econometrics - intro to multiple regression

Upload: madalina-dinu

Post on 06-Jul-2018

229 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    1/36

    Chapter 6Chapter 6

    Introduction to

    Multiple Regression

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    2/36

    Outline

    1. Omitted variable bias

    2. Causality and regression analysis

    3. Multiple regression and OLS4. Measures of fit

    . Sampling distribution of the OLS estimator 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    3/36

    Omitted Variable Bias

    (SW Section 6.1) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    4/36

    Omitted variable bias, ctd. 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    5/36

    Omitted variable bias, ctd. 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    6/36

    Omitted variable bias, ctd. 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    7/36

    Omitted variable bias, ctd. 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    8/36

    Te omitted !ariable bias "ormula#

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    9/36

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    10/36

    $igression on causalit% and

    regression anal%sis

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    11/36

    Ideal Randomi&ed 'ontrolled

    periment

    •   Ideal ! sub"e#ts all follo$ the treatment proto#ol % perfe#t#omplian#e& no errors in reporting& et#.'

    •   Randomized ! sub"e#ts from the population of interest are

    randomly assigned to a treatment or #ontrol group (sothere are no #onfounding fa#tors)

    •   Controlled ! having a #ontrol group permits measuring thedifferential effe#t of the treatment

    •   Experiment ! the treatment is assigned as part of thee*periment! the sub"e#ts have no #hoi#e& so there is no+reverse #ausality, in $hi#h sub"e#ts #hoose the treatmentthey thin- $ill $or- best.

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    12/36

    Bac* to class si&e#

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    13/36

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    14/36

    Return to omitted variable bias 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    15/36

    Te +opulation Multiple Regression

    Model (SW Section 6.,) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    16/36

    Interpretation o" coe""icients in

    multiple regression

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    17/36

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    18/36

    Te O-S stimator in Multiple

    Regression (SW Section 6.) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    19/36

    ample# te 'ali"ornia test score

    data

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    20/36

    Multiple regression in ST/T/

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    21/36

    Measures o" 0it "or Multiple

    Regression (SW Section 6.) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    22/36

    SER and RMSE  

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    23/36

    R , and  2 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    24/36

    R , and   2 ctd.2 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    25/36

    Measures of fit, ctd. 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    26/36

    Te -east S3uares /ssumptions "or

    Multiple Regression (SW Section 6.4) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    27/36

    /ssumption 51# te conditional mean o"

    u  gi!en te included X s is &ero.

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    28/36

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    29/36

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    30/36

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    31/36

    Te Sampling $istribution o" te

    O-S stimator (SW Section 6.6) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    32/36

    Multicollinearit%2 +er"ect and

    Imper"ect (SW Section 6.7) 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    33/36

    Te dumm% !ariable trap

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    34/36

    Perfect multicollinearity, ctd. 

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    35/36

    Imperfect multicollinearity  

  • 8/17/2019 Econometrics - Intro to Multiple Regression

    36/36

    Imperfect multicollinearity, ctd.