Barack Obama: Buyer's Remorse

Statistical work from Paul Lukasiak on Obama's declining support, from February, as people "got to know him."

Buyers' Remorse: How Rank & File Democrats Rejected Obama Once He Was Declared The "Inevitable" Nominee

[Welcome Political Radar readers. This Memorial Day Weekend, try a round of Obama Golf! --lambert]

PART ONE: ALL VOTERS, GENDER, AND RACE

Ever since the media declared that Barack Obama was “inevitable” after February 19th, based on a two week period when an unprepared Hillary Clinton campaign suffered “10 straight losses”, rank and file Democratic voters have been sending a message. Rather than rally ‘round the “inevitable nominee” that message has been a consistent, loud, and clear message to the Democratic Party – DO NOT WANT.

In nearly every demographic category since February 19, Clinton's percentage of the vote has risen, while Obama’s has fallen. This includes Obama’s supposed “strong” demographic categories such as voters with college degrees post-graduate degrees and voters whose income is above the national median. And Clinton beat Obama in the primaries in March, April and May in most of the major categories.

In the aftermath of Super Tuesday, John McCain was anointed by the media as the inevitable nominee – and with good reason. McCain had accumulated 740 of the necessary 1129 “pledged” delegates necessary to clinch the GOP nomination, and all he had to do was win 40% of the remaining delegates against two “non-mainstream” Republican challengers (Huckabee and Paul). Rank and file Republicans accepted McCain as their nominee, and McCain won every contest held subsequent to Super Tuesday with the exception of the Louisiana primary held on February 9th..and that contest he lost by only 1% (43% to 42%).

McCain may not have been the choice of the majority of Republicans, but once he was declared the “inevitable nominee”, rank and file Republicans closed ranks behind McCain. Despite doing virtually no campaigning at all, McCain has been able to garner at least 50% of the vote in every other primary contest held subsequent to Super Tuesday.

But Democratic voters refused to accept the pronouncements of the pundits and “analysts”, and have voted in overwhelming numbers in support of Hillary Clinton. Not only did Clinton pick up the support that Obama lost, Clinton has picked up a lot of the support that, in February, had gone to other candidates. Moreover, the electorate in the Democratic primaries looked a lot more like the “general electorate” in the 2004 Presidential election.

ALL VOTERS, GENDER, RACE—CHANGES BETWEEN FEBRUARY and MARCH- MAY PRIMARIES

Once voters realized that Obama would be the nominee, his support within most demographic categories declined, and declined significantly in most cases. Obama’s support declined overall by 2.3%

CHART 1A

TABLE 1 (data for Charts 1A & 1B)

ALL, GENDER, RACE
 	CHANGE IN SUPPORT	NET CLINTON GAIN/LOSS
 	       CLINTON	OBAMA  Weighed  Unweighed 
all	         +4.5%	-2.3%	 	+6.7%
Male 	         +6.0%	-2.6%	+3.7%	+8.6%
Female 	         +3.1%	-1.2%	+2.5%	+4.3%
White men 	+10.8%	-6.7%	+4.9%	+17.5%
White women 	 +6.0%	-3.6%	+3.5%	+9.6%
Black men 	 -1.4%	+2.0%	-0.3%	-3.4%
Black women 	 -5.1%	+6.1%	-1.2%	-11.2%
White 	         +8.2%	-4.8%	+8.4%	+12.9%
Black 	         -4.1%	+5.4%	-1.8%	-9.6%
Hispanic/Latino  +2.6%	-3.2%	+0.7%	+5.8%

Clinton’s overall support, on the other hand, increased by 4.7%

·
Clinton’s support not only increased in all these categories between February and March-May, but she attracted considerable support from voters who had supported other candidates in February.

Obama did manage to increase his support among African Americans overall by 5.4%, among Black men by 6.7%, and among Black women by 9.6%. But when that data is weighed in terms of the overall electorate, the impact of those changes is small. Thus, Obama’s net increase in support among African American voters represents only 1.8% of the Democratic electorate, while Clinton’s net increase in support from White voters represents 8.4% of the Democratic primary electorate. Clinton’s overall net increase of 6.7% of the primary electorate is a clear sign that "buyers' remorse" has set in.

