Item Quality and Gender-based Differential Functioning in A Geometry Diagnostic Test for Middle Basic Pupils in Plateau State, Nigeria
Résumé
This paper established The Item Quality and Gender-Based Differential Functioning of Geometry Diagnostic Test for Middle Basic Education Pupils in Plateau State, Nigeria. Currently there are no developed existing diagnostic tests available that Middle Basic Education teachers can use to find out the root causes of pupils learning misconception in geometry in Plateau State. Hence, the need to develop one for them that is qualitative and free of bias that is gender-based. To achieve these, the study sought to establish the item indices, item fit, test information function and differential item functioning using both the Classical test theory and the item response theory approaches. The study utilized instrumentation and survey research designs. The population was 3, 618 Middle Basic Education 5 pupils from Plateau Central Senatorial Zone. From this population a sample of 1445 pupils were drawn using clustered sampling method. The Probability Proportionate Sampling (PPS) method was used to sample 24 schools for the study. The instrument was a 50- Items Multiple Choice Geometry Diagnostic Test for Middle Basic Education Five Pupils (MCGDTMBEFP) developed by the researcher. The validity of the instrument was estimated through experts’ judgment, test blue prints, data model fit and differential item functioning while reliability was estimated using test information function. The study was guided by two research questions and one hypothesis. Data analyses were done using the Xcaliber software that employed the use of 3PLM for item indices, Chi-square goodness of fit, z-test and the Maximum Likelihood Estimate. The findings reveal that 64% of the items are within the acceptable difficulty level of 0.30-0.50 (moderately difficult), 74% items discriminated well because they have index value of 0.31 and above and 98% of the items are good because they have guessing index value of 0.25 and below. The chi- square goodness of fit statistics employed for analysis revealed that 100% of the diagnostic items fit the 3PLM. The test information function equalled 3.278 obtained through the use of the maximum likelihood estimate statistics indicated a high reliability of between 0.75 and 0.91. Findings also revealed that 0% or non of the items flagged off differential item functioning based on gender. It is concluded that a middle basic education diagnostic test that is of quality and fair to pupils that is gender-based has been developed to improve assessment practices and recommended among others that the Middle Basic Education mathematics teachers can make use of the instrument in identifying students with geometry learning misconceptions and hence help in organizing remedial teaching.