Chinyere C. OGUOMA, Michael Akinsola METIBEMU and Romy OKOYE (2016)
AN ASSESSMENT OF THE DIMENSIONALITY OF 2014 WEST AFRICAN SECONDARY SCHOOL CERTIFICATE EXAMINATION MATHEMATICS OBJECTIVE TEST SCORES IN IMO STATE, NIGERIA
Mathematics tests used in the assessment of students’ proficiency in mathematics are inherently multidimensional. However, the procedure being adopted by public examining bodies in the scoring of examinees’ performance in the subject is based on Classical Test Theory (CTT), a measurement model that is limited to assessing tests that are unidimensional. This may be one of the reasons why students are consistently performing poorly in Mathematics, as research has shown that assessing multidimensional tests using unidimensional test models compromises test performance. Hence, research in Mathematics education should shift focus from the dominant students, teachers, school related factors and Mathematics achievement towards scoring procedures being adopted by public examining bodies. Therefore, this study assessed the appropriateness of scoring 2014 WASSCE Mathematics multiple choice test using CTT. To achieve this, the dimensionality of the test was assessed. The study adopted causal comparative design. All the year three senior school students (SSS3) of the 274 public Senior Secondary Schools in Imo State that registered candidates for the May/June 2015 Senior School Certificate Examination (SSCE) formed the population for the study. The sample for the study comprised 1142 (SS3) students of 30 schools which were randomly selected from the 274 public senior secondary schools in the State. Data was analyzed using Stout’s Test of Essential Unidimensionality, Bootstrap Modified Parallel Analysis test (BMPAT) and full information factor analysis. Results showed that the test and its items violate unidimensionality assumption (Stout’s Test rejected the assumption that the test was unidimensional, T = 7.3799, p < 0.001, one tailed; BPMAT showed that the second eigenvalues of the real was significantly greater than that of the simulated data set, p=0.0099). Furthermore, full information factor analysis showed that five dimensions underlie the test; the test items revealed within-item multidimensionality. The findings suggest that 2014 WASSCE Mathematics test was multidimensional. It is therefore our conclusion that CTT scoring was not suitable for the measurement of students’ performance in 2014 WASSCE Mathematics test. Hence, it is our recommendation that the use of CTT in the estimation of students’ test performance in SSCE Mathematics tests be stopped.
Chinyere C. OGUOMA, Michael Akinsola METIBEMU and Romy OKOYE (2016)
AN ASSESSMENT OF THE DIMENSIONALITY OF 2014 WEST AFRICAN SECONDARY SCHOOL CERTIFICATE EXAMINATION MATHEMATICS OBJECTIVE TEST SCORES IN IMO STATE, NIGERIA
Mathematics tests used in the assessment of students’ proficiency in mathematics are inherently multidimensional. However, the procedure being adopted by public examining bodies in the scoring of examinees’ performance in the subject is based on Classical Test Theory (CTT), a measurement model that is limited to assessing tests that are unidimensional. This may be one of the reasons why students are consistently performing poorly in Mathematics, as research has shown that assessing multidimensional tests using unidimensional test models compromises test performance. Hence, research in Mathematics education should shift focus from the dominant students, teachers, school related factors and Mathematics achievement towards scoring procedures being adopted by public examining bodies. Therefore, this study assessed the appropriateness of scoring 2014 WASSCE Mathematics multiple choice test using CTT. To achieve this, the dimensionality of the test was assessed. The study adopted causal comparative design. All the year three senior school students (SSS3) of the 274 public Senior Secondary Schools in Imo State that registered candidates for the May/June 2015 Senior School Certificate Examination (SSCE) formed the population for the study. The sample for the study comprised 1142 (SS3) students of 30 schools which were randomly selected from the 274 public senior secondary schools in the State. Data was analyzed using Stout’s Test of Essential Unidimensionality, Bootstrap Modified Parallel Analysis test (BMPAT) and full information factor analysis. Results showed that the test and its items violate unidimensionality assumption (Stout’s Test rejected the assumption that the test was unidimensional, T = 7.3799, p < 0.001, one tailed; BPMAT showed that the second eigenvalues of the real was significantly greater than that of the simulated data set, p=0.0099). Furthermore, full information factor analysis showed that five dimensions underlie the test; the test items revealed within-item multidimensionality. The findings suggest that 2014 WASSCE Mathematics test was multidimensional. It is therefore our conclusion that CTT scoring was not suitable for the measurement of students’ performance in 2014 WASSCE Mathematics test. Hence, it is our recommendation that the use of CTT in the estimation of students’ test performance in SSCE Mathematics tests be stopped.