CHART 1B

ALL VOTERS, & MALE AND FEMALE VOTERS TOTAL SUPPORT, and WEIGHED SUPPORT

As a result of “buyers' remorse”, Barack Obama went from winning the overall vote by 2.7% (49.5% to 46.8%) in the February primaries to losing among all voters in the primaries held from March through May by 4.0% (47.2% to 51.2%).

CHART 2

TABLE 2 (Data for Chart 2)

SUPPORT LEVELS:  ALL VOTERS, MALE, FEMALE
 	         WEIGHED	  UNWEIGHED
 	      Clinton  Obama   Clinton 	Obama
ALL: February	 	 	46.8%	49.5%
ALL: Mar-May	 	 	51.2%	47.2%
MALE: February	17.6%	22.7%	41.4%	53.5%
MALE: Mar-May	20.2%	21.6%	47.4%	50.9%
FEMALE: Feb.	29.4%	26.1%	51.2%	45.6%
FEMALE: Mar-May	31.2%	25.4%	54.3%	44.3%

RACIAL/ETHNIC DEMOGRAPHICS: TOTAL SUPPORT, and WEIGHED SUPPORT

“Buyers' Remorse” is also reflected in the changes in support among racial/ethnic demographic groups.

CHART 3

But while Obama’s numbers among African Americans may look impressive, his improvement as a percentage of the electorate is dwarfed by the gains made by Clinton among White and Hispanic/Latino voters.

And it should be noted, when the numbers drawn from the demographics of the Democratic primaries are adjusted to reflect the overall electorate in the 2004 General Election, Clinton’s numbers are even more impressive (see below).

TABLE 3 (Data for Chart 3)

SUPPORT LEVELS—RACIAL/ETHNIC DEMOGRAPHICS
 	                   WEIGHED	  UNWEIGHED
 	              Clinton   Obama  Clinton 	Obama
WHITE: February	        34.2%	27.6%	52.6%	42.4%
WHITE: March to May	39.5%	24.5%	60.8%	37.6%
BLACK: February	         2.8%	15.5%	15.2%	83.0%
BLACK: March to May	 2.1%	16.5%	11.1%	88.4%
HISPANIC: February	 7.6%	4.2%	63.4%	35.2%
HISPANIC: March to May	 8.0%	3.9%	66.0%	32.0%

RACE v GENDER: TOTAL SUPPORT.and WEIGHED SUPPORT

Perhaps the group that has shown the most “buyers' remorse” is White Male Democrats, who not only rallied to Clinton after Obama was declared the “inevitable nominee”, but deserted Obama in droves. Not even the obvious misogyny of white male voters in Oregon (where Obama won the White Male vote by 2 to 1 (66% to 33%), while winning among White Women by a mere 2% (49% to 51%), could prevent Clinton from racking up major gains among White Male voters.

CHART 4

But in terms of actual impact on the results, Obama’s gains among Black men and women were negligible compared to Clinton’s gains among White males and females.

Obama’s gains among Black males represented just 0.1% (to 6.7% overall) of the Democratic primary electorate, while his gains among Black females represented a gain of 0.6% (to 9.6% overall) of Democratic voters.

Clinton’s gains among White male voters represented 4.9% of the electorate, as she went from negative 1.3% in February (HRC: 12.5%, BHO: 13.8%) to a lead of 3.6% of the electorate. (HRC: 15.6%, BHO: 11.9%)
Clinton’s lead among White female voters in the March-May primaries represented 11.4% of the Democratic electorate overall (HRC: 23.9%, BHO: 12.5%)

TABLE 4 (Data for Chart 4)

