The effects of candidate age in the 2008 presidential election.
Kenski, Kate ; Jamieson, Kathleen Hall
When 71-year-old Senator John McCain had nearly secured enough
delegates to lock up the Republican nomination, 46-year-old Senator
Barack Obama, who was in a battle for the Democratic nomination against
then-Senator Hillary Clinton, told a crowd in Virginia, "This
week--this week, we found out that the presumptive nominee of the
Republican Party will be Senator John McCain. Now, I believe John McCain
is a good man and he is a genuine American hero. And we honor his half
century of service to this nation" (CNN 2008). The compliment
masked an attack. Cloaked as praise was a reminder of his longevity.
McCain's age had been a target by other opponents as well.
"[I]f John takes over the presidency at 72 and he ages three to
one, how old will he be in four years? He'll be eighty-four,"
observed Mike Huckabee supporter and actor Chuck Norris in January 2008
(Fox News 2008).
In this study, we investigate the extent to which the public
believed that McCain was "too old to be president" as well as
the belief that Obama was "too young" using data from the 2008
National Annenberg Election Survey (NAES). We also examine the influence
of these beliefs on candidate evaluations and vote preference. Our
results indicate that more people believed that McCain was "too
old" to be president than believed that Obama was "too
young." Trends from March 2008 to election day show that the
percentage of people concerned about McCain's age increased over
time, while the percentage worried about Obama's youth decreased.
These beliefs were significantly associated with party identification,
such that major party identifiers were more likely to believe that a
candidate's age was problematic if the candidate belonged to the
other party. Newspaper exposure and reading information about the
presidential campaign on the Internet were significantly associated with
believing that age disqualified McCain from the presidency. Candidate
age perceptions were significantly associated with candidate
favorability ratings and vote preference, even when party identification
and a host of other sociodemographic characteristics were controlled.
McCain's Age Questioned
Age stereotyping is a pervasive part of contemporary culture. The
literature is replete with studies demonstrating that age discrimination
and other manifestations of ageism are common in many domains (e.g.,
Hedge and Borman 2006; Myers 2007; Nelson 2005). Riding on the back of
ageism are assumptions that aging is a process of progressive decline
and that being a septuagenarian signals senility. In 2007, Gallup hinted
at the ways in which age might play out in 2008. When asked to assess
various characteristics of a presidential candidate, roughly one in two
reported that being "70 years of age or older" was undesirable
(Carroll 2007).
McCain, of course, was not the first presidential candidate to have
his age translated into an indictment. Republican nominee Bob Dole, the
only nominee to run at an age more advanced than McCain's, faced
such questions and, with them, insinuations of incompetence in 1996 in
his contest against Democratic incumbent Bill Clinton. Abrams and Brody
(1998) observe that news articles about Dole frequently mentioned his
age.
Also problematic for McCain was the fact that the media focused
more on his age than on Obama's. When Obama's experience was
challenged during the presidential primaries by his Democratic rivals,
especially Senator Hillary Clinton, the indictments did not include
slurs about his youth. As he was 46 years old during the Democratic
primaries and 47 years old during the general election, he was not only
more than a decade older than the constitutional minimum requirement to
assume the presidency, but also older than a number of previous holders
of the nation's highest office. By contrast, McCain turned 72 just
before the beginning of the general election campaign, meaning that, if
elected, he would have been the oldest person ever to assume the
presidency in the history of the country.
When it came to age, news coverage was not McCain's friend. A
Lexis-Nexis search of print, broadcast, and cable transcripts from June
i to November 4, 2008 (McCain, 71; McCain, 72; Obama, 46; Obama 47)
conducted under our direction by Annenberg researcher Jacqueline Dunn
reveals that McCain's name was tied to his age 1,390 times and
Obama's to his 898 times. We therefore hypothesized that
[H.sub.1]: Voters were more likely to report that John McCain was
"too old to be president" than to report that Barack Obama was
"too young to be president."
The hypothesis is grounded in news reports that regularly reminded
their audiences, as a February 2008 Newsweek article did, that
"John McCain, 71, will be the oldest president ever elected if he
goes on to win his party's nomination and the White House in
November" (Thomas 2008). To the notion that McCain would be the
"oldest person ever to ascend to the presidency," Time added
the warrant that it used to justify concern about McCain's age.
"He has suffered serious skin cancers over the years, not to
mention brutal physical torture as a prisoner of war." From that
nexus among age, illness, and biography, the article concluded,
"His age and health, therefore, are of legitimate concern to
voters" (Carney 2008, 37).
Some columnists explicitly fielded the notion that aging signaled
"mental decline." "How Old Is Too Old?" read the
headline on Anna Quindlen's February 4, 2008, column in Newsweek.
The piece argued, "It's significant that while the old
mandatory retirement age of 65 has been largely junked, there are still
age limits for jobs like airline pilot or police officer, the kinds of
jobs that require some of the same skills as the presidency--unwavering
mental acuity and physical energy."
