Those who know someone who died in the Iraq War or 9/11 terrorist attacks are less likely to approve of President Bush's performance in office than those with no such connection, according to new research from the University of California, Davis. The pattern holds true for Republicans as well as Democrats, conservatives as well as liberals, and across all races, ages, education levels and incomes.

The research appears in the August issue of American Sociological Review, the journal of the American Sociological Association.

"The notion of blaming one's leader for the death of a family or community member from a terrorist attack or war might seem odd at first," said UC Davis political science professor Scott Sigmund Gartner, the study's author. "But a personal tie to a victim converts abstract, distant costs of international violence into a vivid, personal and profoundly emotional experience, one that has clear, strong and consistent political implications."

Gartner arrived at his conclusions by analyzing results of two large public opinion polls. One was a 2006 Gallup survey that asked a national sample of Americans about their ties to soldiers serving or killed in Iraq. The other was a 2001 Field Poll that asked Californians whether they had lost a friend, family member or business associate in the 9/11 attacks. Both polls also asked about party affiliation, liberal versus conservative political outlook, and support for the president.

Gartner is the author of several earlier studies that looked at war casualties and public support for elected officials. The most recent, published in the April issue of the journal PS: Political Science, showed that members of Congress whose districts faced a disproportionate number of casualties in the weeks before the 2006 midterm elections were punished at the polls. In another study, published in February in American Political Science Review, Gartner argues that public support for elected leaders diminishes as both recent casualties and casualty trends increase.