Print ISSN: 1017-5636
Online ISSN: 2412-558X
"Group impact and educational impact are the quintessence of social consequence. They drag measurement practice into that fourth quadrant". (Willingham, 2002, p. 196)
This paper expands upon concerns expressed earlier in De Lisle & Smith (2004) about the relationship between Eleven-Plus test design and patterns of gendered achievement in Trinidad and Tobago. Following Willingham's (1999), evaluation protocol, it includes (1) a critical analysis of gender fairness issues, (2) an empirical evaluation of gendered impact, and (3) a consideration of proposals for resolving gender fairness issues. Datasets used in the analysis and from the 2001-2003 Secondary Entrance Assessment (SEA) and the 1998-2000 Common Entrance Examination (CEE). The results confirmed that the gender gap was significantly greater for the SEA compared with the CEE, and that both males and females were disadvantaged in different ways by the placement system. The results also confirmed the existence of medium-sized gender differentials across urban-rural educational districts, literacy constructs, and high-low ability groups. A recent proposal to change the way the composite score is calculated did little to reduce the overall female advantage. Moreover, misclassification rates for the corrent remediation cutscore set at 30 percent were relatively high. These fairness issues are not easily resolved, but suggest the need for evidence-based test designs, test validation studies, and a re-examination of the need for selection.