SUPPORT LEVELS--RACE AND GENDER
 	                   WEIGHED	  UNWEIGHED
 	              Clinton 	Obama  Clinton   Obama
WHITE MALES: February	12.5%	13.8%	44.6%	49.2%
WHITE MALES: Mar-May	15.6%	11.9%	55.4%	42.5%
WHITE FEMALES: February	21.6%	13.8%	58.6%	37.4%
WHITE FEMALES: Mar-May	23.9%	12.5%	64.6%	33.9%
BLACK MALES: February	0.9%	6.6%	11.6%	87.3%
BLACK MALES: Mar-May	0.8%	6.7%	10.2%	89.3%
BLACK FEMALES: February	1.9%	9.0%	16.9%	81.2%
BLACK FEMALES: Mar-May	1.3%	9.6%	11.7%	87.3%

ADJUSTING FOR THE GENERAL ELECTION DEMOGRAPHIC DISTRIBUTION

When weighing the data about to reflect the percentage of the electorate voting for each candidate in each demographic category, the sum of all the primaries in which Obama and Clinton went head-to head (i.e., those held from February to May) was used. But that demographic distribution is not similar to that of the General Election in 2004, and it would doubtless be helpful to see how Clinton and Obama would do if the data were adjusted to reflect a General Election distribution.

Unfortunately, there is inadequate exit polling data to account for all the racial/ethnic demographic groups. But the demographic differences among the three larges groups (which comprise 97.6% of the primary electorate, and 96% of the general electorate) are significant. While the primary electorate was 69.4% White, 18.5% African American, and 9.5% Hispanic/Latino, the electorate in the 2004 General Election was 77% White, 11% African American, and 8% Hispanic/Latino.

CHART 5

TABLE 5 (Data for Chart 5)

ADJUSTED MARCH-MAY RACIAL/ETHNIC DATA -- 
2004 GENERAL ELECTION DISTRIBUTION
	                      CLINTON	OBAMA
WHITE March - May Primaries	39.5%	24.5%
WHITE General Election adjusted	46.8%	29.0%
BLACK March - May Primaries	 2.1%	16.5%
BLACK General Election adjusted	 1.2%	 9.7%
HISPANIC March - May Primaries	 8.0%	 3.9%
HISPANIC Gen. Election adjusted	 5.3%	 2.6%
TOTAL March - May Primaries	49.6%	44.8%
TOTAL General Election adjusted	53.3%	41.2%

If the Democratic Primary electorate were the same as that of the 2004 General Election, Clinton’s lead over Obama for the primaries held from March through May would more than double, going from 4.8% (HRC: 49.6%, BHO: 44.8%) to 12.1%. This is mainly due to the under-representation of White voters, and the over-representation of African American voters, in the Democratic Primary electorate.

Given the level of “buyers' remorse” demonstrated by Democratic voters, this adjusted data should serve as a major warning to super-delegates who are considering which candidate to support for the Democratic nomination.

PART TWO of “Buyers Remorse” will examine the “class” Demographics: Income and Education.

NOTE 1 -- METHODOLOGY
All states in which exit polling is available and in which Hillary Clinton and Barack Obama were the only two “major” candidates are included in the data for this survey. (In essence, all primary states from Super Tuesday onward.)

Exit polling data was taken from the CBS News website at http://election.cbsnews.com/campaign2008...

Vote totals were taken from official state websites where available, and when those were not available, totals were based on publicly available news sources. A full list of states and the websites from which the data was gotten can be found at http://www.glcq.com/election08/remorse/r...

To determine overall percentages, each state's exit polling percentages were multiplied by the total number of votes in that state within each category, and then those votes were distributed to Clinton and Obama based on the percentages found in the exit polls. (Total votes included only ballots that were counted as having valid votes.) After each candidate's vote total was determined for each category in each state, each category, and the candidate’s vote in each category, was summed, and a percentage of the total vote for each candidate in each category was derived from those numbers.

Weighed averages were determined by multiplying the percentage for each candidate in a given category by the percentage that category comprised of the overall vote total. (i.e. if category X comprised 40% of the total vote, and Clinton received 60% of the vote in category X, Clinton’s Category support from Category X comprised 24% (60% time 40%) of the electorate.)

