|
2004
PRIMARY TURNOUT LOW
March 9, 2004
GROUPED PRIMARIES LOWER THAN INDIVIDUALS
BETTER SELECTION METHOD NEEDED
Curtis Gans
Contrary to some published reports and with the singular exception
of the New Hampshire Democratic primary which set a new record high, Democratic
turnout in the party's Presidential primaries through Super Tuesday was
generally low - in the aggregate, the third lowest on record.
Republican Presidential primary turnout was the lowest on record. Turnout in
Presidential primaries in those states which held primaries for the nominee of
both parties was the second lowest on record, barely eclipsing the record set in
1996.
In all, an estimated 14,500,000 eligible citizens or 7.2 percent of the
national eligible electorate participated in the Presidential primaries through
Super Tuesday. Only an estimated 10,300,000 citizens or 5.1 percent of the
200,482,000 eligible Americans nationally participated in the selection of Sen.
John Kerry as the Democratic nominee. (Estimations made necessary by more than
one million still uncounted absentee ballots in California.)
A comparatively high turnout, greater than either 2000 or 1996 and perhaps
equaling or exceeding the 58.1 percent turnout of 1992 is expected in November.
But turnout is not expected to reach the 60 percent plus levels of the 1960s.
These are some of the highlights of a
report issued today on Presidential primary turnout (and other issues)
by the Committee for the Study of the American Electorate (CSAE), a
non-partisan, non-profit research organization specializing in voter turnout
issues.
Among the other highlights and lowlights of this report:
-Democratic turnout (an estimated 10.3 million) constituted 11.4 percent of the
eligible electorate in the 20 states which held primaries through Super Tuesday,
higher than the 9 percent which voted in the uncontested 1996 Presidential
primaries and the virtually (after New Hampshire) uncontested primaries in 2000
in which 10.1 percent of the eligible electorate voted. But it was lower than
the turnout for every other Presidential primary season in these states and more
than 50 percent lower than the primary turnouts of 1968 and 1972.
-Republican turnout (an estimated 4 million) constituted 6.6 percent of
the eligible electorate in the 11 states which held GOP primaries a smaller
percentage than in any primary season including the uncontested
Reagan renomination primaries (7.2 percent), the uncontested Nixon
renomination primaries (11.3) and more than 50 percent below the GOP
Presidential primary apex of 14.9 in 1980.
-Turnout in the 11 states which had primaries in both parties was an
estimated 12.6 million or 20.8 percent of the eligible electorate in both
parties eclipsing only the 20.4 percent turnout in 1996, when President Clinton
ran unopposed, and more than a third lower than the primary seasons of 1964,
1972, 1976 and 1980, all of which eclipsed 30 percent turnout rates.
-Democratic turnout reached a legitimate record high in New Hampshire
where 23.5 percent of eligibles turned out as compared to the previous
high of 20.2 in 1992. Three other states could claim "records" but
those
claims are somewhat specious. Arizona had a "record" turnout of 6.4
percent of eligibles but that was in comparison to only one other Democratic
primary held in that state, a virtually uncontested race in 2000 that garnered
2.6 percent of eligibles. Delaware had a "record" turnout of 5.6
percent of eligibles compared to 2 percent turnouts in the uncontested and
virtually uncontested races on 1996 and 2000 respectively. And South Carolina
could claim a "record" of a 9.5 percent
turnout against the only other Democratic primary held in the state in 1992 at
4.4 percent.
-Democratic turnout reached real record lows in Connecticut at 5.4 percent of
eligibles and in New York also at 5.4 percent. Another
"record" low was set in Virginia where only 7.5 percent of eligibles
voted. But there had been only one other Virginia Democratic primary -
in 1988 when 8.3 percent of eligibles voted.
-Vermont, in a posthumous (for his candidacy) show of support for Gov.
Howard Dean, recorded its second highest turnout ever at 16.5 percent of
eligibles as compared to the high of 19.4 in 1984.
-Republican turnout reached legitimate record lows in New Hampshire
(7.2 percent of eligibles as compared to the previous low of 10.6 in
1984), Oklahoma (2.6 versus 5.0 in 2000), Tennessee (2.2 versus 2.4 in
1984), Massachusetts (1.54 versus 1.57 in 1984), Wisconsin (4.0 versus 8.7 in
1984) and California (an estimated 10.7 versus 11.3 in 1984).
