I am looking for a way to convince my participants they "earned" their payout at the end of the experiment. However, I do not want any actual performance fluctuations, because that is beyond the scope of my research. The main thing I want to measure is their reactions to subsequent taxing of the income they receive. Therefore, I want to keep the payout as constant as possible across participants, while still having them think they earned their income through their own effort. My idea was to present them with some sort of multiple choice quiz about a very abstract subject, of which most people have little to no knowledge of. This allows me to convince participants that whichever answer they picked for a question (A, B, C or D) is either right or wrong. This way, the total income of each participant is fixed, while they still believe that what they earned is the result of their own reasoning. My question, then, is if something similar has been done before. I know of some papers that used abstract subjects, but still had variable pay.
There have been many studies in which false feedback was provided to participants (i.e., telling them that they had done well or poorly according to random assignment). Some of this goes back a long way (e.g., the Rotter level of aspiration task).
Your main practical challenge is to develop a task on which people can't tell whether they are doing well or poorly. I don't think a multiple-choice question about an arcane subject is a good choice, because most people will know they don't know the correct answer - and those who do know the correct answer will know that they do!
I think researchers usually employ problem-solving tasks that have a speed element, on which participants believe they are being compared with their peers (e.g., students at the same or a similar university). For example, you could have them work Ken Ken or Sudoku puzzles, solve anagrams, generate antonyms for words, etc. Although they will know how many they've solved, they will have no way of knowing how others have done, so your false feedback will be believable.
You also have an ethical challenge. Deceiving participants can be justified, but your IRB will want you to do so.
Couldn't the peer comparison element be implemented on a quiz as well? Puzzles or other problem solving tasks would be very difficult to program into my current questionnaire template. For example, I could tell participants that their payment for answering a question is based on their peers' performance. This eliminates the need for "arcane" quiz subjects.
Anyway, thanks a lot for your input! I've been struggling to find the right phrasing of the type of methodology I was looking for, and "false feedback" seems to be it. I've already found more articles about this now.
Sounds good. If you're going to use a multiple-choice quiz, I'd suggest making it on a subject that isn't highly specific & knowledge-based. One possibility would be to use a proverb interpretation task, or perhaps one asking people to abstract similarities between two concepts. Make them all fairly difficult, and make sure at least two of the "incorrect" options are highly plausible. That way people won't know for sure how many they got right.
After some deliberation, I've decided to go for a puzzle task. Participants are to complete a Sudoku puzzle as far as they can in 1 minute (or maybe less, yet to be determined). They then get payed according to their performance relative to their peers. This way, I can basically tell them they did better than, say 70% of other participants, and award them accordingly.
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion