Damage Assessment

I apologize for this question being exceedingly vague, but I'm not sure of the correct terminology.

After running a regression analysis, I have found that several factors adversely affect an employees salary. One possible being tenure, or age. The next step for me to complete my reports is to create a "Damage Assessment". By that, I mean I want to know by how many dollars does this variable affect salary.

Does anyone have any ideas?


TS Contributor
The regression analysis should give you the coefficients for each input variable and will tell you how much the output variable (salary) will change with each unit change of the input variable.
Yes, look at the value of the unstandardized slope for the variable age in the output of the regression analysis. If you reconstruct the regression equation by summating the slopes and the constant, you could probably get more useful information for this damage assessment.
Thanks JohnM and JamesMartinn for all your assistance! I would like to be sure that I am reading this right below.

Here is the regression analysis:

Estimate Std. Error t value Pr(>|t|)
(Intercept) 51414.96 4403.51 11.676 < 2e-16 ***
DEPT 6950.10 5781.93 1.202 0.233
EDUCA -2211.08 3193.60 -0.692 0.491
AGE -83.91 97.99 -0.856 0.395
TENURE 480.94 93.90 5.122 2.56e-06 ***
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 7939 on 70 degrees of freedom
Multiple R-squared: 0.3041, Adjusted R-squared: 0.2643
F-statistic: 7.646 on 4 and 70 DF, p-value: 3.599e-05

> rdata

And here is the generalized data:

lm(formula = SALARY ~ RACE + GENDER + AGE + TENURE)

51414.96 6950.10 -2211.08 -83.91 480.94

So am I saying that for each level of change in age, that will have an impact of -83.91 on the salary?


TS Contributor
Yes, but the standard errors are huge. On average, the -83.91 would be the age impact on salary, but with a huge error like this, the model is not useful in predicting individual salaries....