Table 3.

Mediation analyses of the effects of exercise identity on Instagram use and physical activity levels.a

Effects		B	95% CI	β	P value
Instagram PAbphotos—PA levels (METc ≥3/week)d
	Path a: Instagram PA photos—exercise identity	0.10	0.01 to 0.18	.33	.02
	Path b: Exercise identity—PA levels	34.0	11.66 to 56.36	.51	<.01
	Path c: Instagram PA photos—PA levels	–3.65	–9.70 to 2.39	–.18	.22
Instagram PA following—PA levels (MET ≥3/week)e
	Path a: Instagram PA following—exercise identity	3.18	1.02 to 5.33	.48	<.001
	Path b: Exercise identity—PA levels	29.34	7.16 to 51.52	.44	.01
	Path c: Instagram PA following—PA levels	32.75	–129.17 to 194.68	.06	.68
Instagram PA “likes”—PA levels (MET ≥3/week)f
	Path a: Instagram PA “likes”—exercise identity	0.02	–0.01 to 0.06	.21	.17
	Path b: Exercise identity—PA levels	28.6	6.75 to 50.47	.43	.01
	Path c: Instagram PA “likes”—PA levels	0.28	–2.06 to 2.61	.04	.81

^aModel covariates include age, gender, B (unstandardized beta), and β (standardized beta).

^bPA: physical activity.

^cMET: metabolic equivalent of task.

^dIndirect effect: 3.28 (95% CI 0.54 to 8.11); direct effect: –3.65 (95% CI –9.70 to 2.39).

^eIndirect effect: 91.9 (95% CI 36.93 to 174.82); direct effect: 60.53 (95% CI –115.15 to 236.21).

^fIndirect effect: 0.69 (95% CI –0.10 to 2.28); direct effect: 0.28 (95% CI –2.06 to 2.61).