Jobs and Votes

• Review for Unit II Exam
• Review of Unit II Homework Problems.

• Correlation and Regression
  New in 2011 -- based on R program.

• Multiple Regression
  New in 2012 - based on R program and the data file:
  • GasSim.Rdata
    Right click to download; load into R with load("GasSim.Rdata")

• Chapters 8 and 9. Correlation

  • Scatterplots , by Philip B. Stark, University of California at Berkeley.
  • Correlation and Association, by Philip B. Stark, University of California at Berkeley.

  • Correlation and More Correlation , Handouts by Michael Wichura, University of Chicago.
  • Correlation and More Correlation , Notes by Rebecca Nugent, Carnegie Mellon.

  • Review Exercises for Chapter 8 -- Solutions, by Tom McGahagan.
  • Review Exercises for Chapter 9 -- Solutions, (incomplete) by Tom McGahagan.

• Obama's Magic Number? 150,000 Jobs per Month by Nate Silver, Five Thirty Eight blog in the New York Times. February 3, 2012.

• Red State, Blue State, Rich State, Poor State Andrew Gelman summarizes his book by the above title, illustrating the ecological fallacy. Gelman's book of that title studied the relation between income and political preference: wealthier states vote Democratic, but wealthier individuals vote Republican. See also his lengthier article, Rich State, Poor State, Red State, Blue State: What's the Matter with Connecticut? by Andrew Gelman, Boris Shor, Joseph Bafumi and David Park, Quarterly Journal of Political Science, 2007, 2, 345-367. and his own website discussing statistics. The Oct. 16 2009 post is especially interesting.

• Anscombe.R Why you should always plot your data. The Anscombe data is built into R; the file linked above will plot the data. You should RIGHT-CLICK on the link and save the file to your R directory. Run the file with the command

• Ecological.Correlation.R Afile which load two others: US.Smokers.tab and CA.Smokers.tab . You should download all 3 into your R directory, plot and take correlations of smoking and death by US state, then by California county. Follow the directions after you run the Ecological.Correlation.R file with >>> source("Ecological.Correlation.R")

• From Correlation to Regression
  Links to height and weight data, to illustrate basic calculations.

• Chapters 10,11 and 12. Regression
  • Regression and The RMS Error for Regression and The Regression Line , Handouts by Michael Wichura, University of Chicago.
  • Regression Method , Regression Errors and Regression Line , Notes by Rebecca Nugent, Carnegie Mellon.
  • Regression and
  • Errors in Regression , by Philip B. Stark.

  • Review Exercises for Chapter 10 -- Solutions, by Tom McGahagan.
    --- Worked regression examples
  • Review Exercises for Chapter 11 -- Solutions, by Tom McGahagan.
  • Review Exercises for Chapter 12 -- Solutions, by Tom McGahagan.

  • Summary of Economic Projections for the June 22-23 meeting of the Federal Open Market Committee. Note especially the fan of possibilities around the projections for GDP growth, unemployment and core inflation.

  • Prospects for Inflation from the Bank of England's Inflation Report, August, 2011. The Bank of England was the first to use the type of inflation chart that the Fed tried above. It became known as the "River of Blood" for obvious reasons.See charts 5.6 and 5.7 on page 39, and look ahead to the probability charts on page 41.

• Methode der kleinsten Quadrate has a nice picture of Gauss. The English wikipedia article, Least Squares is not as thorough, and a bit too technical.
• Gauss's First Argument for Least Squares by William C. Waterhouse, Archive for the History of the Exact Sciences, v.41, no.1 (March, 1990) is a clear explanation of the basic logic of least squares.
• Regression to the Mean , from The Fifth Down, NFL week 8 Game Probabilities, Oct. 29, 2009. See the guest poster's own blog, AdvancedNFLStats.com and his description of his regression methodology which includes a nice example of his caculation of probabilities. A more detailed example of his methodology is also available.

• Reading a Regression, by Tom McGahagan, University of Pittsburgh at Johnstown.

• Reading a Multiple Regression
  Estimating gasoline consumption by looking at both price and income.

• Sample examination for Unit II. -- Answers
  My apologies that the last question on the exam was not complete. The questions were, given the regression output, to:
  1. Compute a point prediction if prime rate = 10 percent.
  2. Compute a 95 percent confidence interval prediction
  3. Use the given data to discuss any reservation you might have: especially look at the R-squared, the Standard error of the regression, and the standard errors and t-statistics on the coefficients.

  ---

  Retun to Statistics Main page

  ---