# Should you introduce new notation for every regression?

I am currently writing an empirical paper which reports on the results from estimating a number of regression models.

Question: is it standard to introduce new notation for each of these models, or should one stick to $$\beta_0$$, $$\beta_1$$, etc. every time? I am trying not to re-use notation but am running out of Greek letters... I also find introducing the new notation a little pointless since I never actually use the notation to do any symbolic manipulations.

NB I do realise that this is not a question about statistics per se and do apologise for this in advance!

• People will sometimes capture all of the independent variables with a single variable, like Gamma or something, and call it a vector of independent variables. This way they just write down one thing. – John Stud Mar 4 at 13:15
• If the issue is using the same model $Y = X\beta$ but with different predictors, then I would expect to see comparisons naming the variables. In fact, depending on the readership, reminding them of what a linear regression is may not be needed. After all, very many papers cite means and medians without feeling any obligation to remind readers of their definitions. (I recall a university where a course Maths for X turned out to be too demanding for some and so another course Elementary maths for X was added. According to legend, the teacher started by checking familiarity with $+$ and $-$.) – Nick Cox Mar 4 at 14:21
• Addition is not a straightforward concept :) math.stackexchange.com/questions/15869/how-is-addition-defined – afreelunch Mar 4 at 14:51
• Sure, but that is not the point of the story. – Nick Cox Mar 4 at 16:28

If you want to distinguish the $$\beta$$ for the same independent variable, but in different models, you could use the double indices like : $$\beta_{i,j}$$ for the $$i$$th independent variable for the $$j$$th model.
This is an easy question and just using the notation $$\beta_0,\beta_1, ...$$ will be fine for you. This is because the number of the letters is finite. Especially Greek letters, and most of the good letters are taken :).