47

# Chapter 4

RESULTS

# Family Structure

In the original study, Jeynes (2005) identified intact family status as the parental

involvement measure which had the single greatest effect on academic achievement

(p.112). Although that assertion was unsupported by Jeynes’ regression models, the

variable representing family structure produced statistically significant, positive

coefficients in the original study. In this replication of Jeynes’ study, the family structure

variable included in the ELS 2002-2004 data set produced positive, statistically

significant standardized coefficients that nearly matched those in the original study.

# Although the original study did not include a model with unstandardized regression

coefficients, the ELS 2002-2004 data set variable for family structure yielded

unstandardized coefficients which are also positive and statistically significant. The

regression results from this study suggest that, relative to children with different family

structure types, children from intact families have higher test scores, as measured by the

four test variables included in the regression model — after controlling for all other

variables in the regression equation. Family structure held a prominent position in the

original analysis, and this analysis will begin by discussing the family structure variable.

# As indicated in Table 1 below, children from intact families unadjusted mean score

on the standardized math test is 52.29, and the mean score for the reading variable is

# 51.98. These figures which were produced using the ELS 2002-2004 data set, statistically

significantly different from the mean scores Jeynes found in the NELS 1990-1992 data