In these deeply troubling economic times, can religion serve as a bulwark against financial and economic angst, worry and negativity? 

A look at some recent Gallup Poll data suggests that the answer may be yes.  Americans who are the most religious are the least likely to rate the current economy negatively or to say they personally worry about money. 

As usual, I’m going to use church attendance as a summary measure of religiosity.   And I’m using Gallup Poll tracking data for Sept 1-28, involving over 14,000 interviews.

Overall, there is a very statistically significant relationship between church attendance and ratings of the economy.  Those who attend church weekly are 13 points less likely to rate the economy as negative than those who seldom or never attend church.  To be specific, 68% of weekly church attenders can be classified as negative about the U.S. economy, compared to 81% of those who seldom or never attend church.

Now to be sure, some of this is caused by the intervening variable of party identification.  Republicans are both more religious and more positive about the economy than are others.  (Why Republicans are more religious is a fascinating topic for another time; why Republicans are more positive about the economy is most likely due to a loyalty to their Republican president).

But lo and behold, the relationship still holds even when we control for party identification.  Those Republicans who are the most religious are 11 points less likely to worry about the economy than those who seldom or never attend.  The same trend exists with groups of both independents and Democrats.

Religion also seems to protect one from worrying about one's own money.  Thirty-one percent of Americans who attend church weekly said that they worried about money “yesterday”.  That increases to 42% who worried  about money among those who seldom or never attend church.

Religion seems to be a particular antidote against money worry among those with the lowest incomes.  I’m looking here at those who report incomes of only about $12,000 a year or less, obviously at the very bottom end of the income scale.  Only 41% of those in this group who attend church weekly worried about money yesterday.  That jumps to 56% of those who seldom or never attend church.

This presumably comes as no surprise to many highly religious people.  Many religions explicitly embody as part of their theology and practice a focus on providing the faithful with a a surcease from earthly sorrow  (A passage in the Christian New Testament says, for example:  “Let not your heart be troubled: ye believe in God, believe also in me.”).

Of course, some would argue that this isn't necessarily a totally good thing.  Worrying about the economy or one’s financial situation may not be all bad. We humans were most probably designed through evolution to have some anxiety and concerns because such emotions provide the impetus for us to strive to improve and move forward.  Which in turn in theory could improve society.

But regardless of the implications for the long term survival of our species and the planet, the data seem to suggest -- at the surface level at any rate – support for a connection between religiosity, as measured by church attendance, and less worry about earthly economic matters.   

Here’s an update on religion and the vote, based on 13,000 Gallup Poll tracking interviews conducted in the first half of September.

The fundamental divides in American voting behavior based on religious identity appear to remain as strong as ever:

1. Protestants and other non-Catholic Christians remain tilted towards John McCain, by about a 12-point margin.

2. Take blacks out of the picture and focus just on white Protestants/non-Catholic Christians, and we end up with a much higher, 31 point lead for McCain over Obama.  In short, white non-Catholic Christians continue to be one of McCain’s strongest voting blocs.

3. Taken as a whole, Catholics are split down the middle between Obama and McCain.  But that’s deceiving.  Hispanic Catholics skew to Obama by a 64% to 28% margin.  That means, if you do the math, that non-Hispanic Catholics (mostly whites) tilt the other way, supporting McCain over Obama by an 8 point margin.

4. Jews remain strongly in Obama’s camp, supporting him by about a 37 point margin.

5. Mormons, about 2% of the population, are basically McCain’s strongest religious group of all, supporting him over Obama by a 55 point margin.

6. Non Christian, non-Jews – which would include Hindus, Buddhists and the like -- are strongly supportive of Obama. (We don’t have enough Muslims in the September sample to break out separately).

7.  As I’ve pointed out repeatedly, those Americans who have no religious identity and/or who are explicitly atheists or agnostics skew strongly towards Obama by a 41-point margin.

The bottom line here:  Religion continues to be a useful and important predictor of the vote.  Americans’ religious identity is of course intermingled with many other variables, including geography and socioeconomic status, but regardless of the underlying reasons, if we know a person’s religious identity, and in particular if we can match that with his or her ethnicity/race, we can do a pretty good job of predicting how they are going to vote.

