Obama: The Worst President Since WWII?

Source: Mandel Ngan/AFP/Getty Images

Source: Mandel Ngan/AFP/Getty Images

Americans aren’t feeling the presidential love these days. President Barack Obama has been down in approval polls for quite some time now, hovering in the low 40 percentiles for much of 2014, and hitting his all time low of 38 percent job approval in 2011 most recently — according to Gallup. On average, compared with other presidents during this time in their presidency, he isn’t doing that far below the average. However, a new poll from Quinnipiac University listed President Obama as the worst U.S. president since World War II, worse even then the runner up: George W. Bush. Richard Nixon had 13 percent of respondents calling him the worst president, with Jimmy Carter taking 8 percent and George W. Bush with 28 percent to Barack Obama’s 33 percent. Only 8 percent called Obama the best president, while Harry Truman and John Kennedy both received no votes putting them as the worst presidents, and Ronald Reagan received 35 percent of votes for being the best president.

The numbers looked at a pool of 1,446 registered voters, 26 percent of which were Republican, 31 percent Democrats, 35 percent Independents, and 7 percent listed as other. It also showed 45 percent of respondents saying that Mitt Romney would have made a better president than Obama has, with 38 percent saying the nation would have done worse under Romney. “Over the span of 69 years of American history and 12 presidencies, President Barack Obama finds himself with President George W. Bush at the bottom of the popularity barrel,” said Tim Malloy, the assistant director of the Quinnipiac University poll. But is that really what people think?

With Quinnipiac’s data causing such a stir, it’s important to take into account a few things about the available data and about how people respond to polls when viewing some of the more sensational headlines. Slate points out that the poll is largely influenced by how — or more importantly, if — Democrats and Republicans divide their votes. For example, if you view the poll as a partisan split, with Democrats likely to vote positively for Obama, and Republicans to vote negatively, the disadvantages are glossed over. If you look at the questions as a depiction of how divergent preferences are within a party, you have a more accurate picture of why the ultimate splits are somewhat skewed.

Republicans voted highly positively for Reagan across the board, with 66 percent voting for him as the best president. Compare this to Democrats, who have a tendency to live multiple presidents and to have differing preferences, and much like in a congressional election where two strong left or right candidates split the vote, the results are thinned as they are spread out.

The same can be said for negative votes in the reverse, where Republicans are unanimous in who they vote as the worst while others may not be. Loss of Independents are a genuinely negative sign for the president, as is the high number of votes in favor of Romney, but the vote for best and wort president is somewhat skewed. It’s for this reason that state election polls often ask respondents questions on their preference between all candidates, and then comparing one candidate head to head with others individually. Indeed, Quinnipiac did just this with former President Bush and President Obama, finding that 39 percent said Obama is a better president, 40 percent saying he is worse than Bush, and 20 percent saying the two are about the same.

The historical trend is similar with this question. President Obama’s percentage has dropped since 2010 when it was 43 percent saying he was better to 30 percent saying Bush had been worse. In 2011 he was at 46 percent to Bush’s 30 percent. Science of Us discussed the availability heuristic, which sounds complex but is simply referring to how easily a name comes to mind when answering questions. When being questioned by a polling person, the name that most easily comes to mind — even if its not the name you’d chose with time and thought or a written poll — will often bias the answer given.

The actual question asked by those conducting the poll was: “Which of these twelve presidents we have had since World War II would you consider the worst president: Harry Truman, Dwight Eisenhower, John Kennedy, Lyndon Johnson, Richard Nixon, Gerald Ford, Jimmy Carter, Ronald Reagan, George Bush Senior, Bill Clinton, George W. Bush, or Barack Obama?”

Polls are useful and fascinating data sources, but they are often deeply flawed or inherently inaccurate when phrased poorly, often lack objectivity — depending on who conducts the polls — and ultimately people often fail to remember that what they show is opinion, not reality, which isn’t to say that the two don’t sometimes coincide. Anyone who can access Google knows how old Hillary Clinton is. But a poll on how old people think she is can be a useful set of data to have, as we see in the Pew Research table below, and the question is simple and numeric enough to be easier to trust the result of. Asking whether or not that age matters, or is too old, or what degree of disadvantage, becomes a much more complex matter.

Screen Shot 2014-07-02 at 4.29.53 PM

There are all sorts of biases polling groups are aware of when they attempt to design a study. There’s the basic problem of the observer bias; people change their behavior when they’re being observed, which goes hand in hand often with the social desirability and self-serving biases wherein respondents taking part in polls often try to give the answer that makes them appear best or more socially acceptable even if it’s not accurate. The order that names or terms are listed in can have an effect on what respondents choose.

So, while listing the presidents in the order that they have historically been in office may be an objective and neutral way to present them — that’s the order they came in after all — it may still have influence on how people hear the options. The serial position effect is a term that shows how people have a tendency to remember the first and last items on a list, with the middle items fading more quickly from memory. People respond differently to polls taken on paper, in person, or over the phone — as Quinnipiac’s poll was. Question phrasing and leading questions are big concerns in some polls, asking if someone dislikes an idea or person is very different from asking someone’s view on three completely unrelated topics in order to see their opinion of one of those topics more accurately.

All of these things may or may not have effected the Quinnipiac poll specifically, but it’s generally valuable to understand how easily manipulated, or accidentally misleading polls can be. Basically, a completely bias-free poll is nearly impossible, so the best we can hope for, since in the end quantifiable stats still have their uses, is to read what a poll is really saying.

More From Wall St. Cheat Sheet:

Follow Anthea Mitchell on Twitter @AntheaWSCS