Table 1:

Frequency distribution of tweet characteristics

Characteristic	Frequency	High ER Tweets	% High ER	Odds Ratio^{^a}	95% CI^{^a}	P Value^{^a}
Picture
Yes	405	235	58.02%	28.749	17.654–46.816	.000
No	627	23	3.67%	^#
Hashtag
Yes	751	242	32.22%	3.266	1.747–6.108	.000
No	281	16	5.69%	^#
Weekday	910	231	25.38%	0.613	0.348–1.081	.91
Weekend	122	27	22.13%	^#
Time of day
Morning	606	200	33.00%	^#	0.298–0.730	.001
Afternoon	318	47	14.78%	0.467	0.249–1.842	.445
Evening	66	7	10.60%	0.677	0.115–1.773	.254
Night	42	4	9.52%	0.452

↵Note:—# indicates reference category; ER, engagement rate.
↵a From the logistic regression model. The “Month of Tweet” (Fig 1) was also part of the logistic model and did not show any independent statistical significance.