Renato Schibeci

The major purpose of the study reported in this thesis was to investigate the influence of the home, school and peer group environments on high school student attitudes to science. The study was carried out in three main stages. First, a model was developed to guide the investigation. The model, which hypothesised interrelationships among the three educational environments (the home, the school and the peer group) and student attainment, was derived from a conceptual framework developed by Keeves (1972). Second, a number of preliminary investigations were conducted. These were designed to: (a) gather data on science teachers' perceptions of the objectives, especially the attitudinal objectives, of the high school science curriculum; (b) establish the reliability and validity of the affective instruments to be used in the final stage. Third, the model developed for the study was tested using the year 8 cohorts in two different high schools in the Perth metropolitan area. The generalisability of the model to the two different samples was assessed. A number of conclusions may be drawn from this investigation of the influence of educational environments on student outcomes: (1) The influence of background variables is generally not very strong. Typically, values for path coefficients from environmental variables to other variables are low. (2) There is a strong causal chain among the cognitive variables of the study: "general ability" directly influences "initial achievement", which in turn directly influences "final achievement". (3) There is a strong causal chain among the affective variables: "initial science-related attitude" directly influences "general attitude", which in turn directly influences "final science-related attitude". (4) There are links between these two sets of variables (the cognitive set of variables and the affective set of variables), but these links are not as strong as those within the sets. For example, "general ability" influences "initial science-related attitude", but the link is not as strong as those within each set. (5) There are links between the background (independent) variables and the dependent variables, which have path values which are of the same order of magnitude as the 'within' values in (4). The influence of background variables appears to be different for the two schools. This provides limited support for the notion that local influences should be further investigated. Home influences appeared, in general, not to influence science-related attitudes or achievement in any direct, substantial way. Keeves (1972) had arrived at a similar conclusion. However, one variable which appeared to be an important variable was the mother's expectations for the length of the child's secondary schooling; this appeared to be an important influence on the cognitive outcome variables in both samples. Of the peer group variables, the amount of homework reported by the three best friends appeared to be important at both schools. The implications of the results of the study are discussed.