Missouri was also at a "record" low at 2.9 percent of eligibles
voting,
but there have been only two other GOP primaries in the state (in 1988
and 2000, both with turnout rates over 10 percent). The GOP barely
avoided another low primary turnout record in Rhode Island where only
0.34 percent of eligibles cast their ballots, but that actually beat the party's
1984 turnout where only 0.32 percent of eligibles voted.
Democratic turnout was a record in New Hampshire and a near record in the
Iowa caucuses in part because of the enormous amount of time and money spent in
those states, the latter (money) a record, because
competing in those contests involved retail campaigning and mobilization
and because Democrats (along with some independents and Republicans) are united
in their distaste for President George W. Bush, had the time tocomparison shop
and were propelled to the polls by that animus.
That turnout fell off sharply after those contests was not surprising.
The campaign moved immediately (within one week) to grouped contests (see below)
which were underfunded (resources having been severely depleted by Iowa and New
Hampshire), which had to be conducted through television ads and one-shot visits
to get free media coverage, which minimized grassroots mobilization and in which
candidates other than Sen. Kerry could compete in only a limited number of
states. It is not, therefore, surprising that the two highest turnouts in
Democratic primaries were recorded in New Hampshire (23.5 percent) and Wisconsin
(20.5), both the only primaries on their respective days.
It is also not surprising that after a primary campaign in which voter
involvement was limited to six weeks and which was conducted in only 20 of the
50 states and which was a brief wholesale campaign in most of
them, in almost all polls, between 20 and 30 percent of Americans still
do not know enough about either Sen. Kerry or Sen. John Edwards to
determine whether they have a favorable or unfavorable opinion of either. But
this truncated schedule, created by the Democratic Party,
may come back to haunt it, as it gives the GOP five months to define
Sen. Kerry before he has his optimum opportunity to present himself in
the best light at the Democratic national convention.
What continues to be disturbing is the low level of voter turnout -
between 30 and 50 percent lower than turnouts in the 1960s and 70s. What seems
apparent is that the religion of civic duty, which brought large percentages of
the electorate out to vote regardless of what was at
stake, has been severely undermined. Primary turnout should not be as
high as general election turnout. Primaries are for the active and
interested of each party (with, in some cases of open or semi-open
primaries, some augmentation by independents and voters from the
opposition party). But the level of disinterest in these primaries and
in attachment to these parties augurs ill for the American political
system.
HIGHER FALL TURNOUT LIKELY
Turnout levels and trends in Presidential primaries are not a predictor of
general election turnout. Primary turnout increased between 1984 and 1988 (24.0
percent of eligibles to 25.9) but general election turnout fell in 1988 to the
lowest level since 1924. Primary turnout declined between 1980 and 1984 (26.9 to
24.0) but general election turnout increased slightly in the 1984 general
election.
It is virtually inconceivable that general election turnout will not go up in
2004, likely to be equal or higher than the level reached in 1992 when 58.1
percent of the eligible electorate voted.
The Bush Presidency is the most polarizing of any since that of Lyndon
Johnson. Election 2004 will be played out on the field of major issues
and strong emotions. The public will be called upon to adjudicate deep
differences in almost every area of public concern. Both same-sex marriages and
the upcoming trial of Saddam Hussein are likely to be
side-shows buried in a major national debate over the direction of the
nation in both the domestic and foreign arenas.
There are precisely four things which could erode any potential turnout
increase: the economy could start producing a 250 - 300,000 net job gain each
month, the situation in Iraq could stabilize with no further violence and the
emergence of real democracy and large quantities of
weapons of mass destruction could be found in Iraq along with credible
evidence of pre-war ties to Al Qaeda. That would rob the debate of the
three most contentious issues: whether the economic policies of the Bush
Administration were actually benefitting the citizenry; whether the money, lives
and limbs spent in Iraq were necessary; and whether the President can be trusted
in what he says.
The fourth is the tenor of what, now that the primaries are over, will
be a very long campaign. It is in President Bush's interest to run a
relentlessly negative campaign against Senator Kerry (while attempting
to rekindle the memory of the unity behind his leadership created in the
immediate aftermath of September 11, 2001). His campaign would like to make Sen.
Kerry the issue and to depress turnout, whose growth this year will not benefit
the President. It is not in Sen. Kerry's interest to
respond in kind, at least in his television advertising campaign. But should he
do so, the public (and particularly the public that is not strongly identified
with either major political party) may come to believe that neither candidate is
worthy and a larger percentage than now seems likely may eschew the ballot box.
