Multiple regression analyses were utilized to explore the relationships between responses on the PSS-10 and EBPQ.
Statistical adjustments were made to control for age, education, and income. Each eating behavior domain was
regressed on the PSS-10. Perceived stress significantly predicted responses for the emotional eating (B=.29; pb.01)
and haphazard planning domains (B=.22; pb.01) of the PSS-10. Perceived stress accounted for nine percent of the
variance in emotional eating and 11% of the variance in haphazard planning. Perceived stress did not predict responses
for the following eating behavior domains: low fat eating, cultural/lifestyle eating, snacking on sweets, and meal
skipping. Additional regression statistics can be found in Table 4.
Additional regression analyses examined the relationships between responses on the PSS-10 and EBPQ among
participants in the sample that fell into a BMI category of overweight or obese with a BMI greater than or equal to 25.
One hundred twenty-two of 159 participants in the sample met this criterion. Similar relational patterns between the
PSS-10 and EBPQ were found for overweight and obese participants, though the prediction strength was greater. While
perceived stress accounted for eight percent of the variance in emotional eating in the entire sample, it accounted for 11
percent of the variance in the subsample of overweight and obese participants. Similarly, perceived stress accounted for
12 percent of the variance in haphazard planning within the overweight subsample as compared to eight percent in the
entire sample. Within the overweight subsample, perceived stress predicted snacking on sweets (B=.21; pb.05);
however, this prediction was not significant across the entire sample. Other non-significant relationships remained the
same for prediction of low fat eating, cultural/lifestyle eating, and meal skipping.
Mediation analysis was utilized to determine whether the relationships between perceived stress and the eating
behavior domains were mediated by haphazard planning. This analysis was carried out for the only dependent variable
other than haphazard planning that was predicted by perceived stress in the entire sample in the multiple regression