Little Substantive Change in Obama vs. Romney Gender Gap

There has been little substantive change in the gender gap in Americans' voting intentions -- a conclusion I reach looking at over 11,000 interviews we have conducted since April 11. Women continue to skew toward Obama and men toward Romney.

Let’s look at the numbers. We are reporting demographic breaks on a wide variety of variables each week based on three-week rolling averages of well over 8,000 interviews each. Our first report of the demographic breakouts encompassed April 11-May 6, and included more than 11,000 interviews since we began interviewing in the middle of the week (April 11). Thus, this first period of interviewing has over 5,000 interviews with men and over 5,000 interviews with women. All are registered voters.

The gender gap for the period of time April 11-May 6 thus provides an excellent baseline. Here’s what we found:

• Men were 42% for Obama; 50% for Romney
• Women were 49% for Obama; 42% for Romney.

Now, the next three-week rolling average we reported included interviewing conducted April 23-May 13. Since there is an overlap of two of the three weeks, we couldn’t expect much change, and we didn’t find it: Men were 41% Obama; 50% for Romney, and women 50% for Obama; 41% for Romney.

To hone it down a bit more, we can look at our most recent weekly average, for May 7-13. Here we find men 41% for Obama; 50% for Romney, and women 51% for Obama; 40% for Romney.

There has been a little variation. Men in the baseline interviewing (April 11-May 6) were +8 for Romney, and in the most recent week were +9 for Romney. Women in the baseline interviewing were +7 for Obama, and in the most recent week were +11 for Obama. If anything the gender gap has spread out a bit, particularly among women, but this is not what I would call a substantive change.  

Of course, this overall gender gap masks interesting patterns taking place within the data. The big interest relating to the gender gap is among non-Hispanic white voters, given that nonwhite voters will overwhelmingly support Barack Obama regardless of their gender.

For our first three-week period of tracking (actually three weeks and five days as I noted above), white women were 41% vs. 50% for Obama and Romney, respectively. For the next three-week rolling average they were 42% vs. 50%. Men were similarly stable at 32% vs. 59% and then 33% vs. 59%.

Looked at differently, for our most recent three-week rolling average, white women were +8 Romney and white men were +26 Romney.  (All of these data are available here.)

One way of looking at this is to say that the spread in the gender gap among all Americans, regardless of race or ethnicity, is 20 percentage points -- the difference between the 9-point Romney edge among men and the 11-point Obama edge among women. Among white voters, the gender gap is 18 points, the difference between the Romney edge of 26 points among white men, and 8 points among white women.

Much different overall numbers, but basically the same gender gap.

Gallup’s very large sample sizes allow us to track relatively small demographic sub-segments with stability and reliability. So when a dispute comes up surrounding reports of movement within subgroups on voting intentions, these data provide an excellent source of adjudication. And in terms of any reports that the gender gap has disappeared in recent days, we just don’t see it.

More broadly, these large sample sizes provide an excellent way to track the race in general.  So far, our conclusion is that the two major party candidates, Obama and Romney, are essentially tied at around 45% to 46% of the vote each. There are some trend patterns that show one or the other of these moving into the lead for short periods of time, but so far neither of the candidates has been able to mount a sustained lead over the other among registered voters. It remains a close race.  

The Political Impact of Obama's Same-Sex Marriage Announcement

Everyone is asking, and some are speculating, about the political impact of President Barack Obama’s decision to publicly state his personal support for legal same-sex marriage. There is no doubt that Obama’s campaign team spent long hours contemplating the political ramifications of his decision. These discussions almost certainly included an analysis of available polling data. His team's decision that Obama should go ahead with his announcement suggests that in their opinion, on a net-net basis, the potential political benefits of the announcement outweighed the potential political harm.

We know that values issues, like same-sex marriage, are low on an overall priority list for the average American. We just finished our May update wherein we ask Americans to name the most important problem facing the country. Two-thirds mentioned some aspect of the economy. Less than 1% specifically mentioned issues relating to gay rights or gay marriage. Even when we asked Americans in the same poll to tell us what worries them about the state of moral values in this country, very few mention gay issues.

From a big picture perspective, the economy rules. The direction of the economy in the next four months is going to have a much greater impact on the outcome of the election than Obama’s public announcement that he supports legalizing same-sex marriage.

