Term
|
Definition
compares the entire function (model 1) with y=B0 (model 2), meaning it is testing the hypothesis
If the F is not significant, then we cannot say that either model 1 is any better than model 2. It is obtained by dividing the explained variance by the unexplained variance. |
|
|
Term
|
Definition
| (Total Sum of Squares) is the total deviations in the dependent variable. |
|
|
Term
|
Definition
| (Sum of Squared Error) is the amount of the SST that could be explained by the model. |
|
|
Term
|
Definition
| is the ratio SSR/SST (wrong equation?). It captures the percent of deviation from the mean in the dependent variable that could be explained by the model. Aka coefficient of determination. |
|
|
Term
|
Definition
| measures the proportion of the variance in the dependent variable that was explained by variations in the independent variables. Adjusted by the number of observations. |
|
|
Term
| Std error of the estimate |
|
Definition
| measures the dispersion of the dependent variables estimate around its mean. Compare this to the mean of the “Predicted" values of the dependent variable. If the Std. Error is more than 10% of the mean, it is high. |
|
|
Term
|
Definition
| The reliability of our estimate of the individual beta |
|
|
Term
|
Definition
| an estimated value based on current data |
|
|
Term
|
Definition
| when the z value is greater than 2, accept it |
|
|
Term
| major pitfalls in regression analysis |
|
Definition
| 1. Function form. 2. Association versus causation. 3. Measurement error. 4. Identification problem. 5. Simultaneous equations. 6. Multicollinearity. 7. Heteroscedasticity. 8. Auto correlation. |
|
|
Term
| Assumptions of a regression |
|
Definition
| ceteris paribus, reliable data, no pitfalls. |
|
|
Term
|
Definition
| Assumes a linear relationship, such as law of diminishing return. Can be fixed by trying to square or combine variables. |
|
|
Term
| Association versus causation pitfall |
|
Definition
| Be careful to assume causation. |
|
|
Term
| Measurement error pitfall |
|
Definition
| Not measuring what you say you are measuring. |
|
|
Term
| Identification problem pitfall |
|
Definition
| Can measure supply or demand but NOT both |
|
|
Term
| Simultaneous equations pitfall |
|
Definition
| DVs show up as IVs. E.g. Edu is a function of income is a function of edu |
|
|
Term
| Multicollinearity pitfall |
|
Definition
| secondary variables are related. Test IVs by running regressions between them. |
|
|
Term
| Heteroscedasticity pitfall |
|
Definition
|
|
Term
|
Definition
| An important variable is missing. Durbin Watson test. |
|
|