Reporters speculated as well on the tactical disadvantage that
McCain's advanced years carried. After reviewing the Republican and
Democratic ads of the 2008 primary season, for example, The
Atlantic's Jim Fallows (2008) forecast McCain's general
election debate performance by saying, "Worse, he will look and
sound old and weak next to Obama. Ronald Reagan was about McCain's
current age when he ran for re-election against Walter Mondale, but
Reagan looked 10 years younger than McCain does now .... [L]acking
Reagan's outward haleness, he risks coming across like Dole against
Clinton-or, more ominously, his fellow ex-POW [Ross Perot's running
mate] James Bond Stockdale, who turned in a notoriously lost-and
incoherent-sounding performance against [Bill Clinton's running
mate] AI Gore and [George H. W. Bush's vice president] Dan Quayle
in the 1992 vice-presidential debates."
Backdrop assumptions are also at play in our perceptions of youth.
When considering a presidential aspirant, does 47 signal "vigorous,
energetic, brimming with fresh ideas" or "'wet behind the
ears' upstart" and "unready to lead?" Of course,
before one is inclined to ask such questions, age must be made salient
enough to be part of the discussion. For McCain, the media ensured that
conversation happened. As we will document, the Obama campaign did the
same.
Age Theme Present in Obama Ads
Reinforcing an assumption to which the audience is disposed, of
course, is easier than forming a new attitude. And an inference is more
readily prompted when other forces in the culture, such as news reports,
opinion columns, and comedy, are making the same point. Throughout the
post-primary period, the Obama ads built on these cultural factors to
attach negative notions of "old" to the Republican nominee.
The age-related themes in Obama's ads included the following:
"John McCain is blaming Barack Obama for gas prices? The same
old politics." (1)
"John McCain: He's been in Washington for 26 years."
(2)
"John McCain: Same old politics, same failed policies."
(3)
"Lurching to the right, then to the left. The old Washington
dance. Whatever it takes. A
Washington celebrity playing the same old Washington games."
(4)
"For 26 years in Washington, John McCain played the same old
games." (5)
While the McCain campaign argued throughout the general election,
as the Clinton campaign had before it, that Obama was not ready to be
president, those arguments centered not on his age but on his relatively
short time in elective office, his liberal credentials, and his past
associations with suspect groups such as ACORN, a community organizing
group alleged to have falsely registered unqualified individuals to
vote, and individuals such as former Weather Underground leader William
Ayers and convicted Chicago wheeler-dealer Tony Rezco.
In light of the focus on McCain's age in both news and
Democratic rhetoric, McCain's age should have been more salient to
the public than Obama's. We therefore hypothesized,
[H.sub.2a]: Perceptions of McCain being "too old to be
president" increased across the campaign period.
[H.sub.2b]: Perceptions of Obama being "too young to be
president" did not increase across the campaign period.
Beliefs and attitudes, of course, are not just influenced by news
media and candidate advertising. Party identification has long been
shown to be an important factor in candidate impressions and vote
preference (Miller and Shanks 1996). In terms of perceptions about age,
a study by Abrams and Brody (1998) found that party identification was a
primary determinant of negative age evaluations about Republican nominee
Bob Dole in 1996. We suspected that party identification likewise
influenced age perceptions about McCain and Obama in 2008. We
hypothesized,
[H.sub.3]: Major party identifiers were more likely to believe that
the other major party's candidate's age was a liability to
being president. Specifically, Republicans were more likely to believe
that Obama was "too young to be president," and Democrats were
more likely to believe that McCain was "too old to be
president."
Given the news media's attention to McCain's age noted
earlier, we hypothesized that exposure to news media increased
people's belief that McCain was too old.
[H.sub.4]: News consumption was positively associated with
perceptions of McCain being "too old to be president," even
when controlling for demographic and political ideological factors.
We assumed that candidate age perceptions were likely to influence
the candidates' favorability ratings and subsequently vote
preference.
[H.sub.5a]: Believing that McCain was "too old to be
president" was negatively associated with McCain's
favorability ratings.
[H.sub.5b]: Believing that Obama was "too young to be
president" was negatively associated with Obama's favorability
ratings.
[H.sub.6a]: Believing that McCain was "too old to be
president" was associated with vote preference. Specifically, those
who felt that McCain was too old were less likely to vote for him.
[H.sub.6b]: Believing that Obama was "too young to be
president" was associated with vote preference. Specifically, those
who felt that Obama was too young were less likely to vote for him.
Methods
Throughout the 2008 election season, the Annenberg Public Policy
Center at the University of Pennsylvania conducted the National
Annenberg Election Survey, a 13-month-long rolling cross-sectional
survey on the political attitudes and behaviors of the U.S. population.
The telephone surveys began December 17, 2007, and ended November 3,
2008. Telephone interviews were conducted with adults in the United
States each day. Households throughout the nation were selected using
random-digit dialing. Random individuals within each household were
selected for interviewing. The current project utilizes data collected
between March 4 and November 3 (n = 40,704) and between October 15 and
November 3 (n = 4,945). (For a description of the basic contours of the
survey's research design, see Romer et al. 2006).
Variables
Demographics, Respondents were asked to indicate several
demographic characteristics. Demographic variables used in this study
include gender (1 = female, 0 = male), age (years), race (1 = black, 0 =
other), ethnicity (1 = Hispanic, 0 = other), education (years of school
completed), and household income (in thousands).
Political leanings and orientations. Respondents were asked to
indicate their party identification (Republican, Democrat, independent,
or other). Party identification was recoded into two dummy variables,
one indicating Democratic Party affiliation and one indicating
Republican Party affiliation. Political ideology was assessed using a
five-point scale ranging from "very liberal" (1) to "very
conservative" (5).
Information sources. Respondents were asked to indicate how many
days in the past week (0 to 7 days) they had seen information about the
2008 presidential campaign on broadcast or cable television; heard
information about the 2008 presidential campaign on radio shows that
invite listeners to call in to discuss current events, public issues, or
politics; read a newspaper for information about the 2008 presidential
campaign; and seen or heard information about the 2008 presidential
campaign on the Internet.
Candidate favorability ratings. Respondents were asked to evaluate
presidential candidates Barack Obama and John McCain. "Now for each
of the following people, please tell me if your opinion is favorable or
unfavorable using a scale from 0 to 10. Zero means very unfavorable, and
10 means very favorable. Five means you do not feel favorable or
unfavorable toward that person. Of course you can use any number between
zero and 10."
Vote preference. Since its inception in 2000, one of the unique
features of the NAES has been measurement of absentee and early voting.
About 14% of voters cast their ballots before election day in 2000, and
this rose to 20% in 2004 (Annenberg 2005). NAES data show that over 30%
of ballots were cast before election day in 2008 (Annenberg 2008).
Consequently, when assessing vote preference across a campaign, it is
important to combine vote intentions for those respondents who have not
cast their ballots yet at the time of the interview and vote behaviors
for those respondents who have already voted.
Survey participants who had not yet voted were asked,
"Thinking about the general election for president in November,
2008, if that election were held today, and the candidates were John
McCain and Sarah Palin, the Republicans, Barack Obama and Joe Biden, the
Democrats, Ralph Nader and Matt Gonzalez, the independent candidates,
and Bob Barr and Wayne Allyn Root, the Libertarians, and Cynthia
McKinney and Rosa Clemente, the Green Party candidates, for whom would
you vote?" A similar question was asked of those who reported that
they had already cast their ballots, but their question began with,
"In the 2008 presidential election who did you vote for?"
followed by the same response options provided on the vote intention
question. The names of the tickets were randomly rotated. After
combining these question responses into a single measure, the responses
were recoded into a two-party vote preference measure (1 = Obama, 0 =
McCain).
Perceptions of candidates' age. Respondents were asked,
"Do you feel John McCain is too old to be president, or not?"
and "Do you feel Barack Obama is too young to be president, or
not?" For multivariate analyses, "too old" and "too
young" responses were recoded as 1 and "not too old,"
"not too young," and "don't know" responses
were coded as 0.
Analytical Procedures
Data were weighted to take account of the household size and number
of telephone lines in the residence and to adjust for variation in the
sample relating to geographic region, sex, race, age, and education for
the univariate and bivariate analyses. (6) A proportion test was used to
test Hypothesis 1. Trend analyses were conducted to examine the second
set of hypotheses. To understand the relationship between age
perceptions and party identification, cross-tabulations were analyzed
and chi-square tests were conducted. To further examine the party
identification hypothesis (H3) and the news media hypothesis (H4),
logistic regressions predicting age perceptions were conducted. For our
hypotheses regarding the candidate favorability ratings (H5a and H5b),
independent samples t tests were used to evaluate candidate favorability
rating differences between those who held negative age perceptions about
the candidates and those who did not. We also conducted ordinary least
squares regressions predicting the candidate favorability ratings. For
our final set of hypotheses on vote preference (H6a and H6b), we
examined cross-tabulations between age perceptions and vote preference
and conducted logistic regression predicting vote preference.
Results
Respondents interviewed between October 15 and November 3 were
asked about their perceptions of McCain's and Obama's ages.
When asked whether they felt McCain was "too old to be
president," 34.9% said McCain was "too old," 63.5% said
he was "not too old," 1.1% said they didn't know, and
0.5% refused to give a response to the question. When asked whether they
felt Obama was "too young to be president," 11.8% said Obama
was "too young," 87.2% said he was "not too young,"
0.8% said they didn't know, and 0.3% refused to give a response to
the question. A one-sample test of proportion revealed that the
differences in the "too old" and "too young"
percentages were statistically significant (z = 50.352, p < .001),
providing support for Hypothesis 1.
