Table 4.

Unstandardized coefficients from multilevel effects of daily steps on same-day mood (N=79)^a.

Outcome	Model 1b: mood				Model 2c: mood				Model 3d: mood
	β	SE	P value	β		SE	P value	β		SE	P value
Intercept	3.18	1.90	.01	3.22		1.90	.09	3.22		1.90	.09
Daily steps	.06	0.02	<.001	–.07		0.05	.19	–.23		0.12	.06
Age (years)	.04	0.02	.13	.04		0.02	.13	.04		0.02	.14
Sex	–.47	0.40	.25	–.50		0.40	.22	–.47		0.40	.25
Condition	–.50	0.37	.18	–.49		0.37	.19	–.49		0.37	.18
Education	.05	0.07	.49	.05		0.07	.49	.05		0.07	.49
Health	.03	0.01	.02	.03		0.01	.02	.03		0.01	.02
Average steps	–.09	0.08	.29	–.09		0.08	.28	–.09		0.08	.29
Daily steps×sex	—e	—	—	.08		0.03	.01	—e		—	—
Daily steps×age	—e	—	—	—e		—	—	.005		0.002	.02

^aDaily steps were rescaled by dividing the number of steps by 1000; therefore, the β value reflects change in mood per 1000-step increase.

^bModel 1: level 1 variance=2.38 (SD 1.54); level 2 variance=1.60 (SD 1.27). Akaike information criterion=4981.2; Bayesian information criterion=5033.8; log likelihood=–2480.6.

^cModel 2: level 1 variance=2.38 (SD 1.54); level 2 variance=1.60 (SD 1.26). Akaike information criterion=4976.8; Bayesian information criterion=5034.7; log likelihood=–2477.4.

^dModel 3: level 1 variance=2.39 (SD 1.55); level 2 variance=1.60 (SD 1.26). Akaike information criterion=4977.7; Bayesian information criterion=5035.5; log likelihood=–2477.8.

^eOutcome was not included in the model.