I’m not yet seeing any evidence that Sarah Palin had a disproportionate effect on highly religious white voters.   

That’s of interest because many observers had hypothesized that McCain’s selection of Palin would create gains among religious whites who resonate with Palin’s well-publicized conservative positions on value issues like abortion and gay marriage. 

It is true that overall the McCain ticket gained modestly as a result of the convention among white voters. The two candidates started exactly even.  Obama boomed up briefly after his convention, only to lose all of his gains and more by the time the GOP convention was over.  The net effect of the conventions, then, among white voters was a modest gain for McCain.  In comparing polling from the week before the two conventions (Aug 18-24) to the four days immediately after the final convention was over (Sept 5-8), we find that among white voters, McCain picked up 3 points and Obama stayed even.  (McCain gained a little more when non-white voters are taken into account).

Now here are the data by religiosity.

McCain picked up 3 points among white voters who are weekly church attenders (our definition of highly religious whites).  Obama didn’t pick up any points among this group, but didn’t lose any either).  His support levels were unchanged.

Among whites who are mid-range in their religiosity – attending church almost every week or monthly – McCain also gained, 4 points as it turned out.  Very little change in support for Obama; he lost a point.

Finally, among whites who are not religious – those who seldom or never attend church – guess what?  McCain gained 2 points in this group.  Obama stayed even.

Let me summarize:

1. Highly religious whites – McCain gains.   No change in support for Obama.
2. Moderately religious whites – McCain gains.   Little change in support for Obama, who loses a point.
3. Not religious whites – McCain gains modestly.  No change in support for Obama.

This is not the picture we might have expected.  Instead of a disproportionate increase in support among highly religious whites, we find that the gains for McCain appear to be relatively unrelated to religiosity.  McCain gained across the board.  (And, to reiterate, those gains came from undecideds, not conversions from former Obama supporters.)

These data do not allow us to support a hypothesis that McCain’s selection of Sarah Palin as his vice presidential running mate caused a significant shift of highly religious white voters to his support.  We just don’t see it in the data.  It looks like McCain managed to pick up support by drawing in voters from across the religious spectrum.  "Secular gains" we might call them.

Now, that’s not to say that Palin may not have increased energy, commitment and motivation among highly religious whites who were already supporting McCain.  That’s possible.  We did show that enthusiasm moved up among Republicans in general as a result of the convention. And of course, increasing turnout among the faithful can be just as effective a way to win votes as to change people’s minds.

But Palin apparently did nothing special in terms of shifting highly religious whites out of the undecided category to her ticket’s cause.

The Obama campaign team is apparently very conscious of the need to increase market share among religious white voters, and based on all available research, I don’t blame them. 

I reviewed the basic lay of the land here.  The facts are straightforward.  Democrats and Democratic presidential candidate Barack Obama simply don’t do well at all among whites who are highly religious.

News has come forth from the Denver convention that the D’s held an interfaith worship service at the Colorado Convention Center Sunday, part of an apparently major effort to reach out more to religious  voters.

There were a wide variety of faiths represented at the worship service.  But as AP writer Eric Gorski put it:  “Behind the scenes, efforts to attract the religious vote will concentrate largely on Christian ‘values voters.’"

There’s a lot of work to be done here.  If anything, my most recent analysis shows there’s been a slight movement towards McCain among white religious voters in the last week. 

For the week of August 18-24, our Gallup Poll Daily tracking aggregate found 66% of whites who attend church weekly supporting John McCain, with only 24% supporting Barack Obama. That in and of itself gives you a sense for the magnitude of the challenge facing Obama in his effort to reach religious voters.

But when we look back, as noted, we find a slight shift away from Obama. For the 8 weeks of the summer from early June through August 3, 63% of white highly religious voters supported McCain while 26% supported Obama.  That shifted a little in the first two weeks of August to 25% for Obama, 64% for McCain.  Now we have the further shift to the 66% to 24% ratio mentioned above.

