Table 2 Features of Twitter usea among the people who got lost due to dementia

From: How can we better use Twitter to find a person who got lost due to dementia?

 

Total No. (%)

Average no. (SD) per person who got lost

People with dementia who can be found

Yes

No

p-valueb

Yes

No

p-valuec

Types of Twitter users:

 No. of tweet-writers

183

44

 

4.6 (5.0)

3.1 (3.6)

0.258

 From police departments

15 (8.2%)

1 (2.3%)

0.294

0.4 (0.5)

0.1 (0.3)

0.009

 From media organisations

60 (32.8%)

10 (22.7%)

0.265

1.5 (1.7)

0.7 (1.1)

0.051

 No. of retweeters

1807

172

 

45.2 (148.8)

12.3 (18.3)

0.178

 From police departments

7 (0.4%)

0 (0.0%)

0.884

0.2 (0.5)

0.0 (0.0)

0.003

 From media organisations

6 (0.3%)

2 (1.2%)

0.312

0.2 (0.4)

0.1 (0.4)

0.950

 No. of followers, in thousands

12,338

963

 

308.4 (749.7)

68.8 (96.4)

0.031 d

 From police departments

388 (3.1%)

9.3 (1.0%)

<0.001

3.5 (4.7)

0.7 (2.4)

0.010

 From media organisations

6314 (51.2%)

474 (49.2%)

0.256

8.9 (4.8)

4.5 (5.5)

0.013

Types of Twitter posts:

 No. of original tweets

199

50

 

5.0 (5.4)

3.6 (3.7)

0.292

 From police departments

19 (9.5%)

1 (2.0%)

0.143

0.5 (0.8)

0.1 (0.3)

0.005

 From media organisations

66 (33.2%)

10 (20.0%)

0.102

1.6 (1.9)

0.7 (1.1)

0.031

 Tweets with photos

175 (88.0%)

37 (74.0%)

0.024

4.4 (5.3)

2.6 (3.5)

0.177

 Tweets with website links

190 (95.5%)

44 (88.0%)

0.098

4.8 (5.4)

3.1 (3.7)

0.224

 Tweets with hashtags

44 (22.1%)

13 (26.0%)

0.691

1.1 (1.6)

0.9 (1.3)

0.696

 Average length of tweets (no. of words)

   

15.0 (3.8)

17.1 (4.6)

0.136

 No. of retweets

1828

173

 

45.7 (150.1)

12.4 (18.4)

0.176

 Retweets from police departments

8 (0.4%)

0 (0.0%)

0.809

0.2 (0.6)

0 (0.0)

0.044

 Retweets from media organisations

6 (0.3%)

2 (1.2%)

0.308

0.2 (0.4)

0.1 (0.4)

0.950

 No. of retweet / each original tweet

   

10.8 (24.8)

6.1 (9.8)

0.324

  1. Remarks:
  2. aAll tweets were counted until the time before the people who could be found. The average time to find the people was 21.4 h from the first post of original tweet. This was also used as the cut-off times for the people who could not be found in the control group
  3. bp-value of Χ2 test, and significant results (p-value <0.05) were shown in bold
  4. cp-value of independent sample t-test, and significant results (p-value <0.05) were shown in bold
  5. dThe number of followers was transformed to logarithmic scale before performing a t-test