. 2019 Apr 2;4(2):e13634. doi: 10.2196/13634

Table 2.

Analysis of variance of hemoglobin A_1c by blog use status and a second variable.

Second variable	P value
	Model	Blog use	2nd variable	Interaction
Pump use^a	<.001^b	.012^b	.001^b	.22^c
Continuous glucose monitor use^a	<.001^b	.04^b	.03^b	.24^c
Marital status	<.001^b	.45	.006^b	.20^c
Employment	.001^b	.58	.25	.10^c
Age	.002^b	.09	.11	.66^c
Education	.007^b	.04^b	.19	.09^c
Gender	.008^b	.001^b	.23	.17^c
Time since diagnosis	.02^b	.01^b	.10	.07^c
Ethnicity	.03^b	.37	.20	.83^c
Smartphone use	.04^b	.11	.37	.75^c
Race	.01^b	.40	.40	.03
Internet access method	.15	.03^b	.95	.44^c

^aNo significant interaction; significance achieved for ANOVA model and additive effects of blog use and second variable.

^bSignificant at the P<.05 level.

^cNo interaction of the two test variables (P>.05).