Estimation of Item Parameter Indices of NECO Mathematics Multiple Choice Test Items among Nigerian Students

Abstract

The study estimated the difficulty, discrimination, and vulnerability to guessing 2016 National Examination Council (NECO) Mathematics multiple-choice test items. The study employed an ex-post facto design with 276,338 samples. The research instruments used for the study were Optical Marks Record Sheets for the NECO June/July 2016 Mathematics objectives items. The responses of the testees were scored dichotomously. Data collected were calibrated using four parameters logistic model. The results showed that most items in the 2016 NECO Mathematics test were good as their difficulty parameters were within (-2 to 2). For those items, difficulty parameter estimates are considered good. Also, the results indicated that only 21.7% of 2016 NECO Mathematics test items had a very good discriminating power, and the majority of the items had poor discrimination power. The result implies that most of the items were not effective in discriminating between examinees with the required ability and those that lack the required ability. Finally, the result revealed that the 2016 NECO test items were not vulnerable to guessing (i.e., 86.7% of items were good in terms of guessing). The study concluded that the 2016 NECO Mathematics test items were good for difficulty and guessing parameter indices.

Keywords

• item response parameters
• item response theory
• one-parameter logistic
• two-parameter logistic, three-parameter logistic
• item discriminating
• item difficulty
• guessing parameter