Figure 1 presents the percentages of the public believing that
McCain was "too old to be president" and Obama was "too
young to be president" across the presidential campaign. Curve
estimation analyses confirmed that there was a significant linear
increase in the percentage of the public that believed that McCain was
"too old to be president" over the 245 days examined
([R.sup.2] = .190, p < .001). Each day that passed was associated
with a 0.04 % increase in the "too old" belief. Quadratic
([R.sup.2] = .190,p < .001) and cubic ([R.sup.2] =. 195, p < .001)
trends also were found to explain the changes in perceptions about
McCain's age. The increase in the public's perception that
McCain was "too old to be president" provides support for
Hypothesis 2a.
Hypothesis 2b maintained that there was no increase in the
percentages of the public believing that that Obama was "too young
to be president" across the campaign. This hypothesis was supported
by the data, but this does not mean that there was no change across time
in that perception. A significant linear trend ([R.sup.2] = .075,p <
.001) was detected; it was not as pronounced as the linear trend found
for the belief about McCain's age. Each one-day increase over time
was associated with a 0.02 % decrease in the belief that Obama was
"too young to be president." Significant quadratic ([R.sup.2]
= .201, p < .001) and cubic ([R.sup.2] = .202, p < .001) trends
were detected and did a better job of explaining the patterns found the
data than the linear trend. From March to the third week of July, the
belief that Obama was "too young" appeared to increase. As
shown in Figure 1, the belief then decreased during the general election
campaign period.
[FIGURE 1 OMITTED]
As shown in Table 1, Democrats were more likely than members of any
other party identification to believe that McCain was too old to be
president. Of Democrats, 54.9 % said that McCain was too old, in
comparison to 31.3 % of independents and 11.4 % of Republicans.
Republicans were more likely than members of any other party
identification to believe that Obama was too young to be president. Of
Republicans, 17.4 % said that Obama was too young, in comparison to 12.5
% of independents and 6.1% of Democrats. Hypothesis 3, therefore, was
supported in the NAES data. It is important to note that while
Hypothesis 3 was supported, Republicans did not hold Obama's age
against his seeking of the presidency to the same extent that Democrats
held McCain's age against him.
As shown in Table 2, even when demographic characteristics and
political ideology were controlled, Democrats were significantly more
likely to believe that McCain was too old to be president and
Republicans were less likely to believe it. The tendency for Republicans
to contend that Obama was too young to be president, however, was not
significant when demographic characteristics and political ideology were
controlled. Democrats were significantly less likely to believe that
Obama was too young.
Given the amount of coverage in the newspapers about McCain's
age, we predicted that newspaper consumption would be positively
associated with the belief that McCain was too old to be president.
Mixed support for Hypothesis 4 can be seen in Table 2. Each one-day
increase in weekly newspaper reading was associated with a 1.028
increase in the odds of believing that McCain was too old to be
president. Each one-day increase in weekly newspaper reading was also
associated with a 3.7% decrease in the odds of believing that Obama was
too young to be president. Seeing campaign information online was also
positively and significantly associated with believing McCain was too
old to be president. Seeing campaign information on television news was
not significantly associated with candidate age perceptions in the face
of sociodemographic and political controls.
Independent sample t tests provided support for both Hypotheses 5a
and 5b. As shown in Table 3, those who believed that McCain was too old
to be president gave him significantly lower favorability ratings than
those who believed that McCain was not too old or did not know (3.47
versus 6.17;p < .001). Those who believed that Obama was too young to
be president gave him significantly lower favorability ratings than
those who believed that Obama was not too young or did not know (3.51
versus 6.27; p < .001).
Hypotheses 5a and 5b obtained additional support when multivariate
analyses were conducted. As shown in Table 4, even after controlling for
demographic and political characteristics of the respondents,
perceptions about the candidates' ages were significantly
associated with candidate
favorability ratings. Believing that McCain was too old to be
president was associated with a 1.783-point decrease in his favorability
ratings after other variables were taken into account (p < .001).
Believing that Obama was too young to be president was associated with a
1.043-point increase in McCain's favorability ratings after other
variables were controlled (p < .001). A similar pattern appeared when
predicting Obama's favorability ratings. Believing that Obama was
too young to be president was associated with a 1.709-point decrease in
his favorability ratings. Believing that McCain was too old to be
president was associated with a 1.519-point increase in Obama's
favorability ratings.
Hypotheses 6a and 6b maintained that candidate age perceptions were
associated with vote preference. We assumed that respondents who felt
that McCain was too old to be president were less likely to vote for
him; respondents who felt that Obama was too young to be president were
less likely to vote for him. As shown in Table 5, both Hypotheses 6a and
6b were supported in our data. While 50.7% of those who said that McCain
was not too old or did not know about his age being a problem reported
that they voted for him, only 9.4% of those who said that he was too old
ended up voting for him. Of those who said that Obama was too young to
be president, 16.6% voted for him anyway, while 52.6% of those said that
Obama was not too young or did not know about his age being a problem
voted for him.
