Table 3.

Regression results predicting moderate activity at 32 weeks

	Predicting Self-Reported 32-Week Moderate Physical Activity	Predicting Self-Reported 32-Week Vigorous Physical Activity
		Logit			NB
	B(SE)	p	B(SE)	p	B	p
Military Status
Active Duty	17.2 (20.5)	0.40	-0.27 (0.26)	0.30	0.63(0.37)	0.09
Other TRICARE beneficiary	Reference
Body Mass Index
Normal Weight	Reference
Overweight	16.1 (27.1)	0.55	0.19 (0.31)	0.52	0.13 (0.4)	0.76
Obese	-2.7 (23.4)	0.91	0.47 (0.34)	0.17	-0.02 (0.46)	0.97
Baseline activity	0.58 (0.07)	< 0.0001	0.01 (0.002)	0.02	-0.02 (0.01)	0.0002
	Predicting Accelerometer-Measured 32-Week Moderate Physical Activity	Predicting Accelerometer-Measured 32-Week Vigorous Physical Activity
		Logit			NB
	B(SE)	p	B(SE)	p	B	p
Military Status
Active Duty	-15.4 (8.8)	0.08	-2.7(0.81)	0.0008	-25.1(121,081)	1.00
Other TRICARE beneficiary	Reference
Body Mass Index
Normal Weight	Reference
Overweight	6.8 (10)	0.4959	2.06 (0.83)	0.01	1.7 (2.1)	0.41
Obese	-1.8 (11.6)	0.8774	-0.5 (0.95)	0.59	-22.6 (118,166)	1.00
Baseline activity	0.47 (0.07)	< 0.0001	0.62 (0.22)	0.006	0.15 (0.28)	0.60

Models controlled for baseline activity, age, intervention group, and previous live birth. SE Standard error, NB Negative binomial. Zero-inflated negative binomial model consists of logistic model and negative binomial model