The poll includes a drop in Mr. Obama’s favorability ratings, with 36 percent saying they viewed him favorably and 48 percent saying they did not. In April, 42 percent expressed a favorable opinion of him and 45 percent an unfavorable one.
Pretty bad news for the president, you’d think. But wait, don’t worry, there’s an obvious explanation:
That change may have been affected by a reordering of this particular set of questions, which at this point in general election cycles are typically placed near the top of the survey. During the primary season, the questions about favorability were placed lower on the survey after queries about presidential job approval and other topics.
A similar explanation is offered in the poll-results document itself: “This question was farther down in the body of the questionnaire in earlier surveys this year. Favorability questions are typically moved higher at some point during the general election season. Variation in the order of questions can affect the results.”
I’m not a pollster, but there seems no reason to believe that people’s answers to how favorably they view the president (hardly a throwaway question) would be significantly affected by when in a survey they’re asked about it. Obviously responses to poll questions can be somewhat affected by their placement, but there’s no reason to assume or infer that it has a noticeable effect here, or a negative one. In fact, Romney’s favorability ticked up in the same survey, 29 percent to 32 percent — can we assume the former governor actually has experienced an even bigger improvement?
Further, if one examines the April and June surveys, if there really are effects from question ordering, it seems like the change should have helped Obama: The favorability question is now asked as the second or third question, whereas it used to be asked after a litany of questions about, e.g., Obama’s handling of Afghanistan, the state of the economy, and whether the respondent’s financial situation has improved or deteriorated . Surely Obama should benefit from people being asked about their feelings on him before being reminded about all those issues.
(Though credit is due, I guess: Unlike May’s unfavorable poll for the president, this actually did get the prominent billing the Times gives to its monthly polls.)