GoConqr has a new look! Let us know what you think.

Flashcards

Partial Regressions

Robert Gray English 12 cards

Cards in this deck (12)

  1. 1Partial regression coefficents: B0 =predicted y when both x1 and x2 = 0
  2. 2Partial regression coefficents: B1 =predicted change in y as x1 increases by one unit holding x2 constant
  3. 3Partial regression coefficents: B2 =predicted change in y as x2 increases by 1 unit controlling for x1
  4. 4Rx1x2 (correlation)R^2i.12 … when we have more than one predictor, degree to which one x is correlated with all other predictors of the model
  5. 5If x’s are uncorrelated thenyhat = b0 + b1x1 + b2x2
  6. 6Yhat = b0 + b1x1, standardized ZYhat= beta1Zx1
  7. 7Beta1 =Ryx1
  8. 8Yhat =b0 + b1x1 + b2x2 + rx1x2 = 0
  9. 9ZYhatBeta1Zx1 + Beta2Zx
  10. 10Ry1 , … Ry2 Don’t generalize well, why?unstandardized are not biased, standardized are biased, but standardized are nice because you can compare them to other standardized units, (i.e. Beta is bigger = effect is bigger)
    Can interpret them without knowing much about the units of X, and is not dependent on the units of measurement.
  11. 11When do you want to use standardized vs unstandardizedboth are best, if you want to use it applied, unstandardized is better, to compare to other regression variables standardized is better
  12. 12b1 =beta1 SDy/SDx1

Information

Description
Flashcards on Partial Regressions, created by Robert Gray on 12/12/2016.
Created by
Robert Gray
Views
1
Rating

More on Multiple Regression

Feedback

What is it about?

What do you think of the new look?

How do you like it?