Timeline for Comparing two means, finding critical value
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 9, 2016 at 13:31 | comment | added | Ian_Fin | True, but if we change "right on the midpoint" to "quite near the midpoint" I'm not sure that my concern doesn't stand. | |
| Aug 9, 2016 at 13:16 | comment | added | David C | The variables are continuous, so the probability of any given sampled value being exactly equal to the midpoint is zero | |
| Aug 9, 2016 at 13:14 | comment | added | Ian_Fin | Would this still make sense if there was unequal variance? I haven't thought this through fully, but an observation sitting right on the midpoint of two non-overlapping groups with unequal variance would seem more likely to come from the group with greater variance (because it would be fewer standard deviations from the mean of that group) | |
| Aug 9, 2016 at 13:11 | comment | added | David C | If there is no overlap, then this is the best-case scenario. In this case, I would recommend using the midpoint between the highest value of the lower group and the lowest value of the higher group | |
| Aug 9, 2016 at 13:08 | comment | added | Ian_Fin | How would you operationalise the least amount of overlap? If there was no overlap between the two groups then what would you set as the threshold? | |
| Aug 9, 2016 at 12:58 | history | answered | David C | CC BY-SA 3.0 |