Media outlets have repeatedly cited a September 2 Washington Post/ABC News poll to claim that public opinion is split on President Bush's handling of the Hurricane Katrina disaster. But The New York Times, The Washington Post, National Public Radio, and ABC News did not inform their audiences of factors that undermine the poll's reliability and usefulness: The sample was small, and it was taken on the Friday evening of a national holiday weekend.
The Washington Post/ABC News poll found that 46 percent of respondents approve and 47 percent disapprove of Bush's handling of the situation caused by Hurricane Katrina. Yet despite the poll report's own acknowledgment that "[t]he practical difficulties of doing a survey in a single day represent other potential sources of error in this poll," neither The Washington Post nor ABC News noted these problems in their reports on the survey.
Polls with small sample sizes -- in this case, the poll surveyed 501 respondents in one evening -- yield an increased margin of error. Polling organizations typically consider 1,000 to 1,200 respondents the minimum number necessary for a reliable national poll. But a further complication -- the fact that the poll was conducted only on Friday night -- was mentioned in passing in news reports, with no explanation of why this might be problematic. Most pollsters do not conduct surveys on Fridays and Saturdays for the simple reason that many people are not home on weekend evenings to take calls. Those who are not at home have no chance of being included, potentially skewing the sample. For instance, younger people are more likely to go out on the weekends and are thus less likely to be included in a weekend poll. Both Republican pollster Bill McInturff and Democratic pollster Celinda Lake have cited these problems in explaining why they do not poll on Friday and Saturday evenings. Furthermore, as Mark Blumenthal, who operates the weblog Mystery Pollster, has noted, the Post/ABC poll was conducted on the first evening of a holiday weekend, meaning even more people were unlikely to be home.
Nevertheless, the Post wrote that the results of its poll “might offer some cheer to beleaguered White House staffers who feared a stronger negative reaction.” Similarly, the ABC News website titled its article "Poll: Bush Not Taking Brunt of Katrina Criticism." The New York Times reported that the poll revealed “ambivalent feelings” about the president's handling of the crisis, and National Public Radio White House correspondent Don Gonyea remarked that the survey showed “an even split.”