It’s also fair to say that, although this issue dominates news coverage today and tomorrow, it will fade away in the weeks to come.

American presidential politics today is divided along several important fault lines. One of these is values. As my colleagues Jeff Jones and Lydia Saad and I just reviewed, Americans’ religiousness is one of the most important predictors of their partisanship and vote for president -- particularly among white voters. Highly religious and moderately religious white Americans support Romney. Nonreligious white Americans swing totally around and support Obama. Support for legal same-sex marriage follows these same lines. In other words, Americans' religiousness predicts both their politics and their position on same-sex marriage.

Thus, Obama’s public support for legal same-sex marriage will serve to reinforce this pre-existing religious/values divide in American politics. Religious Americans will most likely be reinforced in their support for the Republican presidential candidate -- Mitt Romney. Nonreligious Americans will either be unaffected, or for some, reinforced, in their support for the Democratic candidate -- Barack Obama.

Romney did not fare well with highly religious white Republican voters in the GOP primaries this year. Rick Santorum in our polls, and in exit polling, disproportionately claimed the allegiance of this voting group. These voters swung their support to Romney after he clinched the Republican nomination.  But it is not clear how passionate or enthusiastic that support has been. It is possible that the Obama announcement will serve to reinforce the motivation, passion, and enthusiasm for the Romney campaign among highly religious white voters.

Core, liberal supporters of Obama will be reinforced in their support for the president. But. There is a fascinating exception to this -- black voters, who suffer cross-pressures on moral values issues. Blacks’ support for Obama is, in general, higher than any other group we measure -- 90% say they will vote for Obama in our latest analysis. Blacks, however, do not fall in line with traditional Democrats on this issue of same-sex marriage. I aggregated our last three May Values polls (2010, 2011, and 2012) together to get a larger sample size of black respondents. Overall, 49% of Americans in these three polls said they supported legal same-sex marriage. Overall, 41% of blacks agreed. That puts blacks significantly closer to Republicans’ attitudes on this issue than to Democrats’. This reflects the fact that blacks are the most religious of any major race or ethnic group in America.

It’s highly unlikely that anything will change this vote predilection, including Obama’s announcement. It is possible that Obama’s public stance could, at least in theory, have the impact of reducing black voter motivation and turnout next November. But probably not too possible. Since it can be assumed that most blacks already knew Obama’s general position on this issue, it’s not likely that the announcement will have a material change on anything to do with black voting patterns. But we will have to wait and see.

The other possible impact of Wednesday's announcement could derive from the circumstances of the announcement per se.  Obama clearly made the announcement in response to the focus on the issue that emanated from Vice President Joe Biden's statement on Meet the Press on Sunday that he supported legal same-sex marriage. The question here is whether voters will see Obama's decision to follow suit on Wednesday as a courageous act of leadership and dedication to personal convictions, or as a cave in to political demands. Right now, Obama beats Romney on the dimension of being a strong and decisive leader.  Again, we will have to wait and see how this plays out in the minds of American voters.

Obama and Romney: Too Liberal or Too Conservative?

An interesting question in our recent USA Today/Gallup Swing States poll asked Americans to rate their perceptions of the relative ideological positioning of both Barack Obama and Mitt Romney. The question asked if Obama/Romney were more conservative “than you,” about the same, or more liberal “than you.”

Among swing-state registered voters, the percentage saying each of the two candidates are “about the same" were similar: 34% said that Obama’s ideology was about the same as theirs, while 31% said that Romney’s was about the same. (I should note that 4% didn’t answer when asked about Obama and 10% didn’t when asked about Romney.)

Now this is interesting in and of itself. It means that at the outset roughly six in 10 voters in the swing states believe that Obama and Romney’s ideology is different than theirs. This reflects the polarized nature of American politics today.

But there are very interesting differences in the way these images of the two candidates play out.

Those who believe that Obama’s ideology is different than theirs overwhelmingly say that he is more liberal than they are. In fact, 54% of all Americans say that Obama is more liberal than they are, while 9% say he is more conservative than they are. Looked at differently, if we just examine the pool of those who say that Obama’s ideology is different than theirs, 86% of this group say he is more liberal than they are. Thus, basically in American politics today (or at least in the swing states), you either think that Obama is more liberal than you are, or that his ideology is the same as yours.

