A Longitudinal Investigation of Dropout among Native and non-Native High School Students

Abstract

Most research on secondary school dropout has been cross-sectional, and this has made it difficult to know whether the differences observed between dropouts and persisters are antecedents or consequences of attrition. Furthermore, little is known about the causes of dropout among non-majority students such as native Indians. Investigators also have neglected to consider the utility of using theoretical frameworks that have proved useful in understanding various behaviours. This thesis was directed at learning which variables contribute to dropout among Native and nonNative students and at delineating combinations of variables that are useful predictors of enrollment status. The relative merits of the Ajzen-Fishbein model and a general expectancy-value approach also was of concern. To these ends, a longitudinal, multivariate investigation was undertaken. Information from school records and questionnaires was used to determine which factors distinguished students who had persisted from those who had discontinued 1 year and 3.7 years after the initial survey. The results revealed that dropout can be predicted with considerable accuracy even over several years. Intentions, absenteeism, grade average, and perceived value of education were shown to be the most important predictors. Information in school records enabled good prediction. Of the two models, the Ajzen-Fishbein framework showed the most promise. Although the data were consistent with Fishbein's conceptualization of how intentions are formed, it was shown that the model needs refinement in accounting for actual behaviour. On the basis of the findings from this and other research, suggestions are given for decreasing attrition.