(Sen. Kerry might look to the example of former Colorado Gov. Roy Romer, who in
his re-election campaign announced that he would not run produced negative ads,
spoke about the issues directly in his advertising and won his election
handily.)
But whether the turnout increase is large or small, turnout this fall will not
reach the mid-60 percent levels of the 1960s. There are two concrete indicators
that demonstrate this: Contrary to public impression, the turnout for last
year's California recall election was not, at 47.1 percent of the electorate,
particularly high, despite the deep divisions in the state, the massive media
coverage, the enormous amounts of money poured into it and the star quality of
the elected governor. And the most recent survey of incoming college freshmen (a
group more likely to vote than the non-college attending 18-year-olds),
conducted by the Higher Education Research Institute at the University of
California at Los Angeles showed that interest in politics among this group had
risen slightly to 34 percent from its nadir of 29 percent four years ago.
However, interest in politics among the same group in the 1960s was 60 percent.
Too much has happened over the past 40 years to depress civic engagement: the
erosion of trust in political leadership; the decline in the quality of
education and in the quantity of civic education; the weakening of the nation's
integrating institutions - the schools, the churches, the unions and the
political parties; the fragmentation and atomization of American society -
physically through suburbanization, the Interstate Highway System and the strip
mall, among other things; politically through single-issue and identity
politics, and, most profoundly, by modern media which likes to claim it brings
the world community into the living room but brings the citizen into his or her
living room and fragments shared knowledge along channel lines and websites; the
decreasing coverage of politics and increasing cynicism about it in the media;
the execrable conduct of our political campaigns; and a set of dominant values
that emphasize individual material acquisitiveness, hostility to government and
consumer or libertarian choice, to name but a few.
These deep societal trends will not be reversed in one election, but one can
hope that the scope of the issues in this election will, at least in 2004,
engage a larger slice of the American citizenry.
Below are two commentaries on the primary process the nation has gone
through.
INDIVIDUAL VERSUS GROUPED PRIMARIES
Before 1988 and after the reform commission chaired by then North
Carolina Gov. Jim Hunt which reversed some of the more egregious
excesses of the reform commission chaired by Sen. George McGovern in the
aftermath of the 1968 nomination battle, the Democratic Party and the nation had
settled into a fairly coherent nominating process of sequential individual
primaries over a four month period, beginning in late February and ending in the
first week in June - leavened by the appointment of existing elected and party
office-holders comprising about 20 percent of the convention delegates.
In 1988, Southern Democratic state chairpersons, unhappy with the 1984
landslide defeat of Vice President Walter Mondale and the consequent loss of
many leadership positions in Southern states, attempted to exert some clout on
the selection of the next nominee by grouping their primaries and moving them
forward on the electoral calendar (the first Super Tuesday). The result was a
disaster for them. Of the 15 primaries that day, the Rev. Jesse Jackson won
five, the "northern liberal" candidate Gov. Michael Dukakis won five
and the "Southern" candidate, then Sen. Al Gore won only five.
But this did not stop other states from having the same idea - moving up their
primaries to theoretically exert more influence on who would be the party's
nominee. This led to an ever-shortened campaign and an ever-greater grouping of
primaries, which was further exacerbated this year, when, in the belief that it
was in the Democratic Party's interest to end the primary season as early as
possible, the party eliminated what had been a window of a few weeks between the
Iowa caucuses and the
New Hampshire primary and allowed seven states to hold primaries and
caucuses a week later.
The result has been lower turnout in the primaries which are grouped, due to
campaigns run largely through television advertising with little personal
contact and grassroots mobilization, often with diminished resources to conduct
these campaigns, lesser information about the candidates and consequent
progressively lower voter turnout when
compared to primaries which are held individually.
Since 1988, turnout was higher in grouped primaries than in individual
ones in only one year, 1992.
Using only competitive Presidential primaries in which no incumbent was
seeking renomination:
-Democratic turnout in the 2004 individual pimaries (including the District of
Columbia non-binding primary) averaged 16.2 percent of the eligible vote, while
average turnout in grouped priumaries was 10.3 percent of the eligible vote.
-Democratic turnout in the 2000 individual primaries was 11.3 percent
of eligibles as opposed to 10.6 in the primaries which were grouped.
Republican turnout in individual primaries was 19.1 as compared with
14.3 percent in the grouped primaries. In states which held primaries for
nominees of both parties turnout was 29.6 for individual primaries as opposed to
24.2 for grouped primaries.
