# Regression coefficient interpretation

#### Torben

##### New Member
Hi,

i have run a regression to estimate the impact of couple of variables like growth rate, company size or leverage on the profitability of a firm.

I know that if e.g. the regression coefficient for growth is 0,5 a 1% increase in growth rate would yield a 0,5% increase in profitability if profitability and growth rate are measured in %, too.

My question is: can i also compare the characteristics of a company relative to the average of all comparable other companies?

For example if the average growth rate of a couple of similar companies is 5% and the growth rate of a certain company is 6%: would it be appropriate to conclude, that the profitability should be higher by 0,5 as well? i don't know in how far this interpretation is comparable to the standard if x increases by ... y changes by ...

#### noetsi

##### Fortran must die
do you mean .5? I don't know what 0,5 is.

Regression shows the change in Y given X controlling for all other variables for the whole data set not for a specific case. I have never seen it used the way you are. Which does not mean you can't I just have not seen it used that way.

#### Dason

**** Americans... not realizing that "," is used as the decimal in other countries.

#### jmyles

##### New Member
Your question is a tad vague, but it sounds like you should create another linear regression that will model the average of all other comparable companies. That way you can look at the coefficients of each model and compare that way. If you are trying to see if they're statistically significantly different from each other, that will take a little more work.

#### noetsi

##### Fortran must die
**** Americans... not realizing that "," is used as the decimal in other countries.
I sure never realized that or came across that in articles from other countries. Thanks dason.