Buyers Remorse Part Two--Exploding Class-Related Myths: Income & Education

KEY FINDINGS
When comparing the February primaries to those held in March, April, and May

INTRODUCTION
Ever since Barack Obama was declared by the media and the Obama campaign to be the “inevitable” nominee after winning “ten straight contests” in mid-February, Democrat primary voters have been sending a message – “DO NOT WANT”. In the primaries held after Obama’s string of victories, Hillary Clinton has received 500,000 more votes than Obama (out of 12.7 million cast).

Obama supporters, and media pundits, have dismissively ascribed Clinton’s success during the past three months to racism and “low information voters.” But an analysis of exit polling data show that it is Obama who is gaining ground among voters with the lowest educational levels, while losing ground among the most educated. And Obama is losing ground to Clinton in every income demographic except the very lowest (under 15K/year).

In certain categories, both candidates saw increases in overall support. These dual increases are explained by voters who had preferred a candidate who is no longer in the race (primarily John Edwards) deciding to support to one of the remaining contestants—while the February primaries were essentially two-person contests, 3.5% of the electorate voted for someone other than Obama or Clinton. In the March through May primaries, this “someone else” vote declined to 1.6%.

Were it not for Obama’s increased support among the poorest and least educated voters, Clinton’s 500,000-vote margin over Obama in primaries held since March would be even greater. Neither “racism” nor “lack of knowledge” is driving voters away from Obama; rather it is the realization among rank and file Democrats that Hillary Clinton is far better prepared, more experienced, and more competent than Barack Obama.

(see NOTE 1 for methodology. Charts and accompanying data can be found at http://www.glcq.com/election08/remorse/r... . Full state-by-state exit poll data used in this article can be found at http://www.glcq.com/election08/remorse/r... . Part One of the Buyers Remorse series, which focuses on All voters, and the Race and Gender demographic categories can be found here.)

THE INCOME DEMOGRAPHICS

When compared with the exit polling data from February , Hillary Clinton has outperformed Barack Obama in every income demographic except for the lowest income Americans.

In the March through May primaries, both candidates saw percentage increases among voters in both the “$15K to 30K” and “30K to 50K” demographics, but Clinton did far better:

CHART 1A

In the middle income and upper income brackets, Obama lost support, while Clinton not only picked up the support Obama had lost, she also attracted a significant percentage of the kind of voters who had chosen neither Clinton nor Obama in February.

CHART 1B

When these demographic groups are weighed (see Note 1) to reflect the percentages of the overall Democratic primary electorate, Clinton’s advances among middle and upper income voters far outweigh her loss of support among those making “Under $15K” (Chart 1B),

TABLE 1 (Data for Charts 1A and 1B)

INCOME
 	           CHANGE IN SUPPORT	CLINTON GAIN/LOSS
 	               Clinton	Obama  Weighed  Unweighed
Under $15,000 	        -3.6%	+5.4%	-0.7%	-9.0%
$15,000 - $29,999 	+1.4%	+0.4%	+0.1%	+1.1%
$30,000 - $49,999 	+2.2%	+0.3%	+0.4%	+1.9%
$50,000 - $74,999 	+7.3%	-2.6%	+2.1%	+9.9%
$75,000 - $99,999 	+4.9%	-2.8%	+1.1%	+7.7%
$100,000 or more 	+4.3%	-2.5%	+1.7%	+6.8%

In other words, Clinton’s gains among Democratic voters during the past three months were due primarily to diminished support for Obama and greater support for Clinton among middle and upper income voters – and the media promulgated claim that ‘racist White working class’ voters account for Clinton’s success is a vicious myth that is contradicted by the actual results from the primaries.

EDUCATION:

Perhaps one of the most pernicious myths being spread by the Obama campaign and the media is that Clinton’s success since the beginning of March is due to “low information voters”. Yet, when one looks at educational levels of the electorate, it is Obama who has improved his performance among the least educated voters, while Clinton has done far better among those with the highest levels of educational achievement.

