| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | |||
| 5 | 6 | 7 | 8 | 9 | 10 | 11 |
| 12 | 13 | 14 | 15 | 16 | 17 | 18 |
| 19 | 20 | 21 | 22 | 23 | 24 | 25 |
| 26 | 27 | 28 | 29 | 30 | 31 |
Authors' Committee
Chair:
Matt Blackwell (Gov)
Members:
Martin Andersen (HealthPol)Kevin Bartz (Stats)
Deirdre Bloome (Social Policy)
Andy Eggers (Gov)
John Graves (HealthPol)
Rich Nielsen (Gov)
Maya Sen (Gov)
Gary King (Gov)
Weekly Research Workshop Sponsors
Alberto Abadie, Lee Fleming, Adam Glynn, Guido Imbens, Gary King, Arthur Spirling, Jamie Robins, Don Rubin, Chris Winship
Recent Comments
- Re: Control Groups for Breakfast, Revisited
By: Gary King - Re: Control Groups for Breakfast, Revisited
By: David Kane - Re: Control Groups for Breakfast, Revisited
By: brent
Recent Entries
- Dynamic Panel Models
- Greiner on "Exit Polling and Racial Bloc Voting"
- Choosing variances in general linear models
- Bill Support by Page Length
- Answering "why" questions
- Just in time for "Superfreakonomics"
- Airoldi on "A statistical perspective on complex networks"
- Happy Halloween
- Matching Markets
- Physics of politics
Categories
Blogroll
Brad DeLongCognitive Daily
Complexity & Social Networks
Developing Intelligence
EconLog
The Education Wonks
Empirical Legal Studies
Free Exchange
Freakonomics
Health Care Economist
Junk Charts
Language Log
Law & Econ Prof Blog
Machine Learning (Theory)
Marginal Revolution
Mixing Memory
Mystery Pollster
New Economist
Political Arithmetik
Political Science Methods
Pure Pedantry
Science & Law Blog
Simon Jackman
Social Science++
Statistical modeling, causal inference, and social science
Archives
- November 2009
- October 2009
- September 2009
- August 2009
- June 2009
- May 2009
- April 2009
- March 2009
- February 2009
- January 2009
- December 2008
- November 2008
- October 2008
- September 2008
- August 2008
- June 2008
- May 2008
- April 2008
- March 2008
- February 2008
- January 2008
- December 2007
- November 2007
- October 2007
- September 2007
- June 2007
- May 2007
- April 2007
- March 2007
- February 2007
- January 2007
- December 2006
- November 2006
- October 2006
- September 2006
- August 2006
- July 2006
- June 2006
- May 2006
- April 2006
- March 2006
- February 2006
- January 2006
- December 2005
- November 2005
- October 2005
- September 2005
Notification
Enter e-mail to [un]subscribe:
« Valid Standard Errors for Propensity Score Matching, Anyone? | Main | Another classroom demo: the scientific method »
23 March 2006
Control Groups for Breakfast, Revisited
A few months ago, I wrote an entry entitled The Value of Control Groups in Causal Inference (and Breakfast Cereal). It was a report on a fun experiment I did that worked well both in my daughter's kindergarten class and my graduate methods class at Harvard. There were a fair number of comments posted in the blog, and I also received dozens of other notes from parents and school teachers all over the country with many interesting questions and suggestions.
That correspondence covered four main points:
- Some people suggested a variety of interesting alternative experiments, which is great, but in designing these many forgot that you must always have a control group. That's the main lesson of the experiment: you often learn nothing without some kind of control group, and teaching this to kids (and graduate students!) is quite important.
- Some people didn't squish the cereal enough and the magnet didn't pick up the pieces. It will attract only when squished very well since the bits of iron are very small.
- People then asked why the cereal doesn't stick to the magnet without squishing it up. The reason is the same reason a magnet won't pick up a nail driven into a log, but it will pick up the nail if not in the log.
- Finally, most people asked for other experiments they could run with their kids. For that, which I'm writing up now, please tune in next time!
Posted by Gary King at March 23, 2006 6:00 AM