If you have three variables A, B, C and you want to test whether the hypothesis is that everyone focuses on C, can the expected outcome in a chi-square goodness of fit test then be A: 0 B:0 and C: 100 (or whatever sample size you have) ?
I did my survey in line with the problems in the paper of Heath, Larrick, Wu "Goals as Reference Points" (they take prospect theory value function as an explanation for why people work harder with high goals than low goals i.e. loss aversion and diminishing sensitivity). Here one example where I'm not sure how they statistically tested this:
Charles and David both follow workout plans that usually involve doing 25 sit-ups.
One day Charles sets a goal of performing 30 sit-ups. He finds himself very tired after performing 34 sit-ups and, at most, has the energy to perform 1 more.
David sets a goal of performing 40 sit-ups. He finds himself very tired after 34 sit-ups and, at most, has the energy to perform 1 more.
Who will work harder to perform the 35th sit-up?
N=73; Charles: 18% and David: 82%
Then they state: Our subjects reliably predicted that David would exert more effort than Charles (p < .001 by Chi Square). (due to loss aversion of the prospect theory value function)
How did they get the p-value? They don't state this in the paper but they practically say that since 82% chose Davide this implies that loss aversion takes places and people work harder when they are below the goal. I need to prove this similarly but don't know how.
Hi, Clara:
Your question implies a confirmatory hypothesis, but this is something which chi-square cannot handle, as this test is inherently exploratory.
The minimum expectation (per cell) for chi-square is five:
Article The Minimum Expectation in X Goodness of Fit Tests and the A...
In theory, chi-square can handle that by accepting a null hypothesis. However, that would require the computation of a Type 2 Error. In practice, the only way it can be done is to identify a specific level of relationship and compute Beta (the probability of a Type 2 Error) from that. Otherwise, the you would have a beta distribution.
Thanks! Do you have any other recommendation what statistic test I could use? I'm super lost if I can even use my survey data. I kind of did my survey similar to the problems of reference points in the article by Heath, Larrick, Wu "Goals as Reference Points" and they stated reliability in their paper with (p
You don't want a statistical test. If your hypothesis is exactly 0, 0, 100%, then this is a bold conjecture in the Popper sense, which is good. COnduct you study. If you get any in the first to categories then you reject this hypothesis. This is the common intro to logic example of disproving all swans are white by showing one black swan.
I'm honestly not sure what my hypothesis is. I go in line with the paper by Heath et al. - here they have this problem like I stated above:
Charles and David both follow workout plans that usually involve doing 25 sit-ups.
One day Charles sets a goal of performing 30 sit-ups. He finds himself very tired after performing 34 sit-ups and, at most, has the energy to perform 1 more.
David sets a goal of performing 40 sit-ups. He finds himself very tired after 34 sit-ups and, at most, has the energy to perform 1 more.
Who will work harder to perform the 35th sit-up?
N=73; Charles: 18% and David: 82%
Then they state: Our subjects reliably predicted that David would exert more effort than Charles (p < .001 by Chi Square). (due to loss aversion of the prospect theory value function)
How did they get the p-value? They don't state this in the paper but they practically say that since 82% chose Davide this implies that loss aversion takes places and people work harder when they are below the goal. I need to prove this similarly but don't know how.
You can ask the authors of the paper which you cite all of your theoretical and analytic questions:
https://www.researchgate.net/search.Search.html?type=publication&query=Goals%20as%20Reference%20Points
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size