Controlling for industries

[jamo222]

Hello,
I'm doing a master thesis. I'm regressing CSR performance on Financial performance by OLS and included control variables like sales for size, ROA for profitability etc.
Controlling for industries is important in that kind of study but as my sample is very small (90 observations) my first impression is that there is no sense to do it. Indeed, as my sample is really small, controlling by industries for example with sic code have little sens given the facts that there are a lot of different industries so that I will have only 2,3 firms by industries and the measure for the industry will be poor. I'm not sure about my reasoning.
Can somebody help me with that matter
Thank you very much

 

[Dason]

Have you considered using a larger grouping based on the sic? The codes themselves group themselves quite easily. https://en.m.wikipedia.org/wiki/Standard_Industrial_Classification

 

[hlsmith]

I agree with @Dason recommendation. If you have heterogenous effects between groups and you collapse them into a single group, results could be misleading.

 

[jamo222]

First of all thank you very much for your answers.
I considered your comments and selected the classification made by ICB (industry classification benchmark) which classify companies in 10 industrial group (9 for me as I don't take into account financial's companies). Do you think it's a good way to group my firms?