Skip to main content
. 2021 Jan 11;9(1):e24045. doi: 10.2196/24045

Table 2.

Fisher exact test comparing voice assistant pairs for accurate speech recognition, the presence of a verbal response, and whether the advice was clinically appropriate.

Voice assistant comparison Acc reca, n (%) P value Verb respb, n (%) P valuec Clin app advd, n (%) P value
Comparison 1 .22 .05 .65

Apple Siri 14 (100) 5 (36) 2 (14)

Amazon Alexa 11 (79) 11 (79) 4 (29)
Comparison 2 >.99 .71 >.99

Apple Siri 14 (100) 5 (36) 2 (14)

Google Assistant 14 (100) 7 (50) 3 (21)
Comparison 3 >.99 .45 >.99

Apple Siri 14 (100) 5 (36) 2 (14)

Microsoft Cortana 13 (93) 8 (58) 3 (21)
Comparison 4 .22 .24 >.99

Amazon Alexa 11 (79) 11 (79) 4 (29)

Google Assistant 14 (100) 7 (50) 3 (21)
Comparison 5 >.99 .42 >.99

Amazon Alexa 11 (79) 11 (79) 4 (29)

Microsoft Cortana 13 (93) 8 (58) 3 (21)
Comparison 6 >.99 >.99 >.99

Google Assistant 14 (100) 7 (50) 3 (21)

Microsoft Cortana 13 (93) 8 (58) 3 (21)

aAcc rec: accurate recognition.

bVerb resp: verbal response.

cAfter false discovery rate correction, no comparison was statistically significant.

dClin app adv: clinically appropriate advice.