Melbourne Graduate School of Education - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 2 of 2
  • Item
    Thumbnail Image
    The effects of schools on achievement in science
    Owen, John M (1943-) ( 1975)
    The study sought to identify factors which were based in schools which affected the performance of sixth form students in science in Victorian schools, In order to identify school effects, allowance was made using multiple regression analysis for factors which were shown to contribute to academic performance but were those over which the school had no control. Use was made of information collected during a. study of science achievement by the International Association for the Evaluation of Educational Achievement (IEA). A sample of 37 schools was used the probability of selection of the school was proportional to its enrolment. Within each school, a random sample of students in the sixth form was made to select the students to take part in the testing program. Information collected enabled a predicted score for each school to be made and this was compared with the actual score obtained by averaging the scores of each student in the sample. Two groups of five schools were then selected for comparative study; one group which had performed better than expected and the other which had performed below expectations. The comparison of the two groups of schools to identify school factors was achieved by the study of the responses of teachers, students and school principals on survey instruments. In addition a visit was made to each school to gather further information. These procedures enabled the identification of school characteristics which were seen as contributing factors to the performance of students on tests of science achievement.
  • Item
    Thumbnail Image
    Distinguishing the science content taken by grade 12 students
    Cross, R. J. ( 1977)
    The population of grade 12 students in Australian secondary schools has been steadily increasing over the past two decades. For most of this period the percentage of students at this level choosing science-type courses has been decreasing, and recently the actual number taking physics and chemistry has declined in some states. This study aimed to find a set of variables that would maximize the prediction of grade 12 student science content. Emphasis was directed toward identification of science talented students not opting for high science content in grade 12, and, equally as important, those of low science ability who select predominantly science courses at this level. It was proposed that the variables could be measures of any area likely to be related to the criterion. For example, factors associated with the home, the school, and personal measures were all included. The variable set was then searched for that combination returning optimal criterion prediction. Attention was focussed on six main units of analysis viz males, males of higher science ability, males of lower science ability, females, females of higher science ability, females of lower science ability. The data in each unit was subjected to both discriminant (stepwise and direct) analysis and a process similar to a stepwise regression procedure called the Automatic Interaction Detector (AID). AID employs a branching process using variance analysis to subdivide the sample into subgroups which maximize dependent variable value prediction. The International Association for the Evaluation of Educational Achievement (IEA) conducted a series of tests on a stratified random sample of grade 12 students throughout Australia in 1970. The results, held at ACER, included measures of some 418 variables thirty four of which were selected for this investigation. Included in this group were the results of the four Commonwealth Secondary Scholarship Examination (CSSE) ability tests taken two years earlier. Analysis units were formed on the basis of sex and CSSE - Science score. The results indicate successful science content prediction is possible with the personal or internal variables of science interest, attitudes and abilities, consistently being of greatest importance. The participating external variables vary depending on the unit of analysis. The non-monotonic "State" and "Type of School" factors are predominant in AID analyses.