Skip to main content
. 2021 Mar 5;23(3):e26482. doi: 10.2196/26482

Table 3.

Statistical distributions of the emulated questionnaire answers.

Variable Total tweets, n (%)a COVID-19-related tweets, n (%) P value
Overall 1,015,655 (100) 27,216 (100) N/Ab
User type

Person 913,480 (89.94) 19,869 (73.00) <.001

Organization 102,175 (10.06) 7347 (27.00) N/A
Gender

Male 481,770 (52.74) 11,997 (60.38) <.001

Female 431,710 (47.26) 7872 (39.62) N/A
Age (years)

≤18 346,483 (37.93) 3542 (17.83) <.001

19-29 350,959 (38.42) 5798 (29.18) N/A

30-39 104,228 (11.41) 3640 (18.32) N/A

≥40 111,810 (12.24) 6889 (34.67) N/A
Concern

Economics 142,090 (13.99) 9334 (34.30) <.001

Health 141,176 (13.90) 6152 (22.60) N/A

Politics 73,838 (7.27) 5434 (19.97) N/A

Education 64,799 (6.38) 4284 (15.74) N/A

Entertainment 79,119 (7.79) 1736 (6.38) N/A
Sentiment polarity

Overall (–1 to 1), mean (SD) 0.1067 (0.4647) 0.0659 (0.4941) <.001

Positive (–1 to 0.05) 431,247 (42.46) 11,744 (43.15) <.001

Neutral (–0.05 to 0.05) 318,713 (31.38) 6632 (24.37) N/A

Negative (0.05 to 1) 265,695 (26.16) 8840 (32.48) N/A

aAll values are expressed as n (%), except for overall sentiment polarity, which is expressed as mean (SD).

bP values were calculated for the main variables and not for individual responses.