These are not huge shifts, to be sure.  But with the large sample sizes involved in our weekly tracking, they represent real trends.

More generally, even if we forget the recent worsening of Obama’s position, the bottom line is that McCain is beating Obama among highly religious white voters by well more than a two to one margin.  This is not new.  It’s a challenge faced by all recent Democratic candidates. It is no surprise that Obama is trying to do something about it.

Of course, we’ll see after the dust settles whether or not the Democratic convention has shifted things a little more back in his favor.  As of Monday night, it looked like Americans were actually moving more towards McCain, not what history predicts should be happening after Obama's VP announcement and the heavy duty focus on the Democratic convention.  There is a significant probability, however, that  by the end of the week there will be at least some shift towards Obama overall.  If so, we’ll investigate and see if some of those gains came among white religious voters. 

The back-to-back appearances of Barack Obama and John McCain at Rick Warren’s Saddleback Church in California over the weekend has put a renewed focus on the key voting bloc of highly religious white Americans. 

Notice that I said “highly religious white Americans” here -- while many others have been using the term “evangelicals”.  That’s because I think the term “evangelicals” is very tricky to attempt to define (or as social scientists say, to operationalize).   

I’ve studied this a lot over the years, and even co-authored a book (The Evangelical Voter, written with Stuart Rothenberg) on the subject a while back.  There simply is no agreed-upon definition or way of defining exactly who is and who is not an evangelical.  Everybody who tries does it somewhat differently.

The bottom line, in my experience, is that one can establish a spectrum of definitions of "evangelicals" that can range from the very exclusive at one end to the very loose at the other.

Evangelicals can be defined in terms of their church membership (anyone who identifies with certain denominations can be classified as evangelical), their beliefs (including such things as belief in a literal Bible, beliefs about Jesus), behaviors (church attendance, active proselytizing, importance of religion), experiences (having had a personal “born again” experience, having “accepted Jesus Christ as personal savior”) and self-definitions (do you consider yourself to be an evangelical?).

A researcher so inclined can establish very rigorous criteria, and tighten the definitional noose so much that only a few saintly humans will qualify as truly evangelical.  One research group takes this so far, and imposes so many requirements or hoops that respondents have to jump through to be classified as evangelical, that only about 8% of the population qualifies.  That’s fine if that’s what one wants to do, but that small a group has less interest from a political perspective.

On the other hand, by some definitions almost any Christian is “evangelical”, in the sense that a Christian by definition should subscribe to the basic Christian doctrine that adherents should evangelize, and that adhererents should spread the word of the Gospel.  A researcher could end up saying, in other words, that well over half of Americans are evangelicals, since they adhere to a Christian faith.

Debate over who is or is not an evangelical, and/or analysis of the interrelationships among all of the kind of variables I’ve discussed above is interesting and can be important. 

But in an election season, what I think we’re really interested in is a definition that breaks out a key subgroup of religious American voters who have enough communality that they behave in particular ways relating to their vote.  And for that purpose, one simple definition that works pretty well is to break out non-Catholic white Christians, and to impose on that group some measure of religiosity – most often church attendance or self-described importance of religion.

Note that this removes Catholics from the equation. There’s no right or wrong on this, but I would side with researchers who argue that Catholics share such a different religious tradition and bureaucratic framework from non-Catholic Christians that it makes sense to deal with them as a separate group.

And the race issue. Blacks in the U.S. are largely non-Catholic Christian. They are on average highly religious. Blacks would in general meet many of the evangelical criteria I’ve discussed above.   But politically speaking, it confuses things to put blacks into the equation.  Blacks are one of the most Democratic groups in American politics, regardless of their religious beliefs.  Their largely Democratic vote does not vary much from election to election, and the impact of presidential campaigns is largely confined to affecting black turnout.  So it make sense to treat blacks as a separate category as well.

So we are left with a group I think is quite useful in our pre-election polling:  white, non-Catholic Christians who attend church on a weekly basis, or white, non-Catholic Christians who say religion is important in their daily lives.  The former represents about 15% of the total population at this point (and a higher percent of registered voters), and the latter about 30% of the total population.

