The variable that measures student gender is also a composite, and this variable is
also a dummy variable which will be used to assess the effect of one group as it relates to
the other non-selected group. In this case female = 1 will be referenced against male = 0.
(N = 15,362).
The Regression Model
Because this is a replication of an earlier study, the calculation will consist of a linear
regression (OLS) which will provide regression coefficients to compare student
achievement between students whose parents are more involved in their academic
activities and students whose parents are less involved. Because different types of
parental involvement appear to affect academic achievement in different ways, the study
will compare the effect of the different types of parental involvement. For example, in the
original study, regression coefficients for two parental involvement variables, family
structure and parents’ discussing school activities with their children, were larger and
statistically significant relative to parents’ checking up on their children’s friends and
homework and parents’ attendance at school activities.
This study will make one significant departure from the original research conducted
by Jeynes (2005). Although the original study model included one regression described
as “non-SES” which was compared to another “SES” model, this research will be based
on a regression equation that includes SES as a control variable. All hypotheses are based
on a model which includes SES as a control variable. The main reason for the elimination
of the “non-SES” model is simply a re-evaluation of the original study itself. While