2020.Year1 omitted because of collinearity.

[Jasmin Music]

Hi,

I performed a regression model to check the impact of covid19 on school children´s grades. I control for year fixed effects. The Covid dummy received a 1 if it was in year 2019/2020 and 2020/2021.
But WHY is my year 2020 omitted due to collinearity?
(I also attached my correlation table)

How could I explain this issue tomorrow in my defense?

Thanks a lot in advance and best regards

Jasmin

[William Lisowski]

Because 2020.Year1 is perfectly predicted by the combination of COVID19Dummy and 2019.Year1

• if COVID19Dummy==0 then 2020.Year1==0
• if COVID19Dummy==1 and 2019.Year1==0 then 2020.Year1==1
• if COVID19Dummy==1 and 2019.Year1==1 then 2020.Year1==0

[Jasmin Music]

Hi William thanks for your response. I still try to understand it. But does this cause any problem in my model?

[Carlo Lazzaro]

Jasmin:
as an aside to William's helpful advice, you might be interested in taking a look at

https://en.wikipedia.org/wiki/Dummy_variable_(statistics).
And no, -2020.iYear- omission does not cause any issue with your regression.