Have Computers Messed Up the Unemployment Rate?

Mark Walker fills out an application at a job fair  in Detroit in April. A new academic paper suggests that the unemployment rate appears to have become less accurate over the last two decades as Americans have become less willing to respond to surveys. (Credit Joshua Lott/Getty Images) Mark Walker fills out an application at a job fair in Detroit in April. A new academic paper suggests that the unemployment rate appears to have become less accurate over the last two decades as Americans have become less willing to respond to surveys. (Credit Joshua Lott/Getty Images)
Apparently over the last two decades, according a Princeton study, Americans have become less willing to answer surveys. Trust is a big reason. Trust that the survey is not politically skewed or that personal info and responses are not being secretly stored and desseminated. I know I'm far more suspicious--who isn't?

But the study cited by the New York Times goes further than just social science, it finds a technical problem with the gathering of employment data that goes back to a 1994 change from paper-based to computer-based. The problem is that unemployed workers, who are surveyed multiple times, are more likely to respond to the survey early on and ignore the survey later. This ends up inaccurately weighting the later response, which skews the unemployment rate downward.
—Ron Pechtimaldjian, ThinkAdvisor
The Labor Department's monthly jobs report has been the subject of some wacky conspiracy theories. None was wackier than the suggestion from Jack Welch, the former General Electric chief executive, that government statisticians were exaggerating job growth during President Obama's 2012 re-election campaign. Both Republican and Democratic economists dismissed those charges as silly.
Reprints Discuss this story
This is where the comments go.