Regression.txt

 March 13, 2003

 
 
 

Put Bush-Gore states in Tucson dataset. 

 

Frank: Peer-reviewed. What can you trust?   Special study for the occasion. 


  
 

g604, IUg604g604

 For laptop: FUNCTION F7 cycles through the RGB stuf-- both screens,

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

 FIRST HOUR:  REVIEW AND MORE ON REGRESSION


Yesterday:
 1. Correlation
 2. Regression line (two concepts)
 3. Various terms: show the slide from yesterday
 . 
   That was a great session, I thought.  Mention hte Davis article--- what is
important is the questions, not your plan. Severe wear and tear on you brains,
though--- my plan had been to spend some time on the Japanese judicial system,
partly to provide some nonrigorous relief.


 Plan for today. 

 
 
     Show them how to use Excel to:

0. Input the data.  Start from an ascii data set.  

 1. Graph the data

 2. Find correlation matrix

 3. Do the simple regression, with and without intercept

 

Small Effect size.  R2. Not regression.



 Then do the law firm regressions with less data.

 

 
 Components of r applet. Simple correlation.  Dispersion of data useful.

Restricted range applet: Strength, SAT. Simple correlation. R2. Dispersion of
data useful.


Heterosked.

Autocorrelation


XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

SECOND HOUR: INFO IN DATA.  LIMITED DEPENDENT VARIABLE. DATA PROBLEMS.


  Judges in Japan. Institutions from Chapter 1.

  Do the judges example from Ramseyer and Rasmusen.

Dependent variable a dummy variable

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

 THIRD HOUR 

Skip ahead to Social data. 
 The problem: the idea of social capital, and measuring it. 
 Wallet study. Measuring social capital. 

Use finance data and let them construct their own regressions. Stata.



Data sources, from my data page.  

 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

If there is extra time:  Go to La Porta. Unlikely. 



 

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX