Faith based decision making

The US Bureau of the Census is set to reduce the sample for its critical Survey on Income and Program Participation (SIPP) by more than half, according to Lyndsey Layton in today’s Washington Post.

Reducing the size of the sample increases the level of statistical error. Although the reduced sample will still provide reliable estimates on a national basis, state level data for most states will no longer be available.

The Census Bureau, which oversees the survey, plans to reduce the number of people questioned nationwide from 45,000 to 21,000. The result will mean that detailed data will be generated for just three states -- California, Texas and New York -- instead of the more typical 31 states, said Preston Jay Waite, deputy director of the Census Bureau.

The SIPP is the only instrument providing data to support analysis of program effectiveness.

It is one of the most important surveys the government conducts -- the only large-scale measurement of the impact of Medicaid, food stamps, school lunches, unemployment and other safety-net programs for the poor… Public officials and researchers rely on those findings to determine how government programs are used, which are most effective and which are faltering.

Even Ralph A. Rector, senior research fellow at the conservative Heritage Foundation, opposed the cut.

"The worst outcome is if the money is insufficient for the SIPP and they don't continue" working on an alternative, he said. Then researchers don't get good survey data now or a reliable replacement later, he said.

But the Bush administration is firmly committed to faith based decision making. And they’ve been cutting all those programs that SIPP monitors, anyway, so why would anybody be interested in the impact?