| A | B |
|---|
| r | correlation |
| residual | y - y-hat |
| influential point | outlier with extreme x-value |
| scatterplot | graph of y vs. x |
| residual plot | graph of residuals vs. x |
| Least squares line | y-hat = a + bx |
| explanatory variable | independent variable |
| response variable | dependent variable |
| semi-log transformation | exponential model |
| log-log transformation | power model |
| outlier | point that is an extreme y or x-value |
| def. of slope | amt. that the y-variable changes each time there is an increase of one in the x-variable |
| def. of y-intercept | value of the y-variable when x = 0 |
| extrapolation | using a regression line for prediction outside the range of the original data |
| prediction | main use of a regression line |
| formula for slope of regression line | r(sy/sx) |
| formula for y-intercept | y-bar - b(x-bar) |
| y-hat | predicted value of y |
| linear growth | increases by a fixed amt. |
| exponential growth | increases by a fixed % of the previous total |
| r^2 | % of variation explained by the least-squares regression |
| marginal distributions | row and column totals |
| Simpson's Paradox | the effect of lurking variables |
| two-way table | describes the relaitonship between two categorical variables |
| prospective study | subjects are followed for a long period of time |
| experiment | the only way to establish causation |
