representing a reduction in female reading scores of approximately 75 percent, relative to
the male reference group. In the new study, the coefficient for the 1st wave follow-up
math score is weak B = -.02, and statistically significant (p < .01).
For this research, variables for school location were included in the regression
equation for the purpose of comparing scores between two major categories of schools.
After controlling for all other independent variables in the model, data in the ELS 2002-
2004 data set produced location coefficients for the majority of achievement variables
that were not statistically significant. The single exception was the indicator variable for
urban schools which rendered a coefficient, relative to the suburban reference group, for
the reading score which was weak B = .03 but statistically significant (p < .001).