| Question |
Answer |
| b-values |
Unknown quantities in multiple regression. Tell us about the relationship between the outcome and each predictor. What degree each predictor affects the outcome if the effects of all other predictors are constant. |
| t statistic |
Measure of whether the predictor is making a significant contribution to the model. If the t test associated with a b value is significant (less than .05) then the predictor is making a significant contribution. |
| Beta - B |
Tells the number of standard deviations that the outcome will change as a result of one standard deviation change in the predictor. Provides a better insight into the importance of a predictor in the model. |
| VIF |
Assumption of no multicollinarity. Value less than 10 - no cause for concern. Average of VIF is substantially greater than 1 - regression may be biased. |
| R2 in Multiple Regression |
Measure of how much of the variability in the outcome is accounted for by the predictors. Useful measure of the substantive importance of an effect. |
| Skewness |
Lack of symmetry. |
| Kurtosis |
Pointyness. |
| Positively Skewed |
Frequent scores are clustered at the lower end and the tail points toward higher or positive positions. |
| Negatively Skewed |
Frequent scores are clustered at the higher end and the tail points toward lower more negative scores.
|
| Positive Values of Skewness |
Pile up of scores on the left.
|
| Negative Values of Skewness |
Pile up of scores on the right. |
| Platykurtic Distribution |
Many scores in the tails "Heavy Tailed" and is quite flat. Flat like a PLATEAU. |
| Leptokurtic Distribution |
Thin in tails and look pointy. LEAPT up in the air. |
| r - Pearson Correlation |
Value lies between -1 and 1. Standardized measure commonly used to measure the size of an effect.
|
| r - +/- .1 |
Small Effect
|
| r - +/- .3 |
Medium Effect
|
| r - +/- .5 |
Large Effect
|
| r = +1 |
2 variables are perfectly positively correlated.
|
| r = -1 |
2 variables are perfectly negatively correlated. |
| Multiple Regression |
Seeks to predict an outcome from several predictors.
|
| Adjusted R2 |
Measure of how well a model generalizes. Like the values to be close to R2.
|
| Durbin-Watson |
Measures whether the assumption of independent errors is tenable. The closer to 2 the better. Less than 1 or greater than 3 = ALARM.
|
| ANOVA Table for Regression |
Tests whether the model is significantly better at predicting the outcome than using the means as a "best guess." Look for values less than .05. |
| VIF |
Variance Inflation Factor - indicates whether a predictor has a strong linear relationship with the the other predictors. Value of 10 is worrisome (Myers, 1990) and if average is greater than 1, multicollinearity may be biasing the model. |
| Multcollinearity |
Exists when there is a strong correlation between 2 or more predictors in a regression model. This increases the probability that a good predictor of the outcome will be found non-significant and rejected from the model (a type II error).
|
| Type I Error |
Saying something is when it is not. |
| Type II Error |
Saying something is not when it is.
|
Beef Production
