 logits (SE=0.34) and the student ability measures range from – 1.11 logits to +4.0 logi