Monthly Archives: July 2017

U.S. Military to Trump: “Military Policy isn’t made by Tweet”

By Kent R. Kroeger (Source:, July 26, 2017)

Through a spokeswoman, Iowa Senator Joni Ernst reacted swiftly to President Trump’s ban on transgender Americans serving in the military:

“(Sen. Ernst) has served with people from all different backgrounds and that gender is not a vital indicator of someone’s military prowess. She believes what is most important is making sure service members can meet the physical training standards, and the willingness to defend our freedoms and way of life. Americans who are qualified and can meet the standards to serve in the military should be afforded that opportunity.”

This is from a Republican senator that was seriously considered for the Vice Presidency by the Trump team. Add to the mix that Joni Ernst is a military veteran and is respected in the U.S. Senate for her deep knowledge of U.S. military personnel issues and her strong rebuke of Trump’s new ban becomes a politically significant moment.

Perhaps the most dramatic reaction to Trump’s latest policy-by-Tweet moment came from Utah Senator Orrin Hatch: “I don’t think we should be discriminating against anyone. Transgender people are people.”

That simple but profound comment comes from an 83-year-old man representing the very conservative state of Utah.

The speed with which Republican legislators denounced Trump’s transgender ban tells you everything you need to know about the political ramifications of this sudden policy change — a policy that caught even the Dept. of Defense’s public affairs office by surprise. There is no constituency for this new ban and Trump will find no comfort from his staunchest supporters in the military. They didn’t ask for this ban and they know how destructive this unexpected shift will have on thousands of U.S. military service people who currently serve and have done nothing to warrant this latest presidential fiat.

The Associated Press estimates there are between 2,500 and 7,000 transgender troops currently serving in the military with another 1,500 to 4,000 in the reserves.

Trump is misreading his own base on this one. As this blog presented recently, the transgender bathroom law controversy strongly divides Americans and this division correlates with partisanship and ideology. That should be no surprise.

Over 90 percent of the most conservative (“Right”) voters believe people should use the bathroom based on their birth gender, according the 2016 American National Election Study. In contrast, 95 percent of the most liberal (“Left”) voters think bathroom selection should be based on the gender with which a person identifies.

However, if Trump thinks attitudes on bathroom laws are isomorphic to attitudes on banning transgender people from the military, he will be disappointed.

In opinion research, one of the first things survey professionals learn is that ‘issue framing’ can dramatically alter how people respond to survey questions. The transgender bathroom law issue has been framed in the conservative media as not just a potential privacy invasion, but a palpable threat to the security of young children.

“Do you want a male sex offender using the same bathroom as your young daughter?” asked Joseph Backholm from the Family Policy Institute of Washington.

You can challenge the validity of Backholm’s inference (and I most adamantly do), but you can’t deny its potential power to mold the opinions of average Americans regarding transgender bathroom laws.

The politics around transgender individuals serving in the military has long been presented in a different frame — including within the conservative media that historically has opposed (until recently) U.S. military personnel policy protecting the rights of lesbian, gay and bisexual individuals to openly serve. Past opposition to their openly serving was most often framed around the construct of ‘troop morale and cohesion.’

Allowing lesbian,  gay and bisexual individuals to openly serve has been in place for six years now with no evidence of declines in troop morale, cohesion or readiness.

On Sept. 20, 2011, when the DoD policy of “Don’t Tell, Don’t Ask” (DADT) was lifted, the issue had in many ways already been decided by America’s evolving cultural attitudes and norms that were growing ever more accepting of the LGBTQ community in all aspects of American life.

For transgender individuals, however, the politics around their acceptance has taken longer and remains, as evidenced by Trump’s behavior, somewhat unsettled. The ban on transgender persons from serving was not lifted until June 30, 2016 and, like the ending of DADT,  has had no impact on the readiness of our armed forces.

And that is where framing becomes important to current attitudes on transgender Americans serving in the military. Even Fox News’ Tucker Carlson, who supports the new Trump ban due to its potential “costs” to the U.S. government stemming from possible transition surgeries — which is trivial compared to DoD’s $600 billion total budget — frames the issue around one of economics, not the security of our families.

That framing difference — our family’s security versus economics (and/or troop readiness) – will affect how Americans react to Trump’s new ban. My prediction is that a strong majority of Americans (let’s say, over 60 percent) will oppose Trump’s unilateral action against the transgender service members.

As of now, all we have are the swift and unequivocal comments of a significant number of congressional Republicans who are saying to President Trump: Your ban on transgender Americans serving in our military is unacceptable.


The 2016 ANES dataset, data dictionary, and computer codes (SAS JMP) used in this article are available upon email request to:

About the author:  Kent Kroeger is a writer and statistical consultant with over 30 -years experience measuring and analyzing public opinion for public and private sector clients. He also spent ten years working for the U.S. Department of Defense’s Office of the Under Secretary of Defense for Personnel and Readiness and the Defense Intelligence Agency. He holds a B.S. degree in Journalism/Political Science from The University of Iowa, and an M.A. in Quantitative Methods from Columbia University (New York, NY).  He lives in Ewing, New Jersey with his wife and son.

Democrats won’t ascend until they stop excluding millions of voters

By Kent R. Kroeger (Source:, July 23, 2017)

No matter how many times Democratic strategist Steve Phillips repeats his prediction about the Democratic Party’s ascendancy, it doesn’t make it so.

His latest evidence-based analytic train wreck was published by the New York Times and concludes that the mistake the Democrats (i.e., Robbie Mook, John Podesta, Hillary Clinton, et al.) made in 2016 was that they spent too much time and money worrying about working-class, white males and not enough directed towards getting ‘brown’ (his word) Americans to turnout for Clinton.