Time for an update.  I’ve been analyzing over 22,000 interviews conducted as part of our Gallup Poll Daily Tracking with non-Hispanic whites in the month of July.  I’m looking at the presidential ballot.  Overall, McCain beats Obama among whites by a 50% to 39% margin.  (Obama wins among all registered voters by 3 points because of his strong offsetting performance among everyone else, including in particular blacks and Hispanics).

And, as we’ve been seeing all year, one of the most important variables which helps us understand voting patterns among non-Hispanic whites (about three quarters of the registered voter population), is religion.

Best example (again, from our July Gallup aggregate).  "Is religion important in your daily life?"  Simple question with important consequences. Among those who say “yes”, it’s McCain by 37 points.  Among those who say “no”, it’s Obama by 16 points.

In other words, other than cheating and asking a white registered voter with which party he or she identifies, this is one of the most valuable single questions we can ask to get a handle on presidential voting intentions.

Other questions about religion are almost as valuable. Whites who attend church weekly go to McCain by 36 points. Those who attend nearly every week or monthly tilt towards McCain by 18 points. Those who seldom or never attend church support Obama by 9 points.

And, one’s religious affiliation in and of itself remains an important predictor of vote choice. Among non-Hispanic whites, anyone who says they are a Protestant or identifies with a non-Catholic Christian faith is strongly likely to be a McCain supporter.  In fact, McCain beats Obama among this group by a 57% to 31% margin.  Catholics (remember, these are non-Hispanic Catholics) on the other hand are a little less likely than whites in general to support McCain, giving him only a 6-point margin.  Mormons are overwhelmingly supportive of McCain, who beats Obama by a 74% to 18% margin among LDS members.

Everyone else tilts towards Obama.  Jews, by a 26 point margin, “other religions” by a 23 point margin.  And the roughly 10 percent of white registered voters who have no religious identity tilt towards Obama by 63% to 27%.

I think it’s interesting that it’s Obama’s religion which has received the most news coverage this year.   He's been very heavily identified with the Trinity United Church of Christ.  Some of this identification, of course, has not been positive.  But nevertheless the point has been driven home that Obama is (or has been) an active member of a Protestant, Christian denomination. 

There's been less attention paid to McCain's religion.  He is usually listed as an Episcopalian, although some critics point out that he claims to be Baptist when convenient since he has attended a Baptist church in north Phoenix.

But no matter. Whatever publicity has accrued to Obama's religion has apparently made little difference. This religious divide existed long before these two particular candidates came along -- most probably a reflection of the strong moral and values concerns of the segments of the white electorate who are highly religious, particularly in Christian faiths.  High religiosity usually correlates with stronger opposition to traditionally liberal positions on social/moral issues.  In a nutshell, the Democratic Party has been more open to acceptance of behavioral patterns which deviate from traditional normative patterns – particularly relating to family structures.  The Republican Party has been more protective of traditional structures and behavioral norms, aligning them with the viewpoints of those who are highly religious. And it doesn't look like this is going to change this year.

Billionaires Bill Gates and Michael Bloomberg last week announced a major financial commitment to lowering the rate of smoking around the world.  No doubt the money will go to the traditional mechanisms for reducing smoking – anti-smoking media campaigns, education, and distribution of non-smoking initiatives.

I’ve got a better solution.  Just increase the rate at which people attend church or other worship services. 

A fascinating analysis of over 100,000 interviews Gallup has conducted since January shows that within this country, there is a strong and  direct correlation between smoking and church attendance.  The more often Americans attend church, the less likely they are to smoke.  This is a robust relationship that holds up even when things like education, age, gender and religious type are controlled for.  It just appears that something about being a frequent church attender significantly lowers the probability that one smokes.

This is in the U.S., to be sure.  Gates and Bloomberg are trying to stop smoking in developing countries.  But the principle may well be transferable.

Let's look at some numbers. Overall, 21% of Americans interviewed in our Gallup Daily tracking program this year say that they smoke.  (By the way, that’s down from an all-time high of 45% back in 1954).

