Small sample size

[Sunita Rao]

I would be grateful for some guidance on which statistical model should be used for small samples. I have 51 firms-years i.e. 51 observations in my dataset.

[Carlo Lazzaro]

Sunita:
if you mean that you're dealing with a cross-sectional study with one wave of data only, assuming that your regressand is continuous, I would go for a parsimonious -regress-.

[Sunita Rao]

I have data for the year 2018, 2019, 2020, 2021 and 2022 for 10 companies. Only 1 of these 10 companies has data for 2023. Is that considered a cross sectional study with one wave of data?

[Carlo Lazzaro]

Sunita:
no, this is a panel dataset.
I would get rid of that company that has data for 2023 only and code a parsimonious model.
Then, I'd check whether a panel-wise effect does exist with such a small sample size.