Timeline for What test to use when a chi-squared assumption is violated but my contingency table only has one row?
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 5 at 15:12 | answer | added | kjetil b halvorsen♦ | timeline score: 2 | |
| Mar 5 at 15:09 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
| Mar 5 at 14:47 | history | edited | kjetil b halvorsen♦ | CC BY-SA 4.0 |
deleted 12 characters in body
|
| May 1, 2019 at 2:49 | vote | accept | OxL | ||
| Apr 30, 2019 at 23:49 | history | edited | OxL | CC BY-SA 4.0 |
added observed counts and explained expected counts
|
| Apr 30, 2019 at 23:16 | comment | added | OxL | Thanks for the comments. Frans, I am not familiar with it, from a brief googling it seems to be recommended for 2x2 and I can't see a method for 1x3 tables? | |
| Apr 30, 2019 at 9:42 | comment | added | Glen_b | Out of curiosity (if you're able to say), how do those expected values arise? | |
| Apr 30, 2019 at 9:30 | answer | added | Glen_b | timeline score: 7 | |
| Apr 30, 2019 at 7:29 | comment | added | BruceET | Chi-sq test may be OK. The risk would be that the chi-squared statistic may not have exactly a chi-squared distribution under $H_0.$ But there are ways to get an exact P-value. Can you give observed values? | |
| Apr 30, 2019 at 5:36 | comment | added | Frans Rodenburg | Are you familiar with Barnard's test? en.wikipedia.org/wiki/Barnard%27s_test | |
| Apr 30, 2019 at 5:15 | history | asked | OxL | CC BY-SA 4.0 |