Skip to main content
5 events
when toggle format what by license comment
Mar 25, 2021 at 9:45 vote accept dsbr__0
Mar 25, 2021 at 4:06 answer added Thomas Bilach timeline score: 3
Mar 23, 2021 at 23:17 comment added dimitriy You would use regression like this.
Mar 23, 2021 at 17:16 comment added Thomas Bilach If you want to do this by hand, then yes, you should look at the group averages. But once we obtain more groups and more time periods, then I see no reason to work this out manually. Once we move beyond the two-by-two case, then you should run this in software using a regression formulation; it will give you your standard errors—for free.
Mar 23, 2021 at 11:25 history asked dsbr__0 CC BY-SA 4.0