Beneath Education Production Functions: The Case of Primary Education in Jamaica

Qualitative follow-up was used to add depth to an econometrically based investigation of home and school predictors of achievement in Jamaica (Glewwe, Grosh, Jacoby, & Lockheed, 1995). In the original investigation, theory and research were used to identify potential predictors. Data on these factors, as well as student achievement, were collected on a large, national sample that linked families with the schools their children attended. Although the original choice of relevant factors was grounded in theory and the methodological approach was sound, the findings from this "snapshot" were not always interpretable and were sometimes counterintuitive. For example, why should whole-class instruction be negatively related to achievement, or why were nurse visits negatively associated with achievement but vision screening positively associated with achievement? Researchers, suspecting there was more to the story, identified 4 of the sampled schools for qualitative follow-up. Two of the selected schools (1 urban and 1 rural) reported characteristics associated with higher levels of achievement, whereas the other 2 reported characteristics associated with lower levels of achievement. The purpose of the follow-up was to ascertain the comparability of the 4 schools on factors not assessed by the original investigation and to pursue the significance of the strongest predictor variables. Following extensive interviews with school administrators, teachers, and the board chairmen, the meaning and value of several of the variables were illuminated. For example, some variables were proxies: Nurse visit was an indicator of the absence of local health resources, whereas vision screening was associated with support that was being provided to the school by a service organization (a resource not otherwise measured in the original research). Another variable that was a significant negative predictor of achievement in the original study was the late arrival of textbooks. Although this variable may have been important at the time the original data were collected, its predictive value may not persist. Interviews with administrators and teachers indicated that in subsequent years resourceful school personnel anticipated the late arrival and each year held aside some books for use the following year.