ANOVA using time variables

#1
I am trying to find the difference between 3 groups to determine if the times between authorization to an event is different (longer, shorter). They are measured in hh:mm:ss.

Can you run an ANOVA using time variables. I never ran it before but assume it works. However, my data is not normal and my distributions are not equal. Transformations do not help much. Any idea what I can possible do?

You can view the data here if you wish:

https://docs.google.com/spreadsheets/d/1Fnh4_ko6CJey6q-wIm1sfDKVA42KMXM9MeqO-M3qtfY/edit?usp=sharing
 

bugman

Super Moderator
#2
First I would change the format of the reponse to hh:mm or even just to minutes and convert it back to hours for your write up, since you are just looking at differences in time to approve between groups and I assume the time of data

Yes you can use an ANOVA for this data. Time in minutes is just a continous response variable and you have multiple groups

Can you post some diagnostic plots of your residuals - are you sure the assumptions are not being met?