Accelerated CREP project number: AC1938. *This component contains offline transcription of the final version of the survey. Another file presents survey flow. The survey was distributed via Facebook and presented online on the platform Qualtrics.* **DESCRIPTION OF THE PROCEDURE** **Important note: We accidentally swapped Gerald knowledge control vignette for ignorance control vignette. I.e. there were two Gerald ignorance control vignettes in our study and no Gerald knowledge control vignette.** The first page presented information regarding the research along with a consent statement. Participants then proceeded to a page with instructions and then the first vignette was presented followed by survey items. There were binary items as well as items utilizing a combination of slider and visual analogue scale. Qualtrics wasn't allow to use traditional visual analogue scale. However, it was possible to move the slider with both drag-and-drop and point-and-click process. The scale contained two anchors, no numeric labels and wasn't allow dynamic numeric feedback to show. The only difference from VAS presented in official preprint was therefore additional presence of a slider. Other two vignettes and corresponding items were then displayed in the same fashion. Later on, three open ended items were presented followed by demographic questions (gender, age, years of education, binary questions whether or not was the participant a student during the filling out the questionnaire and if Czech is his or her first language). We asked whether or not Czech language is respondent's first language to distinguish between Czech and Czech speaking Slovak participants and those who might not comprehend the text as needed. If Czech was not his or her first language, we then asked how well versed is the participant in Czech language (four options: 'very well', 'well', 'not so well', 'not at all'). The last part of the survey contained the survey experience questionnaire (scale 1-9 with three anchors and one open ended question). After the response was recorded, we explained the purpose of the study and thanked our participants. The validation option for each item was request response - participants were not required to answer each item in order to complete the questionnaire. If they chose to skip or accidentally skip an item, they however received a notification with an option to return to the page or continue without answering. **SURVEY FLOW** Survey flow was following: Upon entering the questionnaire, respondent was randomly assigned to one of six groups. Each of these group contained every vignette (Darrel, Gerald, Emma) and every condition (knowledge control, Gettier case, ignorance control) once. The order of the vignettes was randomly assigned. **DEPARTURE FROM INSTRUCTIONS** As noted and described above, we accidentaly swapped Gerald knowledge control vignette for ignorance control vignette, this was the most important departure from preprint instructions. We also departed from the instructions in form of visual analogue scale as described earlier and in recording of demographic questions. We decided not to ask about participant's ethnicity, race, country of origin and country of residence. The reason for this is the fact that our participants were expected to have exclusively European descent with Czech and Slovak nationality. Language and cultural differences between these nationalities are insignificant and the population is homogeneous (however we controlled for language differences). Therefore we assumed that asking about mentioned demographic question would be confusing for our participants. We also included a dichotomous question asking whether or not is the participant currently an university student in order to easily distinguish university and non-university participants. We also decided to measure gender as a close ended question with three categories (male, female, other) to prevent nonsensical responses. In our survey flow, participant also didn't always begin with Darrel vignette as suggested in the preprint. There was no pre-programmed randomization in the software we used and this survey flow was a compromise between our options and instructions which provided sufficient randomization.