Further support was found in logistic regression analyses
predicting the two-party vote (see Table 6). Even when demographics,
political characteristics, and candidate favorability ratings were taken
into account, candidate age perceptions were significantly associated
with vote preference (p < .001 for both items). Believing that McCain
was too old to be president was associated with a 4.544 increase in the
odds of voting for Obama, even when controlling for many other factors.
Believing that Obama was too young to be president was associated with a
71.6% decrease in the odds of voting for him.
To demonstrate the impact of believing that McCain was too old on
the probability of voting for Obama, Figure 2 displays the increased
percentage of an Obama vote from holding the belief that McCain was too
old controlling for demographic variables, party identification,
ideology, and candidate favorability ratings by baseline percentages of
intentions of voting for Obama if the "election were held
today." The logistic coefficient of believing that McCain was too
old was 1.514 (see Table 6) and resulted in the impacts shown in Figure
2. If a person had a 25% baseline expectation of voting for Obama, that
expectation changed to a 60.2% expectation of voting for Obama taking
those other variables in account if the person believed that McCain was
too old to be president. In other words, that belief resulted in a 35.2%
increase of having an Obama vote intention. If a person had a 50%
baseline expectation of voting for Obama, that expectation changed to
82% if the person believe that McCain was too old. In other words, that
belief resulted in a 32% increase of having an Obama vote intention. As
shown in Figure 2, if a person's vote intention was already heavily
leaning toward an Obama vote, the impact of believing that McCain was
too old was not as large. The bottom line, however, was that nothing
good could happen for McCain if people believed he was too old to be
president.
Conclusion
In the last three weeks of the general election, the disadvantage
that McCain faced on age perceptions was evident, as approximately
one-third of adults believed that he was "too old" to be
president. Less than 12% of adults felt that Obama was "too
young." The percentage of adults believing that McCain was
"too old" had grown since the primary season. In contrast,
during the same period, Obama had benefited from a decline in belief
that he was "too young." Although negative age perceptions
were strongly influenced by party identification, they exerted
independent, negative, and significant effects on candidate evaluations
and, ultimately, on vote preference. Controlling for a robust set of
sociodemographic, political characteristics, and candidate favorability
ratings, our results indicate that those believing McCain was "too
old" were more than 4.5 times more likely to vote for Obama than
McCain. In other words, subscribing to the belief that McCain was
"too old" not only affected candidate evaluations, it exerted
powerful effects on vote preference in addition.
[FIGURE 2 OMITTED]
At the outset of the article, we surmised that the news media
contributed to voters believing that the Republican aspirant was too old
to be president. Results from the NAES demonstrated that newspaper
exposure and Internet exposure to campaign information were positively
associated with that belief. Newspaper exposure was negatively
associated with the Obama "too young" belief. We suggested
earlier that the Obama campaign facilitated the McCain "too
old" impression through its advertising. We have not tested that
assumption here but consider our overall results consistent with that
narrative.
Candidate age perceptions matter in presidential campaigns,
especially when a candidate is considered "too old." While
this study is limited to the 2008 presidential campaign, we anticipate
that age will matter in future campaigns if opposition campaigns prime
the issue, the news media mention candidate age frequently in their
coverage, and one of the candidates running for the Oval Office is a
senior citizen. Until societal attitudes make fundamental, positive
shifts when it comes to aging, old age will probably continue to be
considered a liability.
References
Abrams, Herbert L., and Richard Brody. 1998. "Bob Dole's
Age and Health in the 1996 Election: Did the Media Let Us Down?"
Political Science Quarterly 113 (Autumn): 471-91.
Annenberg Public Policy Center. 2005. "Early Voting Reaches
Record Levels in 2004, National Annenberg Election Survey Shows."
News release, March 24. http://
annenbergpublicpolicycenter.com/NewsDetails.aspx?myId=67 (accessed April
30, 2010).
--. 2008. "Over One in Nine Citizens Have Already Voted,
Annenberg Survey Shows." News release, October 29.
http://annenbergpublicpolicycenter.com/NewsDetails.aspx?myId=304
(accessed April 30, 2010).
Carney, James. 2008. "The Phoenix." Time, February 4, p.
37.
Carroll, Joseph. 2007. "Which Characteristics Are Most
Desirable in the Next President?" Gallup, September 17. See:
http://www.gallup.com/poll/28693/
which-characteristics-most-desirablenext-president.aspx (accessed April
30, 2010).
CNN. 2008. "Election Results." February 9.
Fallows, James. 2008. "Rhetorical Questions." The
Atlantic, September, p. 50.
Fox News. 2008. "Special Report with Brit Hume." January
21.
Hedge, Jerry W., and Walter C. Borman. 2006. "Age Stereotyping
and Age Discrimination." In The Aging Workforce: Realities, Myths,
and Implications For Organizations, eds. Jerry W. Hedge, Walter C.
