Data analysis

To carry out the aims of this study, a subsample of those living under the FPL and eligible for food stamps was created from the 2007 (n = 10 971) and 2011 (n = 14 928) waves of the NSCH. This poverty subsample represented 16% (n = 95 677) of the 2011 wave, which was a 4% increase from the previous wave in 2007 (n = 91 642). Our first study aim was to examine differences between families participating in SNAP compared to eligible families not receiving SNAP. Families living below the FPL participating in SNAP were compared to those living below the FPL and not participating in SNAP on a range of demographic characteristics, including age of child, child gender, family structure (e.g., single-mother household, 2-parent household), and difficulty affording basic necessities.

Logistic regression and multivariate logistic regression analyses were conducted to assess for differences between the 2 groups across demographics and material deprivation using the 2007 and 2011 data sets. Statistical difference in age of child between the 2 groups was assessed via a t test. However, the variable related to difficulty affording basic necessities was only available in the 2011 data set; therefore, analyses with that variable were confined to that wave of data.

Next, logistic regression analysis was conducted to examine the association between difficulty affording basic necessities and educational outcomes. All demographic variables were included in the models to control for individual- and household-level variables that could be potential confounders. The third and final analysis examined the associations between difficulty affording basic necessities and educational outcomes with SNAP participation added as a potential moderating variable. More specifically, in order to assess for moderation, interaction terms were included in the models described for the second analysis related to difficulty affording basic necessities and educational outcomes.