Naturally, this perception of Obama as being more liberal is driven by Republicans, almost eight in 10 of whom say that Obama is more liberal than they are. Importantly, independents also tilt that way, with 59% saying that Obama is more liberal than they are.

One would imagine that Democrats would be more likely to say that Obama shares their ideology.  This is generally the case, although not overwhelmingly so. About six in 10 Democrats say that Obama is about the same as they are. But, almost a quarter say Obama is more liberal than they are (14% say that Obama is more conservative than they are).

Now, for Mitt Romney, we see a quite a different picture.

All in all, 60% of Americans say that Romney’s ideology is different than theirs. (By way of comparison, remember that 63% of Americans say that Obama’s ideology is different than theirs.) So, just to reiterate, there isn’t much difference between the two candidates on this basis.

But Romney’s image position is, by no means, the mirror image of Obama’s.  It is not the case that Americans tilt just as strongly toward the belief that Romney is more conservative than they are as they do toward the belief that Obama is more liberal than they are.

The data show that 37% of all Americans say that Romney is more conservative than they are, while 23% say he is more liberal than they are. Thus, 62% of the pool of those who say that Romney’s ideology is different than theirs say he is more conservative -- compared to the 86% for the comparable number for Obama. Thirty-eight percent say that Romney is more liberal than they are.

Fewer Democrats say that Romney is more conservative than Republicans say that Obama is more liberal than they are (64% versus 79%). Independents mirror the sample average, tilting toward saying that Romney is more conservative than they are.

Here’s the interesting finding. One-third of Republicans say that Romney is more liberal than they are, while 13% say that he is more conservative.

It is clear, in other words, that Obama has a more sharply etched ideological image than does Romney. Most of those who don’t believe that Obama’s ideology is the same as theirs say he is more liberal than they are. The majority of those who don’t believe that Romney’s ideology is the same as theirs say he is more conservative than they are. But, unlike the case for Obama, there is a sizable minority, many of them Republicans, who say that Romney is not what you would expect -- more liberal than they are.

The key to the election this year -- as is always the case -- is two-fold: motivating turnout among the candidate’s core constituency and gaining support from voters who are not firmly committed to one or the other candidate.

Those voters who are not firmly committed to one or the other candidate are generally more centrist, and not rigidly committed to one ideological side of the spectrum or the other. The fact that Romney, apparently, has a less-fixed ideological image could, in theory, put him in a better position to gain the allegiance of more centrist swing voters. At least, that is one theory. The downside for Romney, of course, is that his more centrist image may dampen enthusiasm among conservative Republicans, resulting in a lower turnout and campaign effort among this group. 

This is the tightrope that both candidates are walking this year.  We've seen this in the last several days for Obama as he and his campaign team attempt to deal with the issue of gay marriage. Going too far in the direction of endorsing legalized gay marriage runs the risk of alienating possible swing voters and also some highly religious nonwhite voters within his Democratic core.  Not going far enough runs the risk of alienating other liberal voters within his Democratic core.  Romney faces the same issues from a different perspective, but does so with a modestly less rigid ideological image than his opponent.

Obama Now Close to George W. Bush's Position in 2004

The sample size of presidents who have sought re-election in the modern era of polling is quite small -- just 10 since World War II.

But even this small handful of contests gives us some contextual clues for understanding the current positioning of the 11th president to seek re-election since World War II -- Barack Obama.

Of the 10 who have gone before him, seven were victorious: Harry Truman, Dwight Eisenhower, Lyndon Johnson, Richard Nixon, Ronald Reagan, Bill Clinton, and George W. Bush. The three who failed were George H.W. Bush, Jimmy Carter, and Gerald Ford -- although Ford was an exceptional case given that he inherited the presidency from a disgraced president (Nixon) only about two years before he had to seek re-election.

Let's look first at the two most recent campaigns in which the incumbent lost: George H.W. Bush in 1992 and Carter in 1980. Our quest is to see how the position of these two incumbents in May of their re-election year compares with where Obama is today.

