Primer On Imputing Census Data
Now that the Census Bureau can't use sampling, they've turned to 'imputing.' The differences are intriguing.
The Wall Street Journal presents a fascinating tutorial and analysis of the impact of "imputing" people and responses. The scope of imputing is vast -- acccording to the Journal, statistics for 2000 total includes 5.77 million people the Census Bureau believes exist but didn't actually count. "When it received no answers from what it believed were occupied addresses, the bureau simply directed its computers to "impute" people, based on various clues, including how their neighbors responded. In some cases, officials acknowledge, they added such imputed residents to the count when they weren't even certain there was a home at a given address. Imputed Americans made up more than 2% of the official 2000 census. In a few states, they made up more than 3%." Note: Registration required.
- Login or register to post comments
- Email this page
- CO2 'Domes' Pose Problems for Cities - Mar 18, 2010
- Finding Solutions for the Colorado River Delta Problem - Mar 17, 2010
- 'Sea-Change' Coming To Transportation Planning - Mar 16, 2010
- The High Costs of America's Aging Water Infrastructure - Mar 16, 2010
- Charting Public Data Via Google - Mar 15, 2010




