CHART 2A

As with income levels, Obama’s gains among those with the least education are far less significant than they appear, because those without a High School degree comprise only 4.9% of the Democratic primary electorate, and Obama’s +13.3% net gain in that demographic translates to a gain of only 0.6% of the electorate itself.

CHART 2B

TABLE 2 (Data for Charts 2A and 2B)

EDUCATION
 	           CHANGE IN SUPPORT	CLINTON GAIN/LOSS
 	               Clinton	Obama  Weighed  Unweighed
Not HS graduate	        -5.9%	+7.4%	-0.6%	-13.3%
High school grad	+4.4%	-0.6%	+1.0%	+5.0%
Some college	        +5.0%	-2.3%	+2.1%	+7.4%
College graduate	+2.3%	+0.5%	+0.5%	+1.9%
Postgrad study	        +3.5%	-1.9%	+1.2%	+5.4%

CLASS ISSUES (INCOME AND EDUCATION)
What happens when these two measures, income and education, are combined?

The media and Obama’s supporters have made much of Obama’s supposed far greater appeal to the “upper middle” and so-call “creative class” elites, while simultaneously attributing Clinton’s success in the March, April and May primaries to racism among the “White working class”. But exit polling data suggests that neither claim has any basis in fact. Although the exit polls themselves do not specifically address the issue of “class”, by looking at trends in income and education data, an approximation of how various socio-economic classes are voting can be determined.

THE ‘UNDERCLASS” – LOW INCOME AND LOW EDUCATION VOTERS FLOCKING TO OBAMA

While the media and Obama’s own supporters have tried to maintain a focus on Obama’s appeal to “upper” class voter, is only among the “underclass” – those voters found on the lowest rungs of the socio-economic ladder – that Obama showed any net increase improvement relative to Clinton.

CHART 3A

But because low income and low education level voters make up a relatively small part of the electorate, Obama’s gains in both categories had minimal impact on the overall vote totals.

CHART 3B

TABLE 3 (Data for Charts 3A and 3B)

THE 'UNDERCLASS' (INCOME AND EDUCATION)
                                SUPPORT LEVELS 
 	                Unweighed Data	Weighed Data
 	               Clinton 	Obama	Clinton  Obama
Under $15K  February	51.3%	45.6%	3.7%    3.3%
Under $15K  Mar-May	47.7%	51.0%	3.5%    3.7%
Not HS grad February	63.6%	33.2%	3.1%    1.6%
Not HS grad Mar-May	57.7%	40.6%	2.8%    2.0%

WORKING CLASS AND MIDDLE CLASS VOTERS –THE MYTH OF “CLINTON’S RACIST SUPPORT” DEBUNKED

As noted above, both Clinton and Obama received additional support from working class/lower-middle class voters (those making more than $15K but less than the $50K. But while Clinton expanded her lead among these income groups, her biggest gains came among the solidly ‘middle income” voters; (those making $50K to $75K. Clinton also improved on her lead against to Obama among those with just a high school degree, but showed even more significant gains (and took the lead from Obama) among those with some College education and/or an Associates Degree.

CHART 4A

Clinton expanded her lead in the $15K to 30K demographic by 1.1%, but because this group makes up only 12.7% of the voters, her gains among these voters represented only 0.1% of the electorate.
Clinton also expanded her lead among those making $30K to $50K, who make up 20.1% of the electorate. Her 1.9% increase translates into a gain of 0.4% of voters
Clinton’s 5% gain among High School graduates lead to her expanding her lead by 1.0% of voters (HS grads made up 20.7% of the electorate.)

In the February primaries, Obama received 2.7% more votes than Clinton, while Clinton received 4.0% more support than Obama in the March, April and May primaries. In other words, the 0.5% combined additional support of the income categories that make up the “White working class” voters whose supposed racism explained Clinton’s recent success made up only 7.6% of her overall gains. (See Note 2) And the net 1% more of voter support from “low information” High School graduates made up less than 1/6 (15.4%) of her increased support over Obama.

