Timeline for Is it possible to change a hypothesis to match observed data (aka fishing expedition) and avoid an increase in Type I errors?
Current License: CC BY-SA 3.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 26, 2020 at 13:47 | comment | added | Sam | @jona link broken | |
| Mar 17, 2015 at 22:47 | answer | added | gung - Reinstate Monica | timeline score: 5 | |
| May 28, 2014 at 16:40 | history | tweeted | twitter.com/#!/StackStats/status/471692610579742720 | ||
| May 28, 2014 at 12:51 | comment | added | keshlam | Data may cause you to change your hypothesis... but in that case you need to start gathering new data from scratch to confirm the new hypothesis. | |
| May 27, 2014 at 17:41 | history | edited | post-hoc | CC BY-SA 3.0 |
deleted 25 characters in body
|
| May 27, 2014 at 16:22 | comment | added | a11msp | I'm wondering whether you'll also find this paper relevant to your question: stat.columbia.edu/~gelman/research/published/multiple2f.pdf. It's not exactly on the same subject, but it addresses one aspect of it. | |
| May 27, 2014 at 16:04 | comment | added | post-hoc | @jona, that's a great paper. I've already read papers by both Ioannidis and Schooler, but Simmons et al wonderfully illustrates the problem. | |
| May 27, 2014 at 16:01 | history | edited | post-hoc | CC BY-SA 3.0 |
added 528 characters in body; edited title
|
| May 27, 2014 at 15:17 | comment | added | Wrzlprmft | Taking another point of view on what has already been said: The essence of the scientific method is to make hypotheses and then try to falsify them for that they may become theories (if the falsification fails). Going on a fishing expedition is a valid way to find hypotheses that are worth falsifying in a later experiment, but you can never make and try to falsify a hypothesis in one go. In particular, if you are open to adjusting your hypothesis, you are not trying to falsify it anymore. Instead, when you adjust, you are falsifying your unadjusted hypothesis and form a new hypothesis. | |
| May 27, 2014 at 13:23 | answer | added | Floris | timeline score: 25 | |
| May 27, 2014 at 11:12 | history | edited | post-hoc | CC BY-SA 3.0 |
added 1 character in body
|
| May 27, 2014 at 9:59 | answer | added | Peter Flom | timeline score: 57 | |
| May 27, 2014 at 9:57 | comment | added | jona | Read this: people.psych.cornell.edu/~jec7/pcd%20pubs/simmonsetal11.pdf | |
| May 27, 2014 at 9:45 | history | asked | post-hoc | CC BY-SA 3.0 |