Borman, and Steven E. Lammlein. Washington, DC: American Psychological
Association, 27-48.
Miller, Warren E., and J. Merrill Shanks. 1996. The New American
Voter. Cambridge, MA: Harvard University Press.
Myers, Jane E. 2007. "Combating Ageism: Advocacy for Older
Persons." In Counseling for Social Justice, ed. Courtland C. Lee.
Alexandria, VA: American Counseling Association, 51-71.
Nelson, Todd D. 2005. "Ageism: Prejudice against Our Feared
Future Self." Journal of Social Issues 61 (June): 207-21.
Quindlen, Anna. 2008. "The Last Word." Newsweek, February
4, p. 64.
Romer, Daniel, Kate Kenski, Kenneth Winneg, Christopher Adasiewicz,
and Kathleen Hall Jamieson. 2006. Capturing Campaign Dynamics 2000 and
2004: The National Annenberg Election Survey. Philadelphia: University
of Pennsylvania Press.
Thomas, Evan. 2008. "What These Eyes Have Seen."
Newsweek, February 11, p. 27.
KATE KENSKI
University of Arizona
KATHLEEN HALL JAMIESON
University of Pennsylvania
(1.) Obama for America, "Old Politics General," July 28,
2008.
(2.) Obama for America, "New Energy Rev," July 30, 2008.
(3.) Obama for America, "Low Road Rev," August 4, 2008.
(4.) Obama for America, "Embrace," August 13, 2008.
(5.) Obama for America, "Never," August 20, 2008.
(6.) The literature is unclear as to whether it is more
conservative or less conservative to weight data when conducting
multivariate analyses that employ a host of demographic controls also
used in the weighting process. Therefore, the unweighted data were used
for the regression models reported in this study.
Kate Kenski is an assistant professor in the Department of
Communication at the University of Arizona. She is the coauthor of
Capturing Campaign Dynamics: The National Annenberg Election Survey and
The Obama Victory: How Media, Money, and Message Shaped the 2008
Election.
Kathleen Hall Jamieson is the Elizabeth Ware Packard Professor of
Communication in the Annenberg School for Communication and the Walter
and Leonore Annenberg Director of the Annenberg Public Policy Center at
the University of Pennsylvania. Among her many books are Presidents
Creating the Presidency and The Obama Victory: How Media, Money, and
Messages Shaped the 2008 Election.
TABLE 1
Candidate Age Perceptions by Party Identification
Party Identification
Republican Democrat Independent Other
Do you feel John McCain is too old to be president, or not?
Yes, McCain is too old 11.4% 54.9% 31.3% 29.3%
No, McCain is not too old 87.5% 43.2% 67.2% 70.7%
Don't know 1.0% 1.4% 0.9% 0.0%
Refused 0.2% 0.5% 0.6% 0.0%
Chi-square = 648.00, df = 9, p < .001
Do you feel Barack Obama is too young to be president, or not?
Yes, Obama is too young 17.4% 6.1% 12.5% 15.8%
No, Obama is not too young 81.7% 93.1% 86.5% 81.2%
Don't know 0.7% 0.6% 0.7% 2.3%
Refused 0.2% 0.2% 0.4% 0.8%
Chi-square = 106.05, df = 9, p < .001
Source: NAES 2008 telephone survey conducted between October 15 and
November 3; data weighted; column percentages given.
TABLE 2
Logistic Regression Predicting Perceptions of Candidate Age
McCain Too Old (1) vs.
Not Too Old/DK (0)
Model 1
B SE Exp(B)
Intercept .478 .308 1.614
Demographics
Female (1 = yes, 0 = no) .315 .075 1.371 ***
Age (in years) .000 .002 1.000
Black (1 = yes, 0 = no) .429 .131 1.536 **
Hispanic (1 =yes, 0 = no) .353 .140 1.424 *
Education (in years) -.050 .017 .951 **
Household income (in thousands) .000 .001 1.000
Political Characteristics
Republican (1 = yes, 0 = no) -1.325 .119 .266 ***
Democrat (1 = yes, 0 = no) .702 .082 2.018 ***
Ideology (1 = very liberal to -.238 .036 .788 ***
5 = very conservative)
Information Sources
Saw presidential campaign
information on TV news
(0-7 days)
Heard about presidential
campaign on talk radio
(0-7 days)
Saw presidential campaign
information in newspapers
(0-7 days)
Saw or read presidential
campaign information on
internet (0-7 days)
N 4054
Cox & Snell [R.sup.2] .157
Nagelkerke [R.sup.2] .218
Percent correct 70.1
McCain Too Old (1) vs.
