Table 1 Data analysis process

1

A master’s student holding a bachelor’s degree in psychology transcribed all 25 h, 9 min, and 2 s of recorded interviews.

2

Another master’s student holding a bachelor’s degree in psychology arranged all transcribed interviews in an Excel spreadsheet for more feasible analyses of content. This Excel spreadsheet did not contain any personal information about the participants.

3

JS and MK reviewed transcribed data and generated a list of rough categories in an open-coding process.

4

JS and MK reviewed rough categories and identified clusters of categories, which they discussed and revised. A preliminary coding list resulted with categories describing the themes of the debriefing myths and containing a description of their contents.

5

While using an iterative process [27], JS and MK moved back and forth between the original transcribed data, their assumptions, concepts of the relevant literature (e.g., single-loop / double-loop learning [70], psychological safety [69]), and the emerging categories. After having processed a portion of the data, the chosen categories were reviewed, and it was examined whether redundant or overlapping categories exist.

6

The final version of the coding scheme and the categories, respectively, were then applied by JS and MK for re-coding the complete data set.

7

For ensuring inter-rater reliability, JS and MK independently coded 15% (238/1591) of the material [78]. Cohen’s Kappa was .91, indicating very good inter-rater reliability [79].

8

Absolute and relative frequencies for all categories were determined.

9

Extracting debriefing myths, i.e., beliefs in contrast to existing scientific evidence