The impact of face coverings on audio-visual contributions to communication with conversational speech

Cognitive Research: Principles and Implications

Table 1 Model summaries for each of the three measures. Intercepts represent the mean of the reference level of each predictor variable (‘Clean’ for the auditory variable, and ‘No mask’ for the visual variable)

	Follow					Effort					Clarity
	Beta	SE¹	Statistic	95% CI¹	p value	Beta	SE¹	Statistic	95% CI¹	p value	Beta	SE¹	Statistic	95% CI¹	p value
(Intercept)	88	1.28	69.0	86, 91	< 0.001	18	1.67	10.6	14, 21	< 0.001	84	1.54	54.8	81, 87	< 0.001
Audio condition					0.002					0.023					0.002
Clean	–	–	–	–		–	–	–	–		–	–	–	–
Surgical	− 0.66	1.02	− 0.653	− 2.7, 1.3		− 0.37	1.31	− 0.280	− 2.9, 2.2		− 0.90	1.28	− 0.706	− 3.4, 1.6
FFP3	0.16	1.02	0.158	− 1.8, 2.2		0.77	1.31	0.589	− 1.8, 3.3		− 1.1	1.26	− 0.878	− 3.6, 1.4
Visor	− 3.2	1.01	− 3.19	− 5.2, − 1.3		3.3	1.31	2.52	0.72, 5.8		− 5.0	1.41	− 3.56	− 7.8, − 2.2
Visual condition					< 0.001					< 0.001					< 0.001
No mask	–	–	–	–		–	–	–	–		–	–	–	–
Mask	− 4.6	1.13	− 4.03	− 6.8, − 2.3		8.1	1.50	5.43	5.2, 11		− 7.4	1.33	− 5.60	− 10, − 4.8
Audio condition * visual condition					0.042					0.81					0.13
Surgical * mask	1.5	1.44	1.02	− 1.3, 4.3		− 0.41	1.85	− 0.219	− 4.0, 3.2		0.60	1.67	0.361	− 2.7, 3.9
FFP3 * mask	− 2.2	1.44	− 1.50	− 5.0, 0.67		1.1	1.85	0.615	− 2.5, 4.8		− 2.3	1.67	− 1.35	− 5.5, 1.0
Visor * mask	1.3	1.44	0.924	− 1.5, 4.1		− 0.42	1.85	− 0.226	− 4.0, 3.2		1.6	1.68	0.940	− 1.7, 4.9
Number of observations	2400					2400					2400