Not Too Old/DK (0)
Model 2
B SE Exp(B)
Intercept .634 .323 1.885 (#)
Demographics
Female (1 = yes, 0 = no) .324 .075 1.383 ***
Age (in years) .000 .003 1.000
Black (1 = yes, 0 = no) .471 .132 1.602 ***
Hispanic (1 =yes, 0 = no) .361 .141 1.435 *
Education (in years) -.063 .018 .939 ***
Household income (in thousands) .000 .001 1.000
Political Characteristics
Republican (1 = yes, 0 = no) -1.316 .119 .268 ***
Democrat (1 = yes, 0 = no) .706 .082 2.025 ***
Ideology (1 = very liberal to -.222 .037 .801 ***
5 = very conservative)
Information Sources
Saw presidential campaign -.024 .018 .976
information on TV news
(0-7 days)
Heard about presidential -.023 .014 .977 (#)
campaign on talk radio
(0-7 days)
Saw presidential campaign .028 .013 1.028 *
information in newspapers
(0-7 days)
Saw or read presidential .038 .013 1.038 **
campaign information on
internet (0-7 days)
N 4054
Cox & Snell [R.sup.2] .161
Nagelkerke [R.sup.2] .223
Percent correct 69.8
Obama Too Young (1) vs.
Not Too Young/DK (0)
Model 1
B SE Exp(B)
Intercept -.807 .438 .446 (#)
Demographics
Female (1 = yes, 0 = no) .021 .106 1.021
Age (in years) .003 .003 1.003
Black (1 = yes, 0 = no) -.843 .335 .430 *
Hispanic (1 =yes, 0 = no) .197 .198 1.218
Education (in years) -.143 .024 .867 ***
Household income (in thousands) -.001 .001 .999
Political Characteristics
Republican (1 = yes, 0 = no) .200 .123 1.222
Democrat (1 = yes, 0 = no) -.782 .152 .458 ***
Ideology (1 = very liberal to .231 .054 1.259 ***
5 = very conservative)
Information Sources
Saw presidential campaign
information on TV news
(0-7 days)
Heard about presidential
campaign on talk radio
(0-7 days)
Saw presidential campaign
information in newspapers
(0-7 days)
Saw or read presidential
campaign information on
internet (0-7 days)
N 4061
Cox & Snell [R.sup.2] .043
Nagelkerke [R.sup.2] .087
Percent correct 89.2
Obama Too Young (1) vs.
Not Too Young/DK (0)
Model 2
B SE Exp(B)
Intercept -.765 .453 .465 (#)
Demographics
Female (1 = yes, 0 = no) .015 .107 1.015
Age (in years) .004 .004 1.004
Black (1 = yes, 0 = no) -.843 .336 .430 *
Hispanic (1 =yes, 0 = no) .159 .199 1.172
Education (in years) -.129 .025 .879 ***
Household income (in thousands) .000 .001 1.000
Political Characteristics
Republican (1 = yes, 0 = no) .197 .123 1.218
Democrat (1 = yes, 0 = no) -.784 .152 .457 ***
Ideology (1 = very liberal to .220 .054 1.246 ***
5 = very conservative)
Information Sources
Saw presidential campaign -.024 .025 .976
information on TV news
(0-7 days)
Heard about presidential -.003 .019 .997
campaign on talk radio
(0-7 days)
Saw presidential campaign -.038 .019 .963 *
information in newspapers
(0-7 days)
Saw or read presidential -.023 .019 .977
campaign information on
internet (0-7 days)
N 4061
Cox & Snell [R.sup.2] .045
Nagelkerke [R.sup.2] .091
Percent correct 89.2
(#) p < .10; * p < .05; ** p < .01; *** p <.001.
Source: NAES 2008 telephone survey; data unweighted.
TABLE 3
Mean Differences in Favorability Ratings by Candidate Age Perceptions
McCain Favorability Obama Favorability
Ratings Ratings
Do you feel John McCain is too old to be president, or not?
Yes, McCain is too old 3.47 *** 7.78 ***
No, McCain is not too old/DK 6.17 *** 4.93 ***
Do you feel Barack Obanta is too young to be president, or not?
Yes, Obama is too young 6.75 *** 3.51 ***
No, Obama is not too young/DK 5.01 *** 6.27 ***
(#) p < .10; * p < .05; ** p < .01; *** p < .001.
Source: NAES 2008 telephone survey conducted between October 15 and
November 3; data weighted.
