Panel data analysis with limited observation in one group and one year

[Cloudy Tech]

Hi everyone,

I need help with some questions below:

1. I have a dataset for 5 years: 2018, 2019, 2020, 2021 2022. However, one important variable is not available in 2020, can I impute the data for the 2020 to do panel data analysis? Or can I do 2 year panel data 2021 2022 analysis?

2. How can I do (2018, 2019) and (2021 2022) difference in difference DiD analysis if I can not use 2020 data - Should I merge each two years then treat as two groups as single period comparison?

3. In one of my variables, which has 4 categories, one category has limited observations (the nature of this group) can I still do all the analysis with xtreg, mlogit, ologit normally

Variable A
Category 1: 0.87%
Category 2: 42.71%
Category 3: 25.10%
Category 4: 31.32%

And the observations in category 1 is only 200 over 30,342 observation in general, each year have only about 50 observations in this category.

Thanks,
Cloudy

[Cloudy Tech]

Hi, anyone can help??

[George Ford]

What year is the treatment?

[Cloudy Tech]

I have had treatment in different years. The policies are in different years based on each state. Thanks.