CHART 4B

Far more substantial gains were achieved by Clinton among the “middle-middle” class, which comprised 21.2% of voters, and those with some educational experience beyond High School, who were 28.8% of voters.

TABLE 4 (Data for Charts 4A and 4B)

MIDDLE/WORKING CLASS (INCOME & EDUCATION)
                                SUPPORT LEVELS 
 	              Unweighed Data	Weighed Data
 	              Clinton	Obama	Clinton	Obama
$15K - $30K  February	51.2%	45.0%	6.5%	5.7%
$15K - $30K  Mar-May	52.6%	45.4%	6.7%	5.8%
High Sch. grad  Feb.	54.3%	40.4%	11.2%	8.4%
High Sch. grad  Mar-May	58.6%	39.8%	12.1%	8.2%
$30K - $50K  February	48.9%	47.2%	9.8%	9.5%
$30K - $50K  Mar-May	51.1%	47.5%	10.2%	9.5%
Some college February	46.9%	49.0%	13.5%	14.1%
Some college Mar-May	51.9%	46.6%	15.0%	13.5%
$50K - $75K  February	45.1%	49.3%	9.6%	10.5%
$50K - $75K  Mar-May	52.4%	46.6%	11.1%	9.9%

“UPPER MIDDLE” AND “UPPER” CLASS – CLINTON PICKS UP CONSIDERABLE SUPPORT AMONG THE “SMART SET”

While Obama’s claim to superior appeal among the upper income/educational level demographics had some validity in February, the results of the March through May primaries tell a different story. While Obama maintained his lead in most of the demographic categories in this segment of the population, that advantage has been substantially reduced as these voters have taken a closer look at both Clinton and Obama.

CHART 5A

Clinton’s gains among these demographic categories represented substantial portions of the overall primary electorate

In other words, despite being declared the “inevitable nominee” in mid-February, not only did Barack Obama fail to increase his support in the “class” based categories his campaign brags about, he lost considerable support in those demographics.

CHART 5B

TABLE 5 (Data for Charts 5A and 5B)

UPPER MIDDLE/UPPER CLASS (INCOME AND EDUCATION)
                                SUPPORT LEVELS 
 	              Unweighed Data	Weighed Data
 	               Clinton 	Obama	Clinton  Obama
College grad February	42.8%	53.1%	10.4%	12.9%
College grad Mar-May	45.1%	53.5%	11.0%	13.0%
$75K - $100K  February	45.8%	50.5%	6.6%	7.2%
$75K - $100K  Mar-May	50.7%	47.7%	7.3%	6.8%
Postgrad study February	42.8%	54.4%	9.2%	11.6%
Postgrad study Mar-May	46.3%	52.5%	9.9%	11.2%
$100K or more February	43.8%	53.6%	10.6%	13.0%
$100K or more Mar-May	48.1%	51.1%	11.7%	12.4%

NOTE 1 -- METHODOLOGY
All states in which exit polling is available and in which Hillary Clinton and Barack Obama were the only two “major” candidates are included in the data for this survey. (In essence, all primary states from Super Tuesday onward.)

Exit polling data was taken from the CBS News website at http://election.cbsnews.com/campaign2008...

To determine overall percentages, each states exit polling percentages were multiplied by the total number of votes in that state within each category, then those votes were distributed to Clinton and Obama based on the percentages found in the exit polls. (Total votes included only ballots that were counted as having valid votes.) After each candidates vote total was determined for each category in each state, each category, and the candidate’s vote in each category, was summed, and a percentage of the total vote for each candidate in each category was derived from those numbers.

NOTE 2: ROUNDING “ERRORS”
In certain cases, rounding “errors” are responsible for differences in reported changes in support described in the text in certain demographic categories from the data found in the tables and charts. For instance, the percentages in February for the $50K to $75K cohort round to 45.1% for Clinton, and 49.3% for Obama, which looks like a 4.2% lead for Obama. But as noted in the text, that lead is only 4.1%, because the actual Clinton percentage was 45.133% and Obama’s was 49.261%, and the difference between the two is 4.128%.