Table 3.
Frequency of studies by ethics-related factors from a systematic review of methodological approaches and ethical considerations for public health research using Twitter data, 2006-2019.
| Ethics-related factors | Frequency (N=367), n (%) | |||
| Level of identification | ||||
|
|
No identifying information | 174 (47) | ||
|
|
Anonymized data and paraphrased tweets | 131 (36) | ||
|
|
Identifiable information and direct quotes | 62 (17) | ||
| Institutional review board (IRB) approval obtained | ||||
|
|
Yes | 119 (32) | ||
|
|
No | 23 (6) | ||
|
|
Not mentioned/unclear | 225 (61) | ||
| Among those with IRB approval (n=119) | ||||
|
|
Exempt | 68 (57) | ||
|
|
Nonexempt | 2 (2) | ||
|
|
Not specified (eg, “approved”) | 49 (41) | ||
| Informed consent of Twitter handler attempted | ||||
|
|
Yes | 0 (0) | ||
|
|
No | 119 (100) | ||
| Any discussion of ethical considerations, including disclaimers | ||||
|
|
Yesa | 109 (30) | ||
|
|
|
Discussion of anonymization process | 53 (49) | |
|
|
|
Extensive discussionb | 5 (5) | |
|
|
|
Other discussion, including disclaimers | 54 (49) | |
|
|
No | 258 (70) | ||
aNote that 3 articles included both an extensive discussion of ethics as well as details regarding their anonymization process.
bThe denominator for the articles that discussed ethics is 109.