INFLUENCE OF PRIMACY, MIDDLE AND RECENCY BIASES ON THE PSYCHOMETRIC CHARACTERISTICS OF THE NATIONAL EXAMINATIONS COUNCIL MULTIPLE-CHOICE MATHEMATICS ITEMS

Abstract

The study determined the difficulty and discrimination of National Examination Councils (NECO) Multiple-choice Mathematics items from 2014 to 2016. The study also investigated influence of primacy, middle and recency biases on the psychometric characteristics (difficulty and discrimination) of NECO Multiple-choice Mathematics items from 2014 to 2016. These were with a view to providing statistical evidence on how the three biases consistently affect item difficulty and discrimination of NECO Multiple-choice Mathematics items across different population sets within the period of three consecutive years. The study adopted the ex-post facto research design. The population for the study comprised 53,986 Senior Secondary School Students (SS III) who registered and sat for the National Examinations Council Mathematics paper III from 2014 to 2016 in Osun state. A sample of 2,311 candidates was selected using multi-stage sampling procedure. Two Local Government Areas (LGAs) were selected from each of the three senatorial districts of the State, using simple random sampling techniques. From each of the selected LGAs, two schools were selected using simple random sampling technique. An intact class of senior secondary students (SS III) was used from each of the 12 selected schools. The instrument used for the study are the NECO Mathematics questions for the three years. Students’ corresponding responses as contained in the Optical Mark Reader (OMR) were scored dichotomously using Keys collected from NECO and calibrated under R software to generate difficulty and discrimination indices. The results of the study indicated that the item difficulty and discrimination parameters of Multiple-choice Mathematics items across the three years. For year 2014: 52(86.7) and 43(71.7) fell under moderate difficulty and discrimination respectively; for year 2015, 57(95.0) and 34(56.7) fell under moderate difficulty and discrimination respectively; lastly for year 2016, 42(70.0) and 12(20.7) fell under moderate difficulty and discrimination respectively. Finally, the result also showed that the primacy, middle and recency biases has no significant influence on difficulty and discrimination levels of National Examination Council (NECO) Mathematics paper III in the year2014 (For the difficulty; F_(2,57) .=0. 412; p>0.05 and For the discrimination; F_(2,57) .= 0.739; p>0.05) and 2016(For the difficulty; F_(2,57) .= 0.023; p>0.05 and For the discrimination; F_(2,57) .= 0.275; p>0.05). And the result further showed that the primacy, middle and recency biases also has no significant influence on difficulty level whilst there is a significant influence in terms of discrimination of NECO Multiple-choice Mathematics items in 2015(For the difficulty; F_(2,57) .= 0.539; p>0.05 and For the discrimination; F_(2,57) .= 6.409; p<0.05). It is concluded that difficulty and discrimination indices of the National Examination Council Multiple-choice Mathematics items are stable and not affected by primacy, middle and recency biases across the three years that were reviewed.