“The country is under conservative assault because Democrats mistakenly sought support from conservative white working-class voters susceptible to racially charged appeals,” intones Phillips. “Replicating that strategy would be another catastrophic blunder.”

Phillips apparently does not recognize the existence of moderate white working-class voters — many of whom voted for Barack Obama in 2008 and 2012 but decided to take their chances with Trump in 2016. Phillips lumps these voters in with conservative white working-class voters and then dismisses them as latent racists not needed for the Democrats to win elections in the future.

The Phillips message to the Democrats is simple: Brown (again, his word) people are your constituency; stop trying to appeal everyone.

Talk about judging someone by the color of their skin. Phillips and his many acolytes use skin tone as a direct proxy measure of someone’s attitudes, beliefs and behavior. Holy shit! Why is that not considered racism?! [Because its being done by a highly educated, well-meaning, New York Times-approved Democrat, you fool. By definition, liberal Democrats can’t be racist.]

Setting aside Phillips’ condescending reliance on racial determinism, the Phillips critique simply ignores the profound variation in attitudes and opinions among America’s minority voters.

As this blog has argued here and here, most African-American and Hispanic voters are qualitatively different from liberal Democrats. On average, these voters are consistently ‘centrists’ on issues such as national security, abortion and LGBTQ rights; and, frankly, don’t share the same values and experiences as the liberal elites Phillips envisions controlling U.S. politics for the foreseeable future. In many areas, African-American and Hispanic voter attitudes have more in common with the Republicans than with the Democrats.

For a deeper dive into the data, please check our earlier post (here), but two charts give evidence of the disconnect between liberals and minority voters.

After identifying ideological voter clusters from the 2016 American National Election Study (ANES), we find six types of voters: Left, Center-Left, Centrists, Libertarians, Center-Right, and Right.  All are comparable in size, each ranging from 14 to 21 percent of the voting population. When we look at the ethnic/racial composition of these six voter clusters, we see some interesting variation (see Figure 1 below).

Figure 1: Racial/Ethnic Composition of Voter Clusters

A plurality of African-American and Hispanic voters are Centrists — they are not ideologically on the Left. This has significant implications on the attitudes, beliefs, policy preferences, and voting patterns of many minority voters. Figure 2 shows the percentage of African-American and Hispanic Centrist and Left voters that supported Trump in the 2016 presidential race.

Figure 2: Trump Support by Race/Ethnicity (Centrists & Left)

Nine percent of Centrist Black (non-Hispanic) voters chose Trump as did 23 percent of Centrist Hispanics. There were very few Trump voters within these racial/ethnic groups on the Left — in fact, less than 0.5 percent. While 9 and 23 percent might not seem like large percentages, the Democrats must ask themselves, could those numbers go up for the Republicans in future elections, particularly if the Republicans can leverage issues where Centrists are more aligned with the GOP than the Democrats?

What are those issues?  Figures 3 to 6 are just a few examples:

Figure 3: Percent saying abortion should never be allowed (by race/ethnicity and ideological cluster)

Figure 4: Percent supporting marriage equality (by race/ethnicity and ideological cluster)

Figure 5: Percent feeling strongly that Transgender Americans should choose bathroom based on the gender they were born with (by race/ethnicity and ideological cluster)

Figure 6: Percent agreeing “a great deal” to allow Syrian Refugees in the U.S. (by race/ethnicity and ideological cluster)

I can summarize these last four charts in one sentence: Centrists, which are a majority-minority segment of the American voting population, hold many different attitudes and beliefs from the ideological (mostly white) Left.

Minority-group Centrists are much more conservative on issues such as abortion, LGBTQ rights, and immigration from the Middle East. As yet, they are still very loyal to the Democratic Party in the voting booth. But can the Democrats assume this will be the case going forward?

I wouldn’t make that assumption.

It is important to remind ourselves about the depth of the problem the Democrats created for themselves. The litany of electoral losses detailing the Democratic Party’s decline have been repeated many times in the past eight months. Along with control of the U.S. Congress and presidency, the Republicans control both chambers of the state legislature and governorships in 24 states — compared to only five for the Democrats. But this is just one point in time. The Democrats’ decline in the state legislatures is a long-term phenomenon (see Figure 7 below).

Figure 7: Number of state legislators by party since 1936

To blame Nancy Pelosi or Barack Obama’s leadership of the Democratic Party is unfair and myopic. With the exception of two temporary spikes in their favor (1974, 2006), the Democrats have been losing state legislators since 1978.

Why the decline? Political observers propose many reasons. The New Yorker’s Noah Rothman suggests the Democrats suffer from a self-imposed radicalism that keeps the base happy but alienates the majority of voters. The Atlantic’s David Graham argues structural disadvantages hold the Democrats down (e.g., gerrymandering, campaign finance laws). The decline of unions and their voter mobilization efforts is another possible reason.

All are probable contributors to the Democrats’ decline,  but there is another possible reason: Long-term American economic prosperity has put more and more voters into an income bracket that may lead to their favoring lower taxes and less government — the bedrock positions of the Republican Party’s agenda.

As seen in Figure 1, the Democrats’  long-term decline in the state legislatures starts in 1978. In the prior 1976 presidential election, Ronald Reagan had challenged a sitting Republican president (Ford) for the party’s nomination. Though he failed, Reagan’s intra-party revolt led to the party’s explicit re-branding, usurping the dominance of the moderate Rockefeller wing of the party and replacing it with an ideologically conservative agenda focused on less government, more freedom, and a strong national defense.

The Republicans under Reagan didn’t move to the “center” — which is what many of the Beltway pundits suggested they needed to do — instead, they made a sharp turn to the right.