George H.W. Bush was clearly in trouble in May 1992 based on his job approval rating of 40% (in a May 7-10, 1992, Gallup poll). Bush had enjoyed an 89% job approval rating just a little more than a year earlier following the successful conclusion of the Persian Gulf War. That “ratings high” was short-lived, however, and the bad economy and perceptions of how well Bush was or was not handling economic matters began to eat away at his ratings. Bush bottomed out with a job approval rating of 29% in late July/early August, and, of course, he went on to lose to Bill Clinton in November (with Ross Perot gaining 19% of the popular vote). Still, despite his ailing job approval rating, Bush was not doing as poorly in terms of the trial heat horse race numbers in May 1992: Gallup had Bush at 35%, Ross Perot at 30%, and Bill Clinton at 29%.

Jimmy Carter was also not faring well in terms of his job approval rating in the spring of 1980. His approval rating was 43% in a May 2-5 poll and fell to a nadir of 31% by June of that year. Still, as was the case for Bush senior, Carter was doing OK in terms of the trial heat ballot in May 1980. He lead Reagan by 38% to 33%, which John Anderson coming in behind at 21%.

Both of these cases suggest that the job approval rating now is more predictive than the trial heat ballot positioning. This follows the general conclusion in Robert Erickson and Christopher Wlezien’s new book The Timeline of Presidential Elections. They found that, in general, the presidential job approval rating is more important as a predictor of the election outcome at this point than the trial heat ballot -- while the ballot becomes more important as a predictor as Election Day approaches.

What about the three most recent presidents who won re-election?

George W. Bush’s position no doubt provides the most potential comfort for Obama’s campaign team. Bush’s job approval rating in a May 2-4, 2004, Gallup poll was 49%, lower than Obama’s current 51%. Bush was also essentially tied with John Kerry in a trial heat poll among registered voters: 45% for Bush and 44% for Kerry, with Ralph Nader picking up 6%. Bush ended up winning over Democrat John Kerry by only three percentage points: 51% to 48%. But he won.

Obama today is in essentially the same position as Bush was then. His job approval rating is slightly higher than Bush's May 49% reading, and although the margin varies, Obama is roughly tied with Romney in the trial heat ballot, with the most recent five-day average showing Obama slightly ahead.

Thus, Obama’s current position looks closer to Bush’s than it does for any of the other incumbent presidents we are analyzing here. As is always the case, of course, every election is different. For one thing, Bush was an incumbent Republican, while Obama is an incumbent Democrat. That means, among other things, that Bush gained a few points in the translation from registered voters to likely voters in the final analysis, which is what usually happens to Republican candidates. Obama is likely to lose some ground against Romney based on differentially higher turnout among Republicans on Election Day.

One of the major issues of 2004 was the Iraq War, with an undercurrent of homeland security and terrorism. The economy is the big issue this year. On the other hand, Bush’s 2004 opponent was a rich, white male in his 60s from Massachusetts. Obama’s 2012 opponent is a rich, white male in his 60s from Massachusetts.

But I think the bottom line is that Obama is in reasonably close enough proximity to Bush’s positioning in 2004 -- based on the job approval and the trial heat ballot indicators -- that he could in theory duplicate Bush’s win.

Bill Clinton was in a better position when he sought re-election in 1996 than Obama is now. Clinton had a job approval rating of 55% in a May 9-12, 1996, poll, and he was significantly ahead of Bob Dole, 47% to 32%, in the trial heat ballot, with Ross Perot getting 19% of the vote. Clinton ended up winning over Dole by a margin of eight points, with Perot picking up 9% of the vote.

Ronald Reagan had a job approval rating of 52% in a May 3-5, 1984, poll. He was only ahead of Walter Mondale by four points, 50% to 46%, among registered voters in a May Gallup poll. Reagan picked up steam as the year progressed, bolstered by an effective campaign strategy, and eventually went on to win over Mondale by 18 points: 59% to 41%.

Overall, we see that Obama is right in the middle of the contextual landscape provided by the last five U.S. presidents who sought re-election. He is not in as bad a position as were Carter and George H.W. Bush in May of their re-election years. He is not in as good a position now as were Reagan and Clinton in May of their re-election years. Obama is just about exactly where George W. Bush was at this point in his re-election year -- and of the five presidents, Bush’s margin over his opponent in the popular vote ended up being the closest. This suggests that a prediction of a close race next November would not be unreasonable.

The big question: What will the trajectory of Obama’s approval and ballot positioning be between now and November? That will in part depend on the trajectory of the economy. Stay tuned.

