Regression Vocabulary

Explanatory Variable

Independant, The thing we think causes (x axis)

Response Variable

Dependant, what we measure

Extraneous Variables

Any variable that affects your results but you are not measuring it

Outliers

Any point that does not follow the general pattern

Describing a relationsuip based on a graph you…

D - Direction (positive or negative)

U - Unusuals (outliers)

F - Form (Linear, nonliner)

S - Strength (strong, moderate, weak,)

+C - Context

Influencial Points

An outlier that when removed changes the relationship

High Leverage Points

An outlier with a very large (or small) x-value compared to every other point

Extrapolation

Predict for a value outside of the date set

To determine a Least Squares Regression Line you would…

Stat, Calc, 8 on calculator

Interpert Slope

For each additional (x w/ unit) the “predicted y” increases/decreases by __ on average

Interpert y-intercept

When “x=0” the Sy is about __ on average

Interpert an influencial point

The point is influencial because when removed the (strength, form, direction, slope) changes from “was” to “because”

Interpert High Leverage

The point is High Leverage because the x-value is bigger (or smaller) than all other points

Correlational Coefficent

|r| > .75 strong

|r| < .5 weak

any where in the middle is moderate

r<0 negative slope

r>0 posiitve slope

Interpert Correlation (r )

There is a “strength” “direction” "linear relationship between “x” and “y”

Coefficent of determination (r²)

__% of variation in “y” is explained in the LSRL with “x”

Standard Deviation for Rituals (S)

Typically each predicted “Y” is __ from the actual “y”

Residuals

mesures difference between the observed response and the predicted response

y - y^

obs - res

Interpert Residuals

The “y” was over/under predicted by [residual]

neg res - over predicted

pos res - underpredicted

How do we know if the linear model is appropriate?

No pattern to residuals is a good fit

Summary Statistics Formulas

Slope (b)

r (Sy/Sx)

Y-int (a)

y- b(x)

y= a+bx