Reagan’s re-branding of the Republicans holds to this day, frayed by time and success, but nonetheless largely intact. Consistency is important in politics. While some Republican candidates are better than others, voters know what they are getting when they vote for a Republican candidate. The Republicans give all of their candidates a factory-setting of “lower taxes, fewer regulations, and the strongest military in the world.”

Like it or not, that message works. But does this suggest the Democrats need to act analogously and resist calls to move to the “center”? Instead, should the Democrats make a clear shift to the left and endorse progressive policies similar to those promoted by Bernie Sanders in his 2016 presidential campaign?

If the new Democratic slogan announced this week by New York Senator Chuck Schumer is an indication — A Better Deal: Better Jobs, Better Wages, Better Future  — they are still struggling to find that message. While standing against corporate mergers, lowering prescription drug prices, and funding apprentice and continuing education programs are admirable policy goals, they lack the digestible theoretical foundation of the Republicans’ free market message strategy.

It is interesting that the new Democratic slogan ignores civil liberties, suggesting that this battle continues within the party’s leadership ranks.

Does that mean the Democrats need to become more ‘centrist’ if they want to compete more effectively with the Republicans? Hell yes! If their goal is to win elections on a more consistent basis, that is.

The now classic Saturday Night Live skit with Tom Hanks as a contestant (‘Doug’) on ‘Black Jeopardy’ was funny because of that well-known reality. And if Democrats want to find the next durable political coalition in this country, they are better off listening to Keeley, Shanice and Doug than anything coming from Phillips’ data analyses or Tom Perez’ Democratic National Committee.

If Rand Paul and his libertarian, non-interventionist cohorts could find a way to accept the legitimacy of government interventions to ensure equal access for all Americans to the ‘American Dream’ and concomitant protections from its vagaries, we might see the rise of a profound and enduring political majority.

But Steve Phillips doesn’t want THAT Democratic majority coalition because it would require real compromise, particularly on the economic and social issues that have been hard-coded into the Democrat’s liberal core as being the essential elements of political enlightenment.

The liberal brain trust with whom Phillips associates, clustered in our coastal metropolis’ and living in socially-approved, class-based segregated communities, think they know enough about working-class whites to know they don’t want them in their ruling coalition.

This arrogant assumption of ideological unity within the broad ranks of the Democratic party addles the Democratic Party’s current leadership — and this arrogance has real consequences.

The Democratic National Committee’s (DNC) underwhelming recent fundraising numbers are additional  evidence that the Resistance is better at scaring Americans away from supporting the Democratic Party than it is at ending the Trump presidency. The DNC’s fundraising is down, while the Republican National Committee is raking in record sums of cash.

It didn’t help the DNC’s finances by recklessly pouring $22 million into a Georgia U.S. House race. Tom Perez’s obscenity-laced tirades against Trump have failed not only to energize the Democratic base, it has turned off donors to the party. Just watch this video of Tom Perez “energizing” the base. The painful look on the faces of those kids standing behind him, forced to listen to his meandering screamfest of a speech, says it all.

Where should the Democrats go from here?

Phillips and other “Turn Lefties” promoting the letting go of white, working-class voters from the Democratic coalition think they doing the equivalent of Reagan’s ideological revolt in the mid-1970s. They are not.

Reagan’s conservative agenda moved TOWARDS the majority of Americans, not away. Reagan had a coherent, empirically-supported organizing theory (‘free markets’) behind his candidacy. Apart from assuming all white, working-class voters are ‘racists,’ the “Turn Lefties” have no guiding principles. For them, it all comes down to assigning voters into ethnic and race categories and adding up the numbers.

The problem is that this strategy works for the Democrats when they run a charismatic, centrist presidential candidate (Bill Clinton, Barack Obama), but fail spectacularly when they run Left-leaning or uncharismatic candidates (Walter Mondale, Michael Dukakis, Al Gore, John Kerry, Hillary Clinton).

By keeping their wallets closed (for now), the Democrats’ big donors are sending a message to the DNC. The endless deluge of Tom Perez and Keith Ellison emails begging them to “save the Resistance and our democracy from the treasonous Donald Trump.

Everyday, like the big Democratic donors, rank-and-file Democrats are rejecting the breathless hype of their compatriots on the Left. American companies are as profitable as ever, American air and waterways have never been cleaner, ISIS is almost defeated and the slightly warmer weather means more available weekends at the nearest beach or park. But despite these good news items, there is still a lot of problems in this country (income inequality, access to health care, deteriorating infrastructure, college costs, etc.) that are often exacerbated by Republican policies.

Running against Trump will work in 2018. The Democrats will, short of an act from God or the Kremlin, regain control of the U.S. House and make inroads in the U.S. Senate. But running against Trump won’t reverse the long-term trend working against the Democrats.

The Democrats need a new message, a new brand strategy, and new leaders (Other than that, they are in good shape!). But writing off white, working-class voters is not a step in the right direction, even if this group’s numbers are in relative decline. By ignoring these voters, the Democrats may also push out core Democratic voters from other race and ethnic categories — and the Democrats can’t afford to do that.


The 2016 ANES dataset, data dictionary, and computer codes (SAS JMP) used in this article are available upon email request to:

About the author:  Kent Kroeger is a writer and statistical consultant with over 30 -years experience measuring and analyzing public opinion for public and private sector clients. He also spent ten years working for the U.S. Department of Defense’s Office of the Under Secretary of Defense for Personnel and Readiness and the Defense Intelligence Agency. He holds a B.S. degree in Journalism/Political Science from The University of Iowa, and an M.A. in Quantitative Methods from Columbia University (New York, NY).  He lives in Ewing, New Jersey with his wife and son.

