- Social media
Twelve active social media services (Facebook, Instagram, LINE, note, Pinterest, Snapchat, TikTok, Twitter, WeChat, WhatsApp, Yahoo!Chiebukuro, and Youtube) were selected for inclusion. We originally examined 16 social media platforms, but four (Ameba blog, cocolog, Kakaotalk, and mixi) were excluded due to having a high amount of inactive users (more than 200 respondents who "used to use but rarely log in now")
- Participants
Two types of questionnaires were separately solicited through Yahoo! Crowdsourcing. We recruited users who had used at minimum one of the target social media services at least once. We paid the participants 3 to 5 yen worth of intra-site points as rewards, depending on the number of questions.
The questionnaire using the psychological safety scale for social media (Q1) was conducted between June 2020 and November 2020. We recruited 600 participants for each social media and received responses from 7200 people. A total of 7198 responses were analyzed, excluding two responses that were excluded as a result of outlier tests.
The questionnaire for the subjective classification of social media (Q2) was conducted in November 2020. About 200 people were recruited for each social media, and 2304 responses were obtained. All responses were included in the analysis. (Facebook n=199, Instagram n=192, LINE n=216, note n=218, Pinterest n=183, Snapchat n=173, Tik tok n=189, Twitter n=195, We chat n=182, Whatsapp n=164, Yahoo! Chiebukuro n=189, Youtube n=204)
- Psychological safety measures
The psychological safety measures consisted of the following seven questions, to which the subjects were asked to respond in a seven-point scale from "1. not at all applicable" to "7. very applicable".
1. When a user makes a mistake, they are often viewed coldly.
2. Users are able to discuss problems and difficulties.
3. Users may reject others because they are different from them.
4. It is safe to take risks on this platform.
5. It is difficult for me to ask for help from other users.
6. There is no user who is willing to waste my efforts.
7. My unique skills and talents are valued and utilized when I work with other users
These items were adapted from the original questionnaire by replacing the following terms: "team members" and "members of the team" in the original scale were changed to "users," and "this team" was changed to "this platform”.
- Operational Statistics
[years of operation]
To assess the size and influence of each social media platform, we obtained statistics pertaining to its global operations. For the number of years in operation, we referred to Wikipedia pages for each platform.
[mAU]
We examined the global data (media that publish mAU: Facebook, Instagram, LINE, note, Pinterest, Snapchat, TikTok, Twitter, WeChat, WhatsApp, Youtube) and domestic data (media that publish mAU: Facebook, Instagram, LINE, note, TikTok, Twitter, WhatsApp, Youtube) for May 2020 or the first fiscal year.
[login frequency]
In the questionnaire, we also asked the frequency of logging in to each social media. We asked the respondents to choose from the following six options: "I have not used social media since I created my account," "I used to use social media but rarely log in now," "I log in less than once a month," "I log in about once a month," "I log in once every few days to once a week," and "I log in almost every day”.
As an ordinal scale, a numerical value from 1 to 6 was assigned in order of decreasing login frequency, with "I have not used the service since I created my account" being 1 and "I log in almost every day" being 6.
- Subjective Classification Indicators
Based on various studies on the concept of social media, seven subjective classification indicators were established to show the characteristics of social media design and user behavior in social media.
- Anonymity [21]: Low anonymity - High anonymity
- Interaction with strangers [16]: Mainly interact with family and close friends - Mainly interact with strangers
- Self-disclosure [22]: Never post information about myself, such as my hobbies, social views, or appearance - Sometimes post information about myself, such as my hobbies, social views, or appearance
- Media richness [22]: Low media richness (difficult to use for real-time communication, unable to communicate non-text information, unable to communicate each other's situations and feelings) - High media richness (able to use for real-time communication, able to communicate non-text information, able to communicate each other's situations and feelings)
- Interaction level [23]: little interaction between users - a lot of interaction between users
- Openness [24]: Cannot be viewed without approval of the contributor, etc. - Anyone can view the postings without the contributor's approval
- Flow model: posts are stored and easily retrieved later. Real-time communication is not smooth - New posts are displayed clearly, and real-time communication is easy. New postings are displayed clearly, and real-time communication is easy, but it is difficult to search postings later.
The SD method (Semantic Differential scale) was used for grading. The words on the left were set to -3 and the words on the right to +3. These were randomly reversed in the questionnaire and treated as reversed items in the scoring.
- Statistics
We conducted the Kruskal-Wallis test and post-hoc Mann-Whitney U test for differences in psychological safety scores between social media. For the objective measures (years of operation, mAU, and number of logins), the association with psychological safety was determined by Spearman's rank correlation coefficient. For the subjective indicators (Anonymity, Interaction with mainly strangers, Self-disclosure, Media richness, Interaction level, Openness, and Flow model), the association with psychological safety was determined by Spearman's rank correlation coefficient, after averaging respondents scores for both surveys by social media platform.