But when the researchers used statistical methods to focus on short-term changes during this time period, they found that infant mortality rates were about 3 percent higher during years in which a Republican was president, compared with the years in which a Democrat was president.

The findings held even after the researchers took into account factors that might affect infant mortality, such as unemployment rates, smoking rates, abortion rates and measures of national education and income level. And the researchers did not consider the first year of any president’s term, they said, because a president’s policies would not be expected to affect mortality in the first year of his presidency.

The researchers acknowledged that their findings could be accidental: infant mortality and the president’s party may be completely unrelated, and instead, what seems to be a link between the president’s political party and infant mortality could actually be due to other factors that shift back and forth over four to eight years.