He’s may be the Worst President Ever and I’d Vote for Him Again

By Kent R. Kroeger (Source:, July 17, 2017)

The TV at the bar is tuned to CNN whose scroll is announcing that President Trump has hit a new approval rating low.

The people around the bar with me start launching anti-Trump quips.

“He can’t do the job,” the waitress says. “I could do it as well as he can.”

The same news scroll continues with other findings from the newest Bloomberg Poll. Health care is considered the most important problem by 35 percent of U.S. adults.

That’s good news for the Democrats, I say to myself. My blogsite,, recently published an analysis of the 2016 American National Election Survey showing that health care is one of the few issues where the Democrats have a strategic advantage over the Republicans (…the other Democrat-favorable issues are climate change and government spending on social programs such as Social Security).

The scroll continues…

Besides health care, unemployment (13%), terrorism (11%) and immigration (10%) also remain high on people’s list of concerns, while U.S.-Russian relations is mentioned by only 6 percent of respondents.

That is bad news for the Democrats. Eight months of 24-7 Trump-Russia coverage and very few people seem to care. Are people burned out from the whole thing?

“I don’t care about the Russia-Donald Trump junior thing,” says Sam, a 29-year-old landscaping contractor from Ewing, New Jersey, who calls himself politically independent but admits he almost always votes Democrat. “Its not good if Trump was getting help from Putin to beat Clinton, but that stuff happens in politics.”

“He’s a con artist,” says Cathy, a 32-year-old waitress and a ‘proud feminist and Democrat.’ “But the election is over. We need to worry about what he’s going to do to health care, not whether or not his son and son-in-law got cozy with the Russians.”

Health care really seems to get people talking, in a way other political issues don’t, except for perhaps war and terrorism.

“I didn’t vote for the guy,” says Ben, a 45-year-old real estate broker – the only registered Republican in my impromptu focus group around the bar at Houlihan’s in Lawrence Township, New Jersey. “But I’m sick of all the negative stuff coming from both parties. I just think our politics is broken and the media feeds off it. I don’t watch the news anymore.”

While there is no evidence (yet) that the cable news networks are seeing a ratings drop, the one constant from the Houlihan’s lunchtime patrons is that ‘Russia’ is a more of a distraction than a major concern to people.

Tim, a retired municipal worker and Marine Corps veteran, may fit some pundits’ perceptions of the classic ‘angry white male’ that put Trump in office. He did, admittedly, vote for Trump largely due to his stands on immigration and terrorism and admits he is ‘pissed off’ at the politicians in Washington, D.C. and Trenton. But he insists he’s not angry.

“He (Trump) says things that typical politicians are scared to say,” says Tim. “Radical Islamic terrorists would love to take away our freedoms and you can’t fight terrorists with nice words and hope.”

Tim voted for Obama twice.

What should worry the Democrats, however, is what Tim says about the next presidential election.

“I think Trump has brought in millionaires and billionaires that don’t know what they’re doing. He may be the worst president ever,” says Tim as he starts his second beer of the afternoon. “But you know what? If I had to choose between Trump and Clinton today, I’d vote for Trump again.”

Tim’s declaration attracts a mix of approving nods and rude rebuttals from the other Houlihan’s patrons.

Donald Trump may be the first truly ‘teflon’ politician — a term typically associated with Ronald Reagan, who weathered years of media attacks and still remained popular among the majority of Americans. In Trump’s case, the term may be even more appropriate. I am talking to people that don’t like him, that think he’s unprepared and unqualified to be president, and are still open to voting for this guy in 2020. That, to my way of thinking, is a ‘teflon’ candidate.

“What if it were Obama versus Trump?” I ask.

“Obama. No question,” Tim replies.

“I’d vote for the good-looking Obama daughter,” yells someone from one of the booths lining the bar.

“Malia versus Ivanka,” someone else chimes in from another corner of the bar. “That’s the election I want to see.”

Sasha Obama is very good-looking young woman as well, but that point is not going to penetrate this crowd at this moment.

“Kid Rock versus The Rock! That would be awesome.”

No, it really wouldn’t.

The informal focus group deteriorates into a bad open-mike night and I ask the bartender for my check.

“Trump versus Wonder Woman,” a waitress says quietly as she walks by.

Yeah, I’d like to see that one. Trump says something rude to her during a debate about ‘jiggly thighs’ and she promptly gives him a vicious leg swipe. Now THAT would be awesome.

I pay my bill and leave the restaurant.

As I drive back to the home office, I wonder, is the lunchtime crowd at an over-priced chain restaurant representative of voting Americans? I suspect not…but, I’m just as certain these opinions are not uncommon either.

Not here in a Democrat-dominated, western New Jersey suburb.

Trump’s polling numbers look grim, but we are more than a year removed from the 2018 midterm elections. More than enough time for the Democrats to screw up. More than enough time to make the 2018 election just as inhospitable to their candidates as the last couple of national elections.

Diana Prince. If you are reading this, please give Tom Perez at the Democratic National Committee a call at your earliest convenience.

About the author:  Kent Kroeger is a writer and statistical consultant with over 30 -years experience measuring and analyzing public opinion for public and private sector clients. He also spent ten years working for the U.S. Department of Defense’s Office of the Under Secretary of Defense for Personnel and Readiness and the Defense Intelligence Agency. He holds a B.S. degree in Journalism/Political Science from The University of Iowa, and an M.A. in Quantitative Methods from Columbia University (New York, NY).  He lives in Ewing, New Jersey with his wife and son.

The Sam Kinison Rule of Politics: Move to the Voters

By Kent R. Kroeger (Source:, July 13, 2017)

The late (great) comedian Sam Kinison was famous for a bit he did about African hunger where he suggested — in primal scream tones — the starving people should just “MOVE TO WHERE THE FOOD IS!!!

