Timeline for Test for paired samples with normal distribution
Current License: CC BY-SA 3.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 28, 2017 at 0:39 | history | edited | Glen_b | CC BY-SA 3.0 |
added 1165 characters in body
|
| May 27, 2017 at 10:35 | comment | added | Glen_b | @宓辰羲 No. The t-test is still correct, the z-test is approximate, but as the sample size increases, the t-test gets closer to a z-test. Ultimately in sufficiently large samples it doesn't matter which you do, but if you truly had normality there's no particular reason to use the z-test over the t-test no matter how big the sample size. With computers to do the calculation, there's no need to replace the t with an approximation. | |
| May 27, 2017 at 10:32 | history | edited | Glen_b | CC BY-SA 3.0 |
added 24 characters in body
|
| May 27, 2017 at 9:50 | comment | added | mcxmcx | Thanks very much for the response@Glen,@Michael. Both of your answers are very helpful for me! I made a mistake that in this case, I should use the paired t test rather than unpaired one. I have another question. For the large sample(>30 data) with normal distribution, is the z.test more appropriate than t.test? | |
| May 27, 2017 at 8:52 | history | edited | Glen_b | CC BY-SA 3.0 |
added 3 characters in body
|
| May 27, 2017 at 8:30 | comment | added | Michael R. Chernick | This is a good answer. I like that you emphasized that the signed rank test is the appropriate form of the Wilcoxon test to use with paired data. I think it would also be good to distinguish that the paired t test should be used rather than unpaired t test. | |
| May 27, 2017 at 6:45 | history | edited | Glen_b | CC BY-SA 3.0 |
added 290 characters in body
|
| May 27, 2017 at 6:20 | history | edited | Glen_b | CC BY-SA 3.0 |
added 19 characters in body
|
| May 27, 2017 at 6:08 | history | edited | Glen_b | CC BY-SA 3.0 |
added 357 characters in body
|
| May 27, 2017 at 5:47 | history | answered | Glen_b | CC BY-SA 3.0 |