Table 6

Multiple linear regression analysis on audience engagement*

Variable	Log-transformed data			Original data
Variable	Estimated intercept	SE	t-value	Estimated exponential	P value
Author types
(Intercept)	2.06831	1.46066	1.416	7.91	0.161
Doctors and science communicators	0.67993	0.52333	1.299	1.97	0.198
Institutional users	1.03571	0.96261	1.076	2.82	0.286
Nurses and other healthcare workers	−0.69053	0.63963	−1.08	0.50	0.284
Video production
Presence of people	4.23693	1.03791	4.082	69.20	0.0001
Background music	−0.31938	0.51654	−0.618	0.73	0.538
Emoji	−0.25708	0.70885	−0.363	0.77	0.718
Animation/flash	0.92154	0.4898	1.881	2.51	0.064
Subtitles	−0.05565	0.98898	−0.056	0.95	0.955
	2.06831	1.46066	1.416	7.91	0.161
Video content DISCERN instrument
Reliability of the videos (questions 1–8)	−0.03388	0.04143	−0.818	0.97	0.416
Quality of treatment choices (questions 9–15)	0.13811	0.06777	2.038	1.15	0.045
Overall information quality (question 16)	0.06258	0.29781	0.21	1.06	0.834

Multiple R² 0.4168; adjusted R² 0.3196; F-statistic 4.289 on 12 and 72 degrees of freedom; p value 4.259e-05. Video content was measured by six questions obtained from reports of Goobie et al. 27 Italic content reflects significant p values.
*Residual SE: 1.776 on 72 degrees of freedom.