Kinison was often crass and his African hunger bit wouldn’t survive the filtering and obligatory over-analysis of today’s 24-7 media, but its core logic would nonetheless survive: If things aren’t working for you, maybe its time to change what you are doing.

That piece of common sense is what I took away from Mark Penn and Andrew Stein’s July 6th New York Times op-ed piece: Back to the Center, Democrats. And their sentiment is hardly new or controversial. But, judging from the reaction on social media, you would of thought they had just endorsed Ivanka Trump for President in 2020.

Here is a just a sampling from Twitter:

Beyond the name-calling and careless misrepresentations of the Penn and Stein piece (Keith, do you really think Hillary could have run as the “Change” candidate in 2008? Seriously?), there is an intellectual dishonesty on display among these Democrats that I fear presages what could become the greatest ‘missed opportunity’ in U.S. political history. If the Democrats do not regain control of the U.S. House in 2018, they can look back at how they reacted to advice like that offered by Penn and Stein (and pollster Doug Schoen and many others, frankly).

I have some issues with the Penn-Stein thesis, but the Stephen Pimpare complaint that they lack sources, reference and data is a weapons-grade level of wrong. If anything, Penn and Stein are too gentle in urging the Democratic Party to take their foot off the progressive-agenda pedal. In suggesting the Democratic Party needs to re-engage with the nation’s political center — where most voters reside — Penn and Stein have the data overwhelmingly on their side.


First, as documented by this blog (here), it is those calling for the Democrats to move more definitively towards the progressive Left that are ignoring or misreading the data.

The two common mistakes made by political writers (such as  Wired’s Issie Lapowsky and The Atlantic’s Peter Beinart) are: (1) cherry-picking the polling data to highlight those issues where the progressive Left is aligned with the majority opinion (gun control, abortion rights, civil rights, increasing the minimum wage), and (2) relying too much on demographic determinism to forecast future electoral outcomes (e.g., Hispanics vote overwhelming for Democrats, as Hispanics increase as a percentage of the voting population, Democrats will be the primary benefactor).

The “Emerging Democratic Majority“-thesis is an example of this second error as it ignores significant opinion diversity within the identity groups most often associated with the Democratic voting base (African-Americans, Hispanics, LGBTQ community, Millennials). More importantly, these analyses under appreciate the opinion dynamics that often accompanies changes in people’s life-stage, income, and age. As we detail below, recent opinion data from African-American and Hispanic voters reveals significant distance between these voters and the ideological Left on a number of Democratic-core issues (Abortion rights, LGBTQ rights, immigration, Syrian refugees).

In the past year, has looked across hundreds of opinion surveys and forecast models related to political attitudes, beliefs and voting behavior in the U.S., and there is one clear conclusion: The average American voter doesn’t fit well into either the liberal or conservative ideological camps.

Yet, they are not apolitical either. They care about politics (when they need to), even if they don’t adhere to any strict ideological rules like those followed by activists from the left or right. While our elected leaders are more partisan than ever in their voting patterns, it is debatable as to whether the American voting population has grown more partisan. Pew Research cites polling data that suggests voters today are, in fact, more partisan. But other academic studies point out the extraordinary consistency over time on the centrist-leanings of American voters.

Our recent analysis of The University of Michigan and Stanford University’s 2016 American National Election Study (ANES) provides more current data on what Americans believe and how it relates to their political choices.

It is not good news for ideologues of the political Left or Right. But for the current discussion, we will concentrate on what the data suggest for the Left.


First, It is helpful to understand the analytic approach we employed with the 2016 ANES data. As noted above, cherry-picking issues for analysis is one of the most common errors committed by political pundits and researchers. While we recognize that some issues are more important than others, in our analysis we chose to summarize voters’ opinions on over 100 equally-weighted issues measured in the 2016 ANES. Our goal was to group voters into relatively homogeneous ideological clusters based on their survey responses on a wide range of social and political issues. We, therefore, did not rely on respondents’ self-reported location on a binary ideological scale (liberal versus conservative).

We do recognize that in choosing some issues and not others in their survey questionnaire, the 2016 ANES researchers engaged in another type of analytic ‘cherry-picking.’ Our response is, ‘You have to work with what you are given.’

For clustering voters, we employed a K-means clustering method using SAS’s JMP 13.1.0 statistical software. For clustering variables, we employed JMP’s Cluster Variable procedure. (Our 2016 ANES dataset and analytic algorithms are available by sending a request to: )


After clustering the 3,649 respondents to the 2016 ANES, we found that only 14 percent of respondents could be classified as consistently “Left” in their opinions (see Figure 1 below). If we include respondents with mostly Liberal/Leftist opinions, 35 percent of American voters can be considered part of the ideological Left in this country.

In contrast, the ideological Right accounts for 15 percent of the voting population, and 33 percent if you include people that are mostly Conservative/Rightist in their opinions. That leaves the remaining third of voters in the ideological center.

Figure 1: The Six Voter Clusters

(Source: 2016 ANES; 3,649 respondents; max. MoE = +/- 4.3%; analytics by


The clusters, by construction, are distinct in their opinions. It is no surprise therefore that the Left is most supportive of civil rights issues (LGBTQ, abortion), immigration, increased government spending on social programs and climate change. In contrast, the Right is less supportive of government spending on social programs and climate change, and more supportive of government action on security, terrorism and stopping illegal immigration.