Tracking the Presidential Race

Check out our recent Election Matters program in which I talked with Temple University Professor Chris Wlezien (co-author with Robert Erikson of Columbia University of the forthcoming book The Timeline of Presidential Campaigns) about forecasting the presidential election. Their research shows that the ballot position of the candidates at this point is not highly predictive of the outcome of the election. Presidential job approval is somewhat more predictive at this point. The predictive ability of the trial heat ballot increases as the election gets closer.

Of course, the reason to measure the trial heat right now is not just to predict the outcome next Nov. 6. We believe that tracking voter sentiments (i.e., voters' responses when asked for whom they would vote if  (“...the presidential election were held today”) helps us understand where the two campaigns are at this point, and what’s happening as they progress. Both Obama and Romney are actively out there attempting to establish positioning, to criticize their opponents, and to gain early momentum. Some observers argue that fixing the images of the candidates in the minds of voters at this point is, in fact, highly important because these initial impressions can take hold and stay in place through next fall. Hence, it’s useful and interesting -- in our opinion here at Gallup -- to monitor how all of that is going.

At this point I would say our tracking, initiated on April 11, shows a quite close race. In fact, if we put together all of the 8,059 interviews we conducted April 11-29, we find that Obama has 46% of the vote of registered voters, and Romney has 46%. In other words, a close race. But there have been shifts over that time period. Obama was ahead in the first several nights of interviewing after April 11, but then Romney moved ahead. Obama regained the lead in our Gallup Daily tracking through last week, but now Romney is moving back out ahead again.

The nightly numbers that form the basis for the tracking averages we report are fairly consistent. Each night’s tracking is a self-contained sample (although the numbers called can be ones left over from previous day's interviewing attempts). Romney was ahead for five out of six of our independent nightly samples from April 14 though April 18, Obama was ahead for seven straight independent nights of interviewing from April 19 through April 26, and now Romney has been ahead for three straight nights from April 27 through April 29. Keep in mind that these are all separate samples. So the conclusion is that the population of registered voters out there across the country has shifted at least marginally several times over the course of the 18 nights of interviewing we have conducted so far.

In terms of our rolling average, one way to look at it is to say that Obama’s share of the ballot has ranged between 43% and 50% over the last 18 days, while Romney’s has ranged between 42% and 48%. That’s a six- and a seven-percentage-point range. Obama’s approval in that rolling average has ranged between 45% and 50% over that time period. Notice that Obama job approval has not varied as much, operating since April 11 in a tighter range, and that is based on a three-day aggregate with a lower sample size. At least at this point, Obama's job approval rating appears to be slightly more stable than his share of the ballot among registered voters.

The conclusion is that the presidential race appears tight, but with some fluidity. Everything we have learned in tracking suggests that this fluidity is fairly normal. Although many voters at the ends of the political spectrum are rigidly fixed in their vote choice, there is always movement in the middle. Not a lot of movement, of course, from a big picture perspective. But there is movement. The reason we track, of course, is to monitor that movement. 

There are differing views on the value of tracking an election race (or tracking the Dow Jones or any other commodity) on a daily basis. The wider the time span over which you track something, the less the volatility and change. So one's decision on tracking depends on one’s perspective on the value of information. Reasonable people disagree on this. Some people think that all that matters is the longer-term, slow change in things like presidential job approval and an election ballot. They would be content to have thirty-day rolling averages that show very slow change. Others (including us here at Gallup) think that it's also valuable to track shorter-term movements in these indexes for the reasons I have given above. (Of course, any consumer of our data can average the reported results into as long a time period as they want to).

Our objective is to find the sweet spot between picking up too much variation and not picking up enough variation. Right now we are using a five-day rolling average, based on around 2,200 interviews with registered voters for each five day period. We have used a three-day rolling average in election tracking in the past and currently still use a three-day rolling average for presidential job approval. We will soon begin reporting candidate choice within detailed voter segments, most likely using longer-term three-week rolling averages, and we may extend our daily election tracking average to a weekly time period at that point.

We are entering a period of the election now when there are fewer specific milestones. There is Romney’s pending vice presidential decision and then the conventions in late August and early September. But with the nominees now fixed -- unless there is a major third party entrant -- some of the excitement and attention paid to the race may drop. Still, all of our experience going back to tracking the 1992 presidential race, suggests there will be change in ballot positioning of the candidates as we go.