-In 1996, Republican turnout for individual primaries was 12.3 percent
of eligibles while turnout in grouped primaries was 8.5.
-In 1988, turnout in states which held primaries for nominees of both
parties had a turnout of 28.7 percent of eligibles compared to a turnout
of 23.0 in grouped primaries. Democratic primary turnout was 16.6 for
individual primaries compared to 15.3 for grouped primaries. GOP turnout was
10.0 in individual primaries as compared to 7.7 in grouped
primaries.
-In the one exception that may prove the rule, Democratic turnout in
1992 individual primaries was 8.0 percent, as compared to 11.9 in grouped
primaries.
The preponderance of the evidence suggests that it is not only the front-loading
but also the grouping of primaries which undermines the
nominating system - a consideration to be taken into account before one
seeks remedy in a system of spaced but group primaries such as the rotating
regional primary system suggested by the National Association
of Secretaries of State.
THERE MUST BE A BETTER WAY
In early January, Sen. John Kerry rolled the dice. Despite being far behind in
all the public opinion polls, he lent his campaign $6.4 million from his
personal wealth and concentrated all his resources on the caucuses in Iowa and
the first in the nation primary in New Hampshire. It was a gamble that paid off
handsomely.
It paid off because Gov. Howard Dean squandered the $41 million he
raised, spreading those resources across the nation; lost his lead in the polls
by a series of ill-thought out statements and relentless press criticism;
engaged in a suicidal battle of negative advertising with and against Rep.
Richard Gephardt and failed to balance his clear and vigorous advocacy on issues
with his moderate and effective record of
governance in Vermont so that he appeared more of a protest candidate
than a potential President. It paid off because Gen. Wesley Clark
abandoned Iowa for New Hampshire and because Sen. John Edwards, who might have
benefitted equally from the collapse of the Dean campaign, did not devote
commensurate resources to the first two contests. (It also paid off because of
Kerry's relentless campaigning on the electability issue, undercutting Dean, and
the week that Sen. Edward
Kennedy spent in Iowa campaigning for Kerry.)
But Kerry's gamble was successful in large measure because of a primary schedule
which gave an enormous advantage to the winner of the first two contests. With
barely a week between the New Hampshire primary and seven contests across the
nation, Kerry could ride a wave of free publicity and augmented funding to
campaign in all states, while his rivals could only concentrate limited
resources on two states, to keep their hopes barely alive. Additional free
publicity and funds for Kerry and limited time and resources for the other
candidates provided enormous advantages for Kerry in sewing up the race on Super
Tuesday.
The Democrats, it says here, got lucky. They emerged from this truncated process
with a capable candidate, a plausible potential President and a unified party.
They did so, in large measure, because the differences between the Democratic
candidates were minuscule compared to their differences with the President and
thus there was a
minimum of internecine hostility. But by truncating the process, they lost the
advantage they had in the daily coverage of criticism of the President and his
policies, propelled themselves into a four-month dead period before the national
conventions in which citizen interest will decline, the opposition will have the
opportunity to define the nature of the race and the party will have to
remobilize its supporters.
In a larger sense, the 2004 nominating process is an accident waiting to happen
and the evolution of this process from the first post-1968 reforms has produced
the least desirable method of selecting the nominees of the two major parties
for the highest office of the most powerful nation in the world.
Why is the process so bad? Let us count the ways:
1. By frontloading and grouping the primaries, it makes it mandatory
for candidates to raise large sums of money early, because of this forces,
candidates to announce and begin their campaigns and fundraising at least a year
prior to the first vote being cast, restricts candidacies to the well-known, the
wealthy or those with access to the moneyed and, as in this year, makes it
desirable for candidates to opt out of the Presidential public financing system.
2. This, in turn, creates a year-long ersatz campaign during which the media,
rather than the electorate, sits in judgement on the candidacies and colors the
outcome.
3. The influence of winning Iowa and New Hampshire becomes exaggerated beyond
any reasonable proportion. Because there is no time between these two contests
and the first grouped primaries, it becomes impossible for any but the winner of
those contests to effectively raise money, counter the free publicity of that
winner and compete effectively in the contests to follow. At least in 2000
President Bush had the individual primaries in South Carolina and Michigan
following his New Hampshire defeat to recoup his losses. One can only imagine
what would have happened if this truncated system had been in place had Pat
Buchanan actually bested the first President Bush in 1992 in New Hampshire.
