. 2017 Jan 29;6(9):519–528. doi: 10.15171/ijhpm.2017.09

Table 3. Bivariate Analysis Results of Negative Binomial Regression for the Relationship Between the Each Independent Variable With Unhealthy Snacking Behavior .

Independent Variable	β Coefficient	Unadjusted P Value
IHPS program (no, yes)	-.10	<.01^a
School type (public, private)	.12	<.01^ab
Snack sold at school (no, yes)	-.10	.07^a
Presence of a SHW (no, yes)	.01	.81
School size (≤150, 151-300, <300)	-.08	<.01^ab
Presence of a canteen (no, yes)	-.03	.46
Gender (girl, boy)	.11	<.01^a
SES level (low, lower-middle, middle, higher-middle, high)	.08	<.01^a
Receiving pocket money (no, yes)	.31	<.01^a
Family structure (dual-parent, single-parent/others)	.14	.04^a
Nutritional knowledge	-.08	<.01^a
Cost/price sensitivity	.03	.08^a
Taste and sensory perception	.12	<.01^a
Social norms pressure	.18	<.01^a
Easy accessibility	.12	<.01^a
Media advertisements	.06	<.01^a
Perceived parental control	-.03	.06^a
Perceived self-efficacy	-.10	<.01^a
Self-body image	-.05	.05^a

Abbreviations: IHPS, Iranian health promoting school; SES, socio-economic status; SHW, school health worker.

^a Those variables showing a significant level < .2 in bivariate analysis.

^b A correlation coefficient higher than 0.50 between school type with school size and IHPS program made a potential risk for severe collinearity and school type was excluded from the multivariate model.