Abstract:
The study compared the Traditional True Score Theory (TTST), and Modern Mental Test Theory
(MMTT) estimated item parameter indices to the ability of examinees in Junior Secondary School
Certificate Examination (JSSCE) in Mathematics. This is with the view of providing empirical
justification on the appropriateness of decisions made statistically from psychometric tests. The study
adopted a descriptive survey design. A sample of 600 students was randomly selected from a population
of 95,419 students who sat for the 2016 JSSCE Mathematics Paper 1 in Oyo State, Nigeria. An adoption
version of the 2016 Oyo State JSSCE Mathematics Paper 1 titled Mathematics Test (MT) instrument
was utilized for information aggregation. Data collected were analyzed using SPSS and BILOG-MG
software packages. The effects indicated that the difficulty indices of Mathematics test items based on
TTST ranged from 0.00 to 0.65. The discrimination indices ranged from 0.00 to 0.41. Fifty-two items
(86.67%) on the MT items had moderate item difficulty (0.200 ≤ p ≤ 0.620). On the other hand, 42
(70%) on the items discriminate poorly (0.10 ≤ D ≤ 0.29). Regarding MMTT results, item difficulty
parameter ranged between 0.216 and 7.988 for 2PLM while the discrimination parameter ranged
between 0.100 - 0.729 respectively. Furthermore, there was a negative relationship between the
difficulty indices of MMTT and TTST models (r = -0.702, p < .01). The concomitance correlation for
discrimination was positive (0.646, p <. 01). The study concluded that the Oyo State 2016 JSSC
Mathematics examination was of a moderate psychometric quality irrespective of the theoretical
measurement model used in the appraisal