TABLE 4
OLS Regression Predicting Candidate Favorability Ratings (0-10)
McCain Favorability Ratings
B SE Beta
Intercept 3.589 .301 ***
Demographics
Female (1 = yes, 0 = no) -.013 .071 -.002
Age (in years) .009 .002 .046 ***
Black (1 = yes, 0 = no) -.705 .138 -.063 ***
Hispanic (1 = yes, 0 = no) .092 .141 .008
Education (in years) .027 .016 .022
Household income (in thousands) .001 .001 .026 (#)
Political Characteristics
Republican (1 =yes, 0=no) 1.343 .092 .209 ***
Democrat (1 = yes, 0 = no) -1.023 .088 -.167 ***
Ideology (1 = very liberal to .417 .035 .164 ***
5 = very conservative)
Candidate Age Perceptions
McCain is too old to be -1.783 .080 -.290 ***
president (1 = yes,
0 = no/DK)
Obama is too young to be 1.043 .113 .112 ***
president (1= yes,
0 = no/DK)
N 4,139
[R.sup.2] .416
Obama Favorability Ratings
B SE Beta
Intercept 7.037 .326 ***
Demographics
Female (1 = yes, 0 = no) .113 .077 .017
Age (in years) -.001 .003 -.006
Black (1 = yes, 0 = no) 1.422 .149 .110 ***
Hispanic (1 = yes, 0 = no) .984 .153 .073 ***
Education (in years) .053 .018 .038 **
Household income (in thousands) .000 .001 .007
Political Characteristics
Republican (1 =yes, 0=no) -1.584 .100 -.211 ***
Democrat (1 = yes, 0 = no) 1.276 .095 .179 ***
Ideology (1 = very liberal to -.773 .038 -.260 ***
5 = very conservative)
Candidate Age Perceptions
McCain is too old to be 1.519 .086 .212 ***
president (1 = yes,
0 = no/DK)
Obama is too young to be -1.709 .123 -.157 ***
president (1= yes,
0 = no/DK)
N 4,127
[R.sup.2] .497
(#) p < .10; * p < .05; ** p < .01; *** p < .001.
Data: NAES 2008 telephone survey; unweighted.
TABLE 5
Vote Preference by Candidate Age Perceptions
McCain Age Perceptions Obama Age Perceptions
McCain McCain Not Obama Too Obama Not
Too Old Too Old/DK Young Too Young/DK
McCain 9.4% 50.7% 66.4% 32.1%
Obama 76.6% 33.1% 16.6% 52.6%
Nader 2.1% 1.8% 1.0% 2.0%
Other 0.3% 0.9% 1.2% 0.7%
Will not vote 2.3% 2.0% 3.6% 1.9%
McKinney 0.5% 0.2% 1.2% 0.2%
Barr 0.5% 1.0% 0.2% 0.9%
Don't know 5.2% 6.6% 6.3% 6.2%
Refused 3.1% 3.6% 3.4% 3.4%
Chi-square = 982.73, Chi-square = 334.79,
df = 8, p <.001 df = 8, p < .001
Source: NAES 2008 telephone survey conducted between October 15
and November 3; data weighted; column percentages given.
TABLE 6
Logistic Regression Predicting Vote Preference (1 = Obama, 0 = McCain)
Model 1
B SE Exp(B)
Intercept 1.275 .744 3.580 (#)
Demographics
Female (1 =yes, 0=no) -.155 .166 .856
Age (in years) .004 .006 1.004
Black (1 = yes, 0 = no) 5.345 .911 209.458 ***
Hispanic (1 = yes, 0 = no) .768 .312 2.156 *
Education (in years) -.025 .038 .976
Household income (in thousands) .001 .002 1.001
Political Characteristics
Republican (1 = yes, 0 = no) -1.160 .204 .313 ***
Democrat (1 = yes, 0 = no) 1.096 .215 2.992 ***
Ideology (1 = very liberal to -.525 .096 .592 ***
5 = very conservative)
Favorability Ratings
McCain favorability rating (0-10) -.794 .052 .452 ***
Obama favorability rating (0-10) .886 .048 2.425 ***
Candidate Age Perceptions
McCain is too old to be president
(1 = yes, 0 = no/DK)
Obama is too young to be president
(1 = yes, 0 = no/DK)
N 3599
Cox & Snell [R.sup.2] .668
Nagelkerke [R.sup.2] .893
Percent correct 95.1
Model 2
B SE Exp(B)
Intercept 1.003 .788 2.726
Demographics
Female (1 =yes, 0=no) -.235 .173 .790
Age (in years) .003 .006 1.003
Black (1 = yes, 0 = no) 4.896 .912 133.783 ***
Hispanic (1 = yes, 0 = no) .995 .326 2.706 **
Education (in years) -.033 .040 .968
Household income (in thousands) .001 .002 1.001
Political Characteristics
Republican (1 = yes, 0 = no) -1.140 .212 .320 ***
Democrat (1 = yes, 0 = no) 1.049 .223 2.855 ***
Ideology (1 = very liberal to -.540 .100 .583 ***
5 = very conservative)
Favorability Ratings
McCain favorability rating (0-10) -.708 .053 .493 ***
Obama favorability rating (0-10) .847 .049 2.333 ***
Candidate Age Perceptions
McCain is too old to be president 1.514 .219 4.544 ***
(1 = yes, 0 = no/DK)
Obama is too young to be president -1.257 .289 .284 ***
(1 = yes, 0 = no/DK)
N 3599
Cox & Snell [R.sup.2] .674
Nagelkerke [R.sup.2] .901
Percent correct 95.5
(#) p < .10; * p < .05; ** p < .01; *** p < .001.
Source: NAES 2008 telephone survey; unweighted.