Table 1 below provides the descriptive highlights of the opinion orientations for each cluster:

Table 1: Opinions of Voter Clusters

Opinion Mix
LEFTPro-choice, pro-amnesty, support LGBTQ rights, support increased gov't spending on social programs, support gov't role in climate change.
CENTER-LEFTSimilar to LEFT, except they are more concerned about security issues (crime, illegal immigration).
CENTERSimilar to LEFT, except much less supportive of LGBTQ rights, abortion rights, and allowing more Syrian refugees into U.S..
LIBERTARIANSimilar to RIGHT, except are generally opposed to foreign military interventions, are much more supportive of civil rights issues (LGBTQ, immigrants) and are concerned that fighting terrorism will compromise civil liberties
CENTER-RIGHTSimilar to RIGHT, except they see a stronger role for the government in addressing poverty and income inequality and are more concerned about climate change than either the Right or Libertarian clusters. This cluster contains is the populist wing of the Republican Party.
RIGHTSupport strong military, support increased intervention in Syria/Middle East to fight terrorism, oppose amnesty, oppose abortion, oppose increased social spending, favor less government, favor lower taxes, oppose gov't role in climate change


When we look at the demographics and voting patterns of these six clusters, America’s political factions are recognizable. The Left, Libertarian, Center-Right and Right are each over 70 percent white (see Figure 2 below). Centrists, conversely, are a majority-minority cluster. In fact, a plurality of African-Americans and Hispanics are Centrists. If that doesn’t shake the foundations of the Democratic National Committee (DNC), nothing will. This is a remarkable conclusion, though hardly the first time this feature of African-American and Hispanic public opinion has been observed. The 2008 passage of California’s Proposition 8, a law that — had it not been struck down by the federal court — would have made it illegal for same-sex couples to marry, occurred in part due to strong support within the African-American and Hispanic communities.

This disconnect on some social issues between the DNC platform and minority voters does not suggest the Republicans have a chance to win a majority of African-American and Hispanic votes anytime soon. They do not. However, the vote turnout among these groups can change election outcomes in many localities and that is where the DNC should be concerned.

Figure 2: Voter Clusters by Race/Ethnicity

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


In terms of education, the Left is by far the most educated voter cluster with just over 30 percent of its members having  an advanced college degree (see Figure 3 below). Centrists are the least educated; however, this a partly a function of this cluster being one of the youngest of the six voter clusters (see Figure 4 below). The Center-Right cluster is similar to Centrists in education with less than 20 percent having a 4-year college degree. The Right has the oldest members and are more likely to be married (see Figure 5 below). Centrists are the least likely to be married.

Figure 3: Voter Clusters by Highest Level of Education

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


Figure 4: Voter Clusters by Age Group

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


Figure 5: Voter Clusters by Marital Status

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


Figure 6 (below) shows the partisan makeup of the voter clusters and provides some validation to the clustering segmentation itself. Over 75 percent of the Left are registered Democrats and over 20 percent are independents. Almost 30 percent of the Center-Right are registered Democrats. Along with the Center-Left cluster, this the segment where we find many of the infamous white, working-class Trump voters. But, based on their self-reported attitudes, it shouldn’t surprise us that these voters would lean toward Trump. They are conservative.

Over 80 percent of the Right are registered Republicans. Not surprisingly, a plurality of independents are in the Libertarian cluster, however four of the six segments are comprised of 20 percent or more independents. Centrists are the second most likely to be registered as Democrats. Obama did much better than Clinton in poaching Center-Right and Libertarian voters from the Republican candidate and he defended his core (Left, Center-Left and Centrist) much better as well.

Figure 6: Voter Clusters by Party Registration

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


Figure 7 (below) further validates our segmentation and shows why Trump did as well as he did. Almost 20 percent of Centrists voted for Trump – a critical inroad for him given the closeness of the election — even though Clinton did even better among Libertarians than Trump did with Centrists. This was a close election and small marginal shifts affected its outcome.

When you look at these same clusters and how they voted in 2012, you see why Obama won in 2012 and Clinton lost in 2016 (see Figure 8 below). Obama held the center together much better than Clinton.

Figure 7: Voter Clusters by Presidential Vote (2016)

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


Figure 8: Voter Clusters by Presidential Vote (2012)

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)



Four issues addressed in the 2016 ANES highlight the significant risk facing the Democrats in future elections:  Transgender bathroom law, abortion, marriage equality and immigration.

At risk is the support of Centrists, the majority of whom are African-American or Hispanic. No, they are not going to vote majority Republican. But their enthusiasm, which drives their voter turnout probabilities, is crucial to future Democratic victories. This fact cannot be over-emphasized: African-American and Hispanic voters are not liberals on social issues – they are centrists. In the aggregate, they hold too many opinions outside the acceptable parameters of the activist Left.

Figures 9 through 12 (below) give support to this conclusion. On the issues of transgender rights, abortion, marriage equality and Syrian refugees, Centrists have more in common with the Center-Right populists than they do with the Left. And its not close.

On the issue of transgender bathroom laws (Figure 9), 50 percent of centrists strongly support laws that say people should use the bathroom of the gender they were born with. That is comparable to the Center-Right (58%) opinion. Libertarians, in contrast, are closer to the Center-Left‘s opinion on the issue.

Figure 9: Voter Clusters by Attitudes Regarding Transgender Bathroom Issue

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


On marriage equality (Figure 10), a similar pattern emerges. Roughly 35 percent of Centrists say there should be no legal recognition of gay or lesbian marriages. That is similar to the Center-Right (27%) and Right (38%). And, again, Libertarians show more congruence with the Center-Left and Left  on this issue.

Figure 10: Voter Clusters by Attitudes Regarding Marriage Equality

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


Abortion is an issue that some have argued has reached a consensus in this country. Americans, if they personally don’t approve of abortion, are not inclined to outlaw the practice. On the contrary, American’s generally support safe and affordable access to the procedure in most cases. Yet, there is a clear disconnect between the Left and Center-Left with Centrists where 25 percent say abortion should never be permitted. That level is closer to the Right‘s position (30% oppose abortion in all cases) than it is to even Center-Right position (15% oppose in all cases).

