Timeline for Should I use a one-sided or a two-sided test when deciding to deploy new system?
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 15, 2021 at 15:24 | history | edited | Noah | CC BY-SA 4.0 |
edited body
|
| Jul 15, 2021 at 12:28 | comment | added | Henry | I agree with your second paragraph. The problem with the first is that some researchers look at a two-sided $p$-value, see it is slightly not-significant at $5\%$ and then develop arguments for using a one-sided test. Instead they should have done some pre-test power analysis and considered their costs of Type I and II errors to decide their critical region before they looked at the data. | |
| Jul 15, 2021 at 2:49 | history | answered | Noah | CC BY-SA 4.0 |