Timeline for Comparing the distribution of few integers
Current License: CC BY-SA 4.0
18 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 7, 2018 at 20:45 | comment | added | BruceET | @JosuéOrtega: Glad to hear of your progress with this! Please let's not get too far into analysis of fake data I used as an exmp in my Answ, and focus on your real data. My experience with paired Likert data is one can't rely on more precision than whether Subject Nr. 15 liked A or B better with maybe some crude idea how much better. In my (continuing) view, such data is usually best handled by a sign test unless you have great faith in small diff's in opinion. // Did a permutation test w/ diff in means as metric, P-val aprx 0.001, but not sure means are meaningful summary for real dta. | |
| Aug 7, 2018 at 20:24 | comment | added | fox | Thank you all. I conducted a signrank test in Stata. The outpout of signrank V2ACC= VLC was Ho: V2ACC = VLC z = -9.332 Prob > |z| = 0.0000 I think this must be a two-sided test, but just wanted to confirm this? | |
| Aug 7, 2018 at 12:57 | comment | added | Scortchi♦ | ... Second, to rank differences you need to be able to take differences - getting around this by first transforming "very good" &c. into ranks isn't always what you'd want to do. The sign test doesn't require a decision about whether '1' for A & '3' for B is a difference of larger or smaller magnitude than '5' for A & '4' for B. | |
| Aug 7, 2018 at 12:57 | comment | added | Scortchi♦ |
An exact Wilcoxon signed-rank test in the presence of ties is provided by the exactRankTests package for R. (For your example it should result in exactly the same p-value as the sign test.) Nevertheless, there are a couple of reasons one might prefer a sign test, though I think you've not quite put your finger on them. First, the null hypothesis for the Wilcoxon signed-rank test is that the ranked differences are symmetric around nought - but asymmetry needn't result from A's being better than B (or vice versa) in any sense. ...
|
|
| Aug 7, 2018 at 8:04 | history | edited | BruceET | CC BY-SA 4.0 |
deleted 25 characters in body
|
| Aug 7, 2018 at 7:53 | history | edited | BruceET | CC BY-SA 4.0 |
added 263 characters in body
|
| Aug 7, 2018 at 7:48 | history | edited | BruceET | CC BY-SA 4.0 |
added 263 characters in body
|
| Aug 7, 2018 at 1:01 | history | edited | BruceET | CC BY-SA 4.0 |
deleted 9 characters in body
|
| Aug 7, 2018 at 0:52 | history | edited | BruceET | CC BY-SA 4.0 |
added 329 characters in body
|
| Aug 7, 2018 at 0:35 | history | edited | BruceET | CC BY-SA 4.0 |
added 3 characters in body
|
| Aug 7, 2018 at 0:28 | history | edited | BruceET | CC BY-SA 4.0 |
added 727 characters in body
|
| Aug 7, 2018 at 0:22 | comment | added | Alexis | Stata's signed rank test implementation adjusts for both zeros and ties out of the box. Article on it: Sribney, W. M. (1995). Correcting for ties and zeros in sign and rank tests. Stata Technical Bulletin, 26:2–4. | |
| Aug 7, 2018 at 0:19 | comment | added | BruceET | True, but here there are too many ties and 0's for the signed rank test. | |
| Aug 7, 2018 at 0:17 | comment | added | Alexis | The sign test hemorrhages statistical power relative to the sign rank test because it discard information about the (ranked) magnitude of the differences. | |
| Aug 7, 2018 at 0:06 | history | edited | BruceET | CC BY-SA 4.0 |
added 68 characters in body
|
| Aug 6, 2018 at 23:50 | history | edited | BruceET | CC BY-SA 4.0 |
added 408 characters in body
|
| Aug 6, 2018 at 23:44 | history | edited | BruceET | CC BY-SA 4.0 |
added 408 characters in body
|
| Aug 6, 2018 at 23:23 | history | answered | BruceET | CC BY-SA 4.0 |