Are Hispanic Catholics going to split from the Democratic Party over this issue? We don’t think so. But given the party’s recent treatment one of Omaha’s Democratic mayoral candidates, Heath Mello, who had a mixed record on abortion rights, it also appears the party doesn’t tolerate diversity of opinion within its own ranks.

Figure 11: Voter Clusters by Attitudes Regarding Abortion

(Source: 2016 ANES; 3,649 respondents; max. margin of error = +/- 4.3%)


No issue demonstrates how disconnected the Left is from the rest of America as does the willingness to accept more Syrian refugees into the U.S.  Eighty-percent of the Left supports allowing more Syrian refugees into the U.S. (see Figure 12). Only 32 percent of Center-Left voters and 12 percent of Centrists share that opinion. And I’ll let you guess where the Center-Right and Right come down on this issue.

Understand. I am not making an argument that the U.S. should stop accepting more Syrian refugees. This author absolutely is willing to stand against the majority opinion on this issue. Sometimes, right is right. George W. Bush, Barack Obama and Hillary Clinton (among others) created that mess and, in my opinion, it is our moral obligation to take on its consequences. That opinion, however, is not the opinion of most Americans. Its not even close.

The exercise presented here is to objectively assess the degree to which the ideological wing of the Democratic Party is in aligned with its own partisans and the American people in general. Since persuasion has become a lost art within both parties, assessing the party’s alignment with its core constituencies is critical to its future electoral success.

Figure 12: Voter Clusters by Attitudes Regarding Allowing More Syrian Refugees into U.S.



Should the Democrats move to the Center? To answer this question, so far, we’ve grouped people with similar opinions on a wide range of issues and looked at their demographics, their voting behavior, and their opinions on a selection of issues.

But the Penn-Stein thesis requires more rigorous testing. First, does a two-dimensional ideology construct (Left-Right) represent the real world? That is, does it even make sense to talk about ‘moving to the center’?

Second, if the two-dimensional construct does offer some value, is the distance between the Left and the other voter segments large enough to suggest the Left needs to move closer to the center. Figures 13 to 17 are a preliminary attempt to address these two questions.

Third, do voters really choose candidates based on issues, or are other factors such as trust and likeability more important in their vote calculus? On this question, we will rely on previous research but’s own analysis of the 2016 ANES shows trust was the most powerful predictor of vote choice in the 2016 presidential race — though factors such as a voter’s party identification, presidential approval, economic assessments, and attitudes regarding Syrian refugees and transgender bathroom laws were also significant predictors.


As mentioned earlier regarding our methodology, we clustered not only the voters in the 2016 ANES (6 clusters), we also clustered the issues into 18 sub-clusters that we further combined into five major issue clusters:

  • Role of Government (social spending, deficits, role of gov’t in health care)
  • Security (national defense, terrorism, crime, spending on police)
  • Immigration (building a border wall, amnesty, importance of English)
  • Civil Rights (marriage equality, gender equality, race relations)
  • Climate change (belief in global warming, fracking, regulation)

If the Left-Right ideology spectrum is meaningful, we would expect the six voter clusters to fall predictably within the issue indexes, particularly the most ideologically consistent voter clusters (Left and Right).  And that is what we find in the graphs below — though,  with respect to Libertarians and Centrists, there was more variability in their mapping locations. On some issues (Role of Government), the Libertarians are more aligned with the Right. On other issues (Civil Rights), they are more aligned with the Left. 

The Libertarian example suggests it takes more than two-dimensions to summarize public opinion. Nonetheless, if we look at Figure 13 as an example, two-dimensions still work pretty well. On immigration (the y-axis), the Left is far removed from the next closest cluster (Center-Left) on this issue. In this case, high scores on the y-axis indicate strong support for immigration policies that give illegal immigrants a path to citizenship and allow for more open immigration policies in general. The same is true for the security issue (x-axis), except now the Libertarians are closest to the Left, which is still far removed from the other five voter clusters.

Notice that the Right and Center-Right are very close to each other on both immigration and security, and both are far removed from the Libertarians, Centrists, and Center-Left. For the Republicans to gain a consensus on either issue, they will need significant support from Libertarians (and maybe even Centrists). The Left, on the other hand, has a tougher road ahead in building a consensus position on immigration and security.

Figure 13: Voter Clusters by Immigration and Security Dimensions

(Source: 2016 ANES; 3,649 respondents; analytics by


Figure 14 is similar to Figure 13, only the Security Index has been replaced with the Civil Rights Index. It shows that the Right and Center-Right are still in alignment on civil rights issues and immigration and the Left is far removed from even the Center-Left on both indexes. The correlation of the voter clusters plotted by the Immigration and Civil Rights Indexes is 0.94, indicating support for the utility of the two-dimensional ideology construct.

Figure 14: Voter Clusters by Immigration and Civil Rights Dimensions

(Source: 2016 ANES; 3,649 respondents; analytics by


Figure 15 looks at the relationship between the Role of Government and Civil Rights Indexes. Though the correlation remains high (0.76), there is a bit more scatter and the Libertarians are closer to the Right on the Role of Government Index. In contrast, the Center-Right, Centrists, and Center-Left are closer to the Left than they are to the Right.

This is good news for the Democrats.