But the percentage of smokers is only 12% among those who attend church once a week.  Smoking rises to 15% among those who attend almost every week.  Then 22% for those who attend once a month, 26% for those who seldom attend church, and finally 31% among those who never attend church.

So knowledge of an Americans’ church attendance alone predicts smoking pretty well.  This is a straight linear relationship the likes of which we don’t see all that often in survey data.

The relationship between smoking and church attendance seems to hold up within almost every subgroup of the population I can come up with. For example, there's education.   Generally speaking the smoking rate goes down as education goes up.  But there is a relationship between church attendance and smoking within every sample of Americans created by education.

Smoking overall is very low among Americans who have post graduate education.  But within this highly educated group, 4% of those who attend church weekly smoke, compared to 10% of those who never attend church.

The relationship is stronger among those with high school educations or less.  In this group, smoking goes from 17% for those who attend church weekly to an amazing 48% among those who never attend church.   Yes, that’s right. No typo.  Almost half of Americans who have high school educations or less and never attend church, are smokers.

In terms of gender, smoking goes from 10% among women who attend church weekly to 29% among women who never attend church, and from 14% among men who attend weekly to 33% among men who never attend.

Finally, age.  This is very interesting. It has been well-established that the older one is, the less likely one is to smoke.  This is attributable either to wisdom (the older one gets, the more one has learned about the negative effects of smoking) or to nature’s taking its grim toll among smokers as they get older, ruthlessly reducing the number of surviving smokers.

Smoking is, as an example, much lower among Americans who have survived to be 70 years of age and older.  But even here church attendance matters. Only 5% of those 70 and up who attend church each week smoke.  But triple that percentage, 15% of those who are 70 and older and who never attend church smoke.

Among the youngest group of Americans, 18 to 29, the smoking rate is 14% among those who attend church weekly but rises to a whopping 37% among those who never attend.  This seems to be a critical finding.  If we can get our young people into church, it seems, we may be able to prevent them taking up the nicotine habit.

The effect occurs within every major religious grouping. Protestant, Catholic, Mormon, other Christian religion, non-Christian religion, those with no religion.  In every instance, those who attend church the most are the least likely to smoke. 

(There aren’t a lot of Mormons who never attend church, but among that group, 34% smoke. Among the much larger group of Mormons who dutifully attend every week, the smoking rate is at a miniscule 2%.)

The exact mechanisms behind this relationship is unclear.  To my knowledge there’s nothing specific in the Bible about smoking.  However, many evangelical Protestant faiths have had as part of their normative structure a prohibition (or at least frowning upon) social sins such as smoking, drinking, gambling and dancing.  This is obviously true of Mormons as well.  So one plausible explanation is that highly religious Americans are more subject to religious training/dogma that discourages indulgences such as smoking.

But we also know that church attendance is correlated with higher self-reports of happiness, lower stress, lower anger, and lower sadness.  And at the same time smokers are less happy, and more likely to report being angry, stressed and sad.  So it’s possible that smokers engage in their habit as a surcease from their negative mental state and/or negative emotions. Church goers find their surcease from these negative emotions in their religion, and therefore may not need the palliative impact of nicotine as much.

Whatever the reasons, Bloomberg and Gates might well sit up and pay attention.  One key to reducing the rate of smoking – at least using the American situation as the example -- may well be to get people more into religion. 

This may come as no shock to those who look around them at worship services they might attend,  but there is a strong correlation between age and religion. Older Americans are decidedly more religious than younger Americans. Some new Gallup analysis sheds some very specific light on the pattern.

I’ve just finished analyzing over 145,000 Gallup Poll Daily Tracking interviews conducted from January to mid-July this year.  We have several measures of religiosity in the on-going interviews.  One of the more straightforward is the basic question “What is your religious preference?”.  We can look specifically at the percent of respondents who said that they had no religious identity and/or were atheists or agnostics. 

Exactly 11.9% of all Americans, regardless of age, tell an interviewer in response to the basic question that they have no specific religion with which they identify, or that they are an atheist or agnostic.

Since we have so much data, I was able to examine these responses across 72 different subgroups of the population, each created by current age at the time of the interview.  In other words, the percent who have no religious identity within each age group from those who were 18 years of age to those who were 89 years of age. Year by year.

