1 year ago
#64524
Navot Naor
Running a robust mixed model ANOVA using statsmodel anova_lm
I have a data set of users behavior which I want to run series of ANOVA's on, based on there outcome of assumption checks. For situations where the assumption of homogeneity isn't met I would like to conduct a robust mixed model ANOVA.
My data is structured like this:
| user_id | Groups | test_id | avg_payment | variable | value | Condition |
|---|---|---|---|---|---|---|
| 21000010749909 | control | 1 | 0 | avg_daily_net_rev | 0 | before |
| 21000021190553 | control | 1 | 0-9 | avg_daily_net_rev | 80 | before |
| 21000021497845 | control | 1 | 9-40 | avg_daily_net_rev | 48 | before |
| 21000020054046 | test | 1 | 40-80 | avg_daily_net_rev | 14 | before |
| 21000021354449 | control | 1 | 80-100 | avg_daily_net_rev | 7 | before |
| 21000010749909 | control | 1 | 0 | avg_daily_net_rev | 0 | after |
| 21000021190553 | control | 1 | 0-9 | avg_daily_net_rev | 76 | after |
| 21000021497845 | control | 1 | 9-40 | avg_daily_net_rev | 30 | after |
| 21000020054046 | test | 1 | 40-80 | avg_daily_net_rev | 48 | after |
| 21000021354449 | control | 1 | 80-100 | avg_daily_net_rev | 71 | after |
I would like to compere values between the 2 groups (control and test) and the 2 times (before and after).
What is the correct why to set it up using statsmodels anova_lm?
Thank you
statsmodels
anova
mixed-models
robust
0 Answers
Your Answer