PCAP-13 2007: Factors Contributing to Performance in Mathematics and Science is the first of a series of research projects in which the PCAP 2007 data set is used to examine questions of interest to educational policy-makers and practitioners in Canada. It focuses on the mathematics/science components of PCAP 2007 and, in particular, on the factors contributing to performance in these two subjects. 

The comparative results reveal statistically significant differences in student performance across jurisdictions and language groups. Although such differences are of considerable policy interest, further analysis reveals that differences among jurisdictions are small compared to differences between students and schools. This has major implications for the statistical modelling of jurisdictional and language differences.

Some significant findings include the following:

Student characteristics

  • Grade level of students and number of books in the home (a measure of socioeconomic status or SES) are the most consistently positive predictors of achievement in both mathematics and science.
     
  • Speaking French as a home language is a positive predictor.   
     
  • A gender effect favouring males is found for mathematics and science. 
     
  • Students born in Canada do less well in mathematics than those born outside of Canada; however, the opposite is true for science.  
     
  • Student enjoyment of school, enjoyment of reading, and student perception of being a good reader are positive indicators of both mathematics and science performance.
     
  • The results for variables such as homework time and days absent from school indicate that more time invested in learning is associated with higher achievement. 

 

School characteristics

  • Students in private schools have higher performance. 
     
  • Schools in larger communities have higher performance in mathematics, but not in science.
     
  • Class size seems to make little difference to achievement in either mathematics or science.

 

Jurisdictional effects

  • The jurisdiction/language-level effects account for only 3 per cent of the total variance in achievement in mathematics and science.
     
  • Adding student- and school-level variables aggregated to the jurisdiction accounts for more of the variance than using the student- and school-level variables at their measured levels.
  • Specific jurisdiction-level variables are not statistically significant, which is related to the small number of units for analysis (15, including language breakdowns) and the subsequent large number of standard errors.

 

Equity of school systems within jurisdictions

  • Some jurisdictions have managed to reduce disparities among students and schools more than others.
     
  • The differences in school-level variance are greater in mathematics than in science. 
     
  • The effect of SES on achievement is similar across jurisdictions.