The result is a negative linear relationship, with the percent having “no religion” dropping steadily as age gets higher. 

It’s a little messier in the youngest age range.   Twenty percent or more of the group of Americans under 30 have “no religion”.  But the exact percent jumps around a bit in this younger group. In fact, the single year of age in which the percent “no religion” is highest is not 18, the youngest age included in our polling sample, but 23.  In other words, if you want to pick one year in which lack of religion peaks in America today, it’s age 23.  Ironically enough – 23% of those 23 years of age say that they have no specific religious affiliation or identity, or that they are atheists or agnostics.

Beginning at age 30 this “no religion” percent begins to decline.  By age 40 it’s down in the lower teens, and by age 62 it’s in the single digits. It keeps getting lower. Of all of those we interviewed over the past six months who were 89 years of age (over 150 of them),  only 2% said that they had no religious identity whatsoever. 

So, to modify the old saying a bit, while there may be no atheists in foxholes, it appears that there are no atheists in retirement homes either.

The question here is whether or not this major age difference is a factor that is always with us, or one that is peculiar to our time in history.

As sociologists tell us, the older age group in a cross-sectional slice of a population reflects two factors.  First, that old age segment was born and raised in a previous era, and thus subject to different cultural and environmental influences.  Those now in their 80s in the U.S. were all born roughly between 1918 and 1928.  This group, quite obviously, for the most part experienced some wildly different things than were experienced by people in today’s younger cohorts.  Eighty year olds lived through the depression as young men and women, went through World War II (many having served in the military in that war or in Korea), came of age and married in the 1940s and 1950s and were in fact the parents of the now-fabled baby boom. 

At the same time, these people are now old, and are experiencing the same issues and factors that affect all old people, whenever they were raised.  They have health issues, face death, are retired, have grandchildren, and face all the other things that go along with being old (good and bad). 

So older Americans can be different from younger Americans either because they were raised in a different era, or because they are simply at a different phase of life.   Older Americans, in other words, are more religious than younger Americans today because: a. They were raised in a more religious time and their early socialization and life experiences led them to be religious, b. They are older and facing death and that fact (along with others) encourages one to be religious, or c. All of the above.

An analysis by National Opinion Research Corporation/University of Chicago scholar Tom W. Smith presented at the annual meetings of the American Association of Public Opinion Research this May suggests that c. may be the correct answer.  Younger Americans have always (well, at least back through the 1970s at any rate) been less likely to identify with a religion than those who are older.  But it appears that the percent who say "no religion" has been significantly increasing among the youngest Americans today compared to its increase among older Americans. 

So older Americans are relatively more religious than younger Americans because they are old -- a standard pattern, but also it appears because the young people of today are particularly less likely to be religious than was the case for the young people of yesterday.  (This suggests, of course, that as the young people of today age, they will become more religious, but most likely will never be as religious as their parents and grandparents).

By the way, as is well known, young people of today are much more likely than those older to support Obama for president. In fact, only 39% of those 65+ in our latest Gallup sample support Barack Obama for president compared to 65% of those 18-29. So older Americans are both more religious and less likey to support Obama. Exactly what's behind this political relationship is the subject for a future discussion.

The impact of religion on the presidential race this year is pretty well established, at least to this point in the contest. As my colleague Jeff Jones outlined here, the largest religious group in America – non-Catholic Protestants and other non-Catholic Christians -- favor McCain over Obama.  American Catholics skew slightly towards Obama over McCain.  Jews and those with no religion are very strongly in favor of Barack Obama.  There’s basically a continuum from Protestants (and other non-Catholic Christians) to Catholics to Jews to those with no religion, from highest to lowest, in support for McCain. 

Of course, these are pretty large groups.  We can refine the analysis.  One important variable is race/ethnicity. 

Most blacks in America today are non-Catholic Christians (with their largest denominational affiliation being Baptist).  The vast majority of blacks – at or above the 90% level-- tell us they are going to vote for Obama in November.  When we remove blacks from the equation, the McCain advantage in the non-Catholic Christian group gets even larger.