4. A six week campaign is not an adequate length of time to evaluate and
re-evaluate the merits of a potential candidate and should one establish his or
her primacy within those six weeks and later be found to have major defects,
there is no reasonable democratic recourse, such as a late entering candidacy.
5. By grouping the primaries after Iowa and New Hampshire, it emphasizes the
least desirable forms of campaigning - 30-second ads -
and de-emphasizes the most desirable - personal contact and grassroots
mobilization.
6. The six-week campaign leaves the majority of the nation out of the
decision-making process altogether and, to a great extent, without knowledge
about the personalities, character, qualifications and policy
positions of the nominees, unless they are sitting Presidents.
7. The current system renders the national conventions of both major
parties as relatively meaningless exercises in political propaganda, inviting
the media to do what they have been progressively doing - reducing and
eliminating coverage.
8. This system leaves the nation with a five month dead period before the major
party conventions, diminishing political interest and forcing a remobilization
in the fall.
In 2000, the Delaware Republican Party chairman offered a plan which
would address all of these concerns. What he proposed was that the primary
process be re-elongated to what it had been before - a four month campaign and
scrutiny of the candidates. What he proposed was that the states with the fewest
delegate votes have an initial month between mid-February and mid-March to hold
their primaries. Those with the second fewest delegate votes would have the next
month to hold their primaries and the ensuing months would allow those with the
second most and the most delegates to hold their primaries.
Under such a system, candidacy would not necessarily be limited to the
rich, the famous and the connected, but would allow a modestly financed
candidate to enter the race and, if successful, raise money for ensuing
contests. It would end the long ersatz campaign. It would reduce or
eliminate the worst of the grouping of primaries and permit a much greater
amount of retail campaigning. It would permit reasoned scrutiny
of the candidates and allow for re-evaluation. It would permit candidates to
determine their own geographical strategies and allow for late entering
candidacies. It would help to sustain voter interest throughout the election
year.
Under such a system, those with primaries in the first and second tier would
likely winnow out the field, leaving the ultimate decision to the
larger states, or if that fails to produce a winner, the national convention
which would continue to have a bloc of party leaders. It would put the
decision-making in the hands of the voters rather than the media. And it would
likely allow all candidates to stay within the Presidential financing system,
rather than opting out.
This plan was supported by the rules committee of the Republican Party
in 2000 and by the Republican National Committee, before being waylaid by
nominee George W. Bush who did not want to have a rules floor fight on the
national convention floor. It was supported, for a while, by a majority of the
Democratic rules committee before Vice President Al Gore, at the behest of some
of his staff and some large state party leaders, scotched the plan.
Every Presidential election year since 1988, attempts have been made to
right the wrongs of the previous nominating process. And every year, the
system has been made worse. It is time to resurrect the Delaware plan and
provide a durable basis for selecting the people who will lead the most powerful
nation in the world.
NOTES AND SUMMARY CHARTS
1. What is Turnout: Turnout should be a simple calculation in which the
numerator is the number of votes cast and the denominator is the number
of citizens eligible to vote. But because of various anomalies in election
statistics, some of which are outlined in detail below, this calculation is more
complicated. By common usage, the numerator in every Presidential election year
is the vote for President (even though that
tally is usually about one percentage point lower than the actual number
of citizens who go to the polls. It is lower because many states, although an
ever-diminishing number, do not keep records of all those who go to the polls,
the total ballots cast). In mid-term elections, the numerator is the total of
votes for the statewide race in each state which draws the highest number of
votes and the aggregate total of votes for U.S. House of Representatives in
those states which do not have statewide races. (This total tends to be between
1 and 1.5 percentage points lower than the actual total ballots cast but is used
for the same reasons - that many states do not compile total ballots cast
figures.)
Turnout is NOT the percentage of those registered who voted. There are
three basic reasons for this: a. Using registration as a denominator does not
account for the whole of the electorate, including those who are not registered.
Thus, it gives a false picture of true citizen engagement. b. Changes in
registration law can dramatically affect the figures. If the nation adopts, as
it did, a registration law that provides for national mail registration,
registration at motor vehicle bureaus and at social service agencies,
registration will go up but turnout of those registered will decline
artificially by a greater amount than it does when using the entire eligible
electorate as a denominator. c. Registration figures are subject to the
fluctuations of election administration. If a state conducts a thorough purge of
its registration lists close to