0

I am running an OLS regression in R studio but my variables have different scales. For example variables such as my dependent variable is the log return of a stock index(too small) and my covid and vaccination variables are too big. Is there any way I can report my results in a more readable form, I mean avoiding all those e- values.

Coefficients:
                     Estimate Std. Error t value Pr(>|t|)    
(Intercept)        -9.424e-05  3.076e-04  -0.306 0.759413    
DummyFinancial     -1.604e-03  9.691e-04  -1.655 0.098554 .  
DummyLock           9.241e-04  1.030e-03   0.898 0.369870    
DummyHousehold      3.465e-03  9.327e-04   3.715 0.000227 ***
DummyTravel        -3.430e-03  1.245e-03  -2.755 0.006084 ** 
DummyHealth        -8.503e-05  8.198e-04  -0.104 0.917433    
Covid.Cases        -1.080e-07  1.079e-07  -1.001 0.317370    
Deaths              2.348e-05  4.823e-05   0.487 0.626697    
Hospitalizations    4.646e-06  1.410e-05   0.329 0.741987    
Vaccination         3.089e-08  2.175e-08   1.420 0.156235    
stringency          1.707e-04  1.361e-04   1.254 0.210337    
STOXX50E            4.687e-01  8.725e-02   5.372 1.21e-07 ***
GDAXI               6.040e-02  8.517e-02   0.709 0.478556    
EXRUSA              1.806e-01  9.939e-02   1.817 0.069776 .  
EXREU               1.322e-01  1.292e-01   1.023 0.306742    
Volatility.Germany -1.779e-02  5.370e-03  -3.313 0.000994 ***

I tried to standardize my variables with the Z-score, but as soon as I did it my constant become as follows:

**(Intercept)         4.3181e-16  1.9641e-02  0.0000 1.000000**
Ben Bolker
  • 211,554
  • 25
  • 370
  • 453
Sebastian Merizalde
  • 3
  • 2
  • When you say they're too small or too big... what does that mean? – Bill O'Brien May 02 '22 at 14:56
  • this belongs on [CrossValidated](https://stats.stackexchange.com), but the immediate answer to your question is that the intercept of a set of centered numeric variables is zero, by default ... (your variables are all numeric because you have generated dummies by hand rather than using R's machinery to do so ... is there a reason why this is a problem for you?) – Ben Bolker May 02 '22 at 15:08

0 Answers0