As these variables
Quinlivan et al.
142 2011 The Authors
Australian and New Zealand Journal of Obstetrics and Gynaecology 2011 The Royal Australian and New Zealand College of Obstetricians and Gynaecologists; 51: 141–146
were categorical in nature, except the age of participants,
Pearson chi-squared tests or Fisher exact tests for small
expected cell size were applied. Student’s t test was
employed to test for any differences in ages between groups.
Because gestational diabetes was a categorical variable, a
comparison of proportions between groups was carried out
using the unweighted logistic regression technique with
adjustment for confounding factors. Weight gain and
birthweight were continuous variables; hence, linear
regression techniques were applied to compare any
differences between the intervention and control groups. To
test the hypotheses, a significance level of 5% was used in
the study.