Participants education levels were grouped into three levels: (1) Limited/standard grade (grouped together due to small sample size in limited education group)—for those who finished education at age 16 or under; (2) Higher/advanced higher for those who finished education at age 17 or 18; and (3) Tertiary for those who finished education at 19 or over.
Although there were no differences in scale scores by level of education, the sample size for the lower education segment was small.
To control for outliers and maintain the full sample size, we winsorized age, education, partner age, partner education, number of children, and number of pregnancies using the last value carried forward method to impute missing and extreme values.
Assuming an 8-month post-baseline retention rate of 66%, a total sample size of 450 per education condition will need to be recruited/consented at baseline (n = 900).
*Logistic Model adjusted for age, sex, race, income, education **Total sample size for analysis 6375, variables unknown drinking status (n= 225) and unknown physical activity (n= 97) were excluded from final analysis due to small sample size.
The final sample sizes for the education analyses were restricted to those older than 20, and for occupation to those older than 16 and in the 2003-2004 cycle.
Future research, however, should aim to replicate these findings as our sample size among women with no formal education was small and our confidence interval was wide.
The sample size allowed for six matching variables: age, education, marital and working status, menopausal status, and existence of comorbidities.
Sample size limitations in the 1993 survey for the higher education category were overcome by combining the complete secondary education and higher education categories in the analyses presented in this paper.
The multivariate results exclude 71 persons with missing information on education level, for a final sample size of 9,686.