Similarly, some of the overall Catholic skew towards Barack Obama is caused by Hispanic’s strong tilt in that direction. Remove Hispanics from the group of Catholics, and those left are slightly in favor of John McCain over Obama.

Then there is importance of religion.  Not all Catholics or Jews or Baptists are creatred alike.  Some take their religion seriously.  Others identify with the religious group in name only. Not shockingly, an in-depth analysis I recently completed confirms that holding everything else constant, the more religious an American is, the more likely he or she is to support McCain.

Looked at differently, the less religious a person is, the more likely that person is to support Obama. 

You ask:  “How do you measure how religious a person is?”  At this point, we rely on two trusty measures:  “Is religion important to you in your daily life?” and “How often do you attend church, synagogue, mosque or other worship services?”

But it really doesn’t seem to matter.  However you look at it, those who score higher on a religion scale are more likely to vote for McCain.

My analysis shows there are some subtleties in the data (as there almost always are).  If you are black, or Hispanic or have no formal religious identification, then it doesn’t matter how you score on our religion scales.  You have a high chance of voting for Obama.  Period.

But among most of the rest, religiosity is an important predictor, regardless of the brand name religion to which you adhere.

Of significant interest is the finding that this relationship holds even among American Jews.  Those Jews who say religion is important to them in their daily lives (about 40%) break about even in their support for McCain and Obama.  The majority of Jews who aren’t religious skew heavily towards Barack Obama. 

All of this means that we can do a pretty good job of predicting whom you are supporting for president by looking at you and asking a few totally non-political questions relating to religion.  If you are white, Protestant, and say religion is important in your daily life, we can predict you are voting for McCain.  If you are Hispanic and have no formal religious identification, you are most likely voting for Obama.  And so forth.

Highly religious Christians have formed a solid GOP block in the past, and McCain’s strategists would no doubt like to keep it that way, while at the same time expanding his appeal to less religious groups.

Obama’s people know their only strength among highly religious people comes within the black and Hispanic segments, groups for whom race/ethnicity appear to override the impact of religion. This state of affairs appears to be working for Obama, to the degree that he is leading McCain modestly in most recent national polling. Still, Obama's team would no doubt like to increase his penetration into more religious voting groups in the months ahead, something that looks like it’s going to be a tough challenge.   

1. The percentage of “swing” voters in this election is higher than at any point in 2004.  Furthermore, the percentage of swing voters who have a favorable view of the two major party candidates is much higher now than was the case in 2004.  This suggests that it may be easy for voters to switch back and forth between candidates this year.

2. Catholics tilt towards Barack Obama over John McCain by about a 4-point margin, but this Obama advantage is totally a result of the strong Obama skew among Hispanic Catholics.  Among non-Hispanic Catholics, McCain reverses the field and has a slight advantage over Obama.

3. American Jews strongly tilt towards Barack Obama, by over a two-to-one margin.  However, there is a strong impact of religiosity among Jews, just as there is among Protestants and Catholics.  The roughly four in ten Jews who say that religion is an important part of their daily lives are significantly more likely to vote for John McCainthan those who say it is not an important part, with the result that the two candidates split roughly even among this group. 

4. Half of Americans say they are “very” concerned that John McCain, if elected president, would have policies that are too similar to those of George W. Bush. In contrast, 30% of Americans are very concerned that Barack Obama would go too far in his effort to change away from Bush’s policies.  Of interest to McCain’s campaign consultants is the fact that Bush’s job approval rating among McCain voters is only at 55%.  This leads to the question of exactly how prominent a role George W. Bush should take in the McCain campaign, including in particular what Bush's role should be at the St. Paul Republican convention in early September.

5. There is no clear guidance from recent elections as to which of the two candidates is most likely to pick up steam between now and Election Day.  In late June, 2004, John Kerry was ahead of George W. Bush by one point among registered voters.  Kerry ended up losing to Bush by 3 points.  In late June 2000, George W. Bush led Al Gore by 7 points among registered voters.  Gore win on to win the popular vote, albeit by a very small margin.