Figure 15: Voter Clusters by Civil Rights and Role of Government Dimensions

(Source: 2016 ANES; 3,649 respondents; analytics by


Figure 16 may be the most intriguing of these bi-plot graphs as it shows the potential contradictions between security issues and the role of government.  For many voters, it requires some intellectual gymnastics and perhaps a little temporary amnesia to rationalize support for big government with respect to security but dismiss its relevance on most other issues. Its certainly not impossible to believe the government is good at national defense, but not much else. But Figure 16 implies that rigid ideological arguments breakdown somewhat with respect to the guns versus butter debate.

Figure 16: Voter Clusters by Role of Government and Security Dimensions

(Source: 2016 ANES; 3,649 respondents; analytics by


Finally, ideological consistency returns in Figure 17 which plots the voter clusters by the Role of Government and Climate Change Indexes. The good news for Democrats here — and it really is good news — is that the Right is the most distant cluster from mainstream opinion on climate change. Now, if only the Democrats can make climate change policy an important factor in how people vote.


Figure 17: Voter Clusters by Role of Government and Climate Change Dimensions

(Source: 2016 ANES; 3,649 respondents; analytics by

Though we haven’t addressed the relative importance of the various issues on how people vote, The Gallup Company provides a nice summary over time on that question. You can access their data here.

Generally, economic issues (jobs, economic growth) and national defense tend to emerge near the top of voters’ lists of most important issues. But those lists do vary by election year and the value of predicting what those “most important” issues will be cannot be underestimated. Ironically, with the current strong economy and significant progress in fighting ISIS, the 2018 midterms may see post-materialist issues like civil rights and climate change rise up on voters’ agenda, potentially benefiting the Democrats in addition to the likely importance of the ongoing Trump-Russia investigation and the effects of the health care debacle the Republican Congress is about to unleash on the American public.


The Penn and Stein thesis carries a large assumption about the relationship between candidate’s issue positions and voters. This is called the ‘representative democracy’ model of voting. In making their vote decision, do voters line up the candidates’ issue positions with their own and choose the candidate closest to them on the most important issues?

It appears the answer is ‘yes’ — for most voters.

Political scientist Jon Krosnick, a leading expert on American voting behavior, says the political psychology research on American voters offers an optimistic conclusion about the American democracy. “Most Americans vote according to the principles of representative democracy,” says Krosnick. “but guardianship democracy and performance appraisal are (voting) approaches alive and well, too.”

A guardianship democracy is where voters vote for the most competent or intelligent candidate. In this case, trustworthiness and experience are drivers of vote decisions and, for some voters, their policy preferences may not be well-aligned with their preferred candidate. Performance appraisal voting is similar in that voters keep incumbents in office when things are going well (e.g., the economy) and vote for challengers when things aren’t going well.

So, can we conclude that Penn and Stein’s advice to the Democrats is built on the valid assumption that voters seek candidates they agree with on the most important issues?

Not so fast.

Recent research by Stanford’s Toni Rodon (here) offers evidence from European voting data that centrist voters are less likely  to vote when the parties don’t show clear ideological differences. If the parties aren’t different enough, why vote? Though Rodon presents his conclusion as debunking conventional wisdom, Ronald Reagan’s pollster, Richard Wirthlin, discovered something similar over 40 years ago when  Reagan was preparing to run against President Gerald Ford, a moderate Republican.

In his 1975 speech to Young Americans for Freedom, Reagan urged Republicans to “raise a banner, not of pale pastels, but bold colors.” His prescient observation rose from Wirthlin’s research showing many undecided voters, all else equal, are more attracted to candidates with a clear policy agendas.

If this voting dynamic is prevalent with a significant number of voters (and we think it is), it would argue against running centrist candidates and support the conclusion that the Democrats need to be the clear progressive alternative to the Republican Party.

This is a seductive argument and may help explain why “moderate” Democratic presidential candidates (Dukakis, Gore, Kerry, Hillary Clinton) failed to capture the imagination of enough American voters.


That is the challenge the Democrats will face if they choose to follow Penn and Stein’s advice. To make that move poorly could make the Democratic Party and its candidates appear indecisive and triangulating. That is not an attractive position to be in. Ask Hillary Clinton or John Kerry.

However, talented communicators like Bill Clinton and Barack Obama have proven the centrist strategy can work for Democrats. Americans will vote for candidates they disagree with on some important issues if that candidate presents a clear vision of where they want to lead this country.

But we have not seen any data that supports the conclusion that a political party can be too far outside mainstream opinion and consistently win elections at all levels of government. Unfortunately, that is where we think the Democrats stand today. They are out-of-sync, not just with independents, but with some of their core constituencies (African-Americans, Hispanics, and the working-class). On some issues, such as national security, abortion, and LGBTQ rights, these core constituencies are closer to the Republicans in their attitudes and policy preferences.

Should the Republicans ever learn how to exploit this fact, even the advantage the Democrats have going into the 2018 midterms could be in jeopardy. The good news for Democrats is that the Republicans have shown no ability or commitment to make this adjustment. If anything, the Trump presidency has set the Republican Party back in this regard.

Regrettably, too many Democrats are showing a similar disinterest in appealing to the moderate predispositions of a large proportion of their core supporters. It makes me think…somewhere…in some other dimension…Sam Kinison is screaming at the Democrats, MOVE TO THE VOTERS!


The 2016 ANES dataset, data dictionary, and computer codes (SAS JMP) used in this article are available upon email request to:

About the author:  Kent Kroeger is a writer and statistical consultant with over 30 -years experience measuring and analyzing public opinion for public and private sector clients. He also spent ten years working for the U.S. Department of Defense’s Office of the Under Secretary of Defense for Personnel and Readiness and the Defense Intelligence Agency. He holds a B.S. degree in Journalism/Political Science from The University of Iowa, and an M.A. in Quantitative Methods from Columbia University (New York, NY).  He lives in Ewing, New Jersey with his wife and son.