APPENDIX A
Causal Correlations and the Correlation Coefficient
When asking the Three Questions, you need some very basic statistical capability you can learn right here—nothing fancy. You can do wonders with a correlation coefficient and an R-squared. With these two analytical tools, you can credibly disprove, in many instances, that two events have any connection to each other. It’s easy. All you need is an Internet connection and Excel.
To start, let’s get some data to compare. For an easy exercise, we can see how much one stock is correlated to the market over 10 days. (Ten days isn’t enough time to tell you anything about anything but will give us short columns of data to work with.)
Step 1
Go to Yahoo! Finance (http://finance.yahoo.com). (If you’re Internet savvy, feel free to use whatever source you’re comfortable with. Just be sure you know how to download or copy and paste into Excel.)
- Click on the link to the S&P 500, which is featured prominently—usually across the top of page.
- Click on the link for “Historical Prices”—on the left-hand side of the page.
- Select “Daily” prices, select a short time frame (use any time frame you like, just choose the same time period during Step 2—I used January 1, 2006, through January 10, 2006), and click “Get Prices.” Now select “Download to Spreadsheet.”
- An Excel spreadsheet will pop up with index data for the dates in question.
- Copy and paste the “Date” and the “Adjusted Close” columns into a new Excel spreadsheet page. You don’t need ...