TABLE 3.

Adjusted Association between State Laws Allowing Exemptions or Substitutions and Selected^a School Practices, 2014.

			Adjusted Prevalence of School Activity if State Law…
School-Level Substitution or Exemptionb (No. Schools in Analysis)	AOR	95% CI	Allows Substitution/Exemption, %	Prohibits or Does Not Address Substitution/Exemption, %
Substitution activities
Participation in school sports (n = 225)	3.59*	1.33–9.68	26.0	10.8
Participation in school activities other than sports (e.g., JROTC, band; n = 314)	8.52***	2.90–25.03	52.3	13.8
Participation in community sports activities (n = 311)c	4.30*	1.43–12.96	18.6	5.9
Exemption activities
Positive/passing/high physical fitness test scores (n = 315)d	4.67**	1.49–14.62	20.7	6.5
Participation in vocational training (n = 222)e	5.92*	1.04–33.68	19.5	4.7

Each row corresponds to a separate multivariable logistic regression model to examine the association between the school-level PE substitution or exemption practices and the corresponding state law allowing such practices. A significant association with an AOR greater than 1 means that schools in states with a given substitution or exemption law are significantly more likely to engage in the given substitution or exemption practice. Models controlled for school level, majority race, locale, free/reduced-price lunch eligibility, enrollment, and region, as categorized in Table 1. Adjusted prevalences were computed from these models showing the average predicted probability of having each exemption or substitution if all schools were in states that did versus did not allow the given PE substitution or exemption.

^aTwo school practices were not included in this analysis: 1) state laws did not address community service as a PE substitution/exemption, and 2) the estimates of the analysis linking state law to enrollment in other courses were imprecise.

^bThe grouping of school practices by substitution/exemption is based on how the state laws refer to each practice as being either a substitution for PE or an exemption from PE (Table 2 and Appendix).

^cBecause no schools in the analytical sample in the Northeast allowed this substitution, this model did not control for region.

^dBecause only one state allowed fitness tests as an exemption, this model was run with “any exemptions” allowed for the state law predictor rather than the fitness test exemption specifically.

^eThis model did not control for free/reduced-price lunch eligibility or region, as these were both perfect predictors.

^*P < 0.05.

^**P < 0.01.

^***P < 0.001.

AOR, adjusted odds ratio; CI, confidence interval.