[Add] psychological safety measure

[JamesPHoughton]

Probably from: Edmondson, Amy. 1999. “Psychological Safety and Learning Behavior in Work Teams.” Administrative Science Quarterly 44 (2): 350–83.

[JamesPHoughton]

Adapted in Nicole Abi-Esber, Alison Wood Brooks, and Ethan Burris. “Feeling Seen: Leader Eye Gaze Promotes Psychological Safety, Participation, and Voice.” as:

1. During this conversation I felt if I made a mistake, the other people in the group would hold it against me. (R)
2. During this conversation I felt the other people in the group were able to bring up problems and tough issues.
3. During this conversation I felt the other people in the group sometimes rejected others for being different. (R)
4. During this conversation I felt it was safe to take a risk with the other people in the group.
5. During this conversation I felt like it was difficult to ask the other players for help. (R)
6. During this conversation I felt like no one would deliberately act in a way that undermined me and my efforts.
7. During this conversation I felt like my unique skills and talents were valued and utilized while working with the other people in the group.

(responses ranging from 1= strongly disagree to 7 = strongly agree)

We could probably factor out "During this conversation I felt..."

[sky1113]

I'm not sure if you wanted this survey checked as well, but it looks to be very similar to the questions in the original survey and has not been "adapted" much.

[sky1113]

Also, it looks like these new surveys mostly use 7-point scales as opposed to the 5-point ones we were previously using. I wanted to confirm that I should implement a 7-point scale and ask if changing our 5-point scales to 7-point scales would be something you would be interested in seeing.

[sky1113]

• [x] Create a new folder in surveys/src/surveys w/ descriptive, easy-to-read name superSpecialSurvey/
• [x] Create SurveyJS .json file (e.g. superSpecialSurvey/superSpecialSurvey.json)
• [x] Create a .bib file citing your references for the survey (always: references.bib)
• [x] Create a storybook markdown file to document (e.g. superSpecialSurvey.stories.mdx)
  • [x] Survey purpose
  • [x] Expected behavior
  • [x] Design rationale
  • [x] APA-style bibliography
• [x] Create file with function to aggregate survey data (e.g. superSpecialSurvey.score.js)
• [x] Create PR and link to this issue

[JamesPHoughton]

A few notes, some of which are probably things to follow up with other researchers about - I'll add to our discussions.

• The order of the disagree scales doesn't precisely invert the order of the agree scales (Disagree vs Somewhat disagree)
• I'm realizing that most people will probably skip the tex and go straight to the bubbles, and so miss where it says "During this conversation I felt"... so we may want to make that more prominent
• Do we want to have the "Psychological Safety Assessment" title visible to the participants? Does it change the outcome if respondents know what construct we're trying to get at?

[JamesPHoughton]

From the original paper:

The "feeling seen" paper makes several adjustments:

1. The language replaces "members of the team" with "people in the group", to generalize to situations in which there isn't really a team (ie, experiments with convened groups that have not formed teams).
2. The questions are put in past tense to refer to what has occurred in the experiment.
3. The focus is shifted to "other people in the group" rather than all members of the team. This may move the focus away from the respondent, although it isn't clear that this was an explicit choice.

The first two adjustments are clearly needed for using in the online experiment context. The third I am not sure we should keep. It is valid for a participant's own experience of (e.g.) sharing problems and tough issues to influence their perception of psychological safety in the group. On the other hand, an individual's ability to bring up problems and tough issues is also related to "Voice", a separate construct. On balance, I'm leaning more towards sticking as close to the original language as possible, given adjustments 1 and 2.

On the other hand, I still wonder if the adjustments that adapt the survey from an existing group to a convened group are enough. Do people in this context have enough experience with one another to answer about what would happen if you made a mistake? That's a liability of the measure.

Updated questions

1. "If you made a mistake in this group, it was often held against you."
2. "People in the group were able to bring up problems and tough issues."
3. "People in the group sometimes rejected others for being different."
4. "It was safe to take a risk with people in the group."
5. "It was difficult to ask other people in the group for help."
6. "No one in the group would deliberately act in a way that undermines my efforts."
7. "My unique skills and talents were valued and utilized while working with the other people in the group."