Table 2.

Satisfaction measures, with Friedman significance tests for differences among conversational assistants. P values were adjusted using the Benjamini-Hochberg procedure to decrease false discovery rate.

Item	Anchor 1	Anchor 7	Median (interquartile range)
			Overall	Alexa	Siri	Google Assistant	P value
How satisfied are you with the conversational interface?	Not at all	Very satisfied	4 (1-6)	1 (1-2)	6 (4-6)	4 (2-5)	<.001
How likely would you be to follow recommendations given by the system?	Not at all	Very much	4 (2-6)	2 (1-3)	6 (5-7)	4 (2-6)	<.001
How much do you trust the conversational interface?	Not at all	Very much	4 (2-6)	1 (1-3)	6 (5-6)	4 (2-6)	<.001
How easy was talking to the conversational interface?	Very easy	Very difficult	5 (2-6)	6 (2-7)	4 (2-6)	5 (3-6)	.05
How much do you feel that the conversational interface understood you?	Not at all	Very much	3 (1-5)	1 (1-3)	5 (4-6)	3 (2-5)	<.001
Did you think you were interacting with a person or a computer?	Definitely a person	Definitely a computer	7 (6-7)	7 (7-7)	7 (6-7)	7 (6-7)	.05