50

appear in both data sets, the new data set includes an additional math variable. Although

the variables were described in the methodology section, it is helpful to restate how the

variables available in the data set create an overall limitation to the study.

# As its name indicates, data for the Education Longitudinal Study (ELS 2002 2004) is

collected at regular intervals, occurring every two years. Theoretically, this practice

allows researchers to evaluate the extent to which outcomes such as academic

achievement are affected by activities or conditions such as parental involvement or

family structure which occurred two years before. The first year of data included in this

data set this study was 2002, and the follow-up, or first wave data was collected in 2004.

# Data released at the time this study was initiated included only one outcome variable

which is differentiated from the other math variable by its name — the 1^{st }wave follow-

up math variable. The two math variables discussed in this analysis are differentiated by

name and the time in which data was collected — the 1^{st }wave follow-up math variable

data was collected in 2004, and data for the measure identified simply as the math

variable was collected in 2002. The predictive ability of this study is limited because data

for all outcome variables other than the 1^{st }wave follow-up math variable were collected

in the same year as the independent variables, 2002.

# For the key achievement variable, the 1^{st }wave follow-up math variable, the mean

scores are 45.44 for children from intact families, 35.16 for children whose parents

cohabitate, 38.11 for widowed parents, 34.33 for students whose parents are separated,

38.97 for children from divorced parents, and 31.97 for students whose parents never

married. For the 1^{st }wave follow-up math variable, the disparity between married and