Paul, Patricia2022-01-182022-01-1819901549https://hdl.handle.net/2139/53648Scores of 4,950 candidates, randomly selected from a total of 49,843 14 to 15-year-old junior secondary school leavers, from 30 schools in Trinidad and Tobago, were examined to determine psychometric properties and technical characteristics of 14+ mathematics examinations during 1985-1988. Examination papers for the period were also analysed. Results showed that: 1) content validity of the examination papers was fairly satisfactory; 2) inter-item reliability was high, ranging from .71 to .87; 3) less than 50 percent of the school leavers were successful in the examination in any one year; 4) item difficulty ranged from 11 percent to 78 percent; 5) discrimination indices ranged from .24 to .35; 6) there were better performances in the arithmetic and algebraic based items on the papers than on the geometry and statistics based items; 7) problems presented mainly numerically and symbolically were more often answered correctly than those with verbal content; 8) there were negligible differences in students' scores on the "traditional" and "modern" mathematics items; and 9) the scores on items of cognitive levels--knowledge, comprehension, application, and analysis--were similarSecondary school leaversM.Ed.