Compare the preceding four simple linear regression models to determine which model is the preferred model. Use the Significance F values, p-values for independent variable coefficients, R-squared or Adjusted R-squared values (as appropriate), and standard errors to explain your selection.
Regression1 | Regression 2 | Regression 3 | Regression 4 | ||||
Significant F Values | 2.31E-31 | 1.06E-36 | 2.70E-07 | 0.079243652 | |||
p-values for independent variable coefficients | 2.31E-31 | 1.06E-36 | 2.70E-07 | 0.079243652 | |||
R squared | 0.601403011 | 0.662202192 | 0.16417861 | 0.020666911 | |||
Standard Errors-y | 0.850103399 | 0.633569376 | 1.81204468 | 0.649109434 | |||
Standard Errors-x | 0.139580202 | 0.181911517 | 0.25245326 | 0.038379067 |
We prefer regressions with lower p-value and f-value. Both the
regressions have very low f-value and p-value, so both the
regressions are significant with significant independent
variable.
So until now we are unable to decide which regression to
prefer.
Now, we prefer the regression with lower standard errors.
Regression 1 has lower standard error for x while regression 2 has
lower standard error for y.
So we still can't decide which one to prefer.
Now, we know that a better regression will have high R squared
value.
Regression 2 has much higher R squared value than regression
1.
So regression 2 is preferred.
Thank you
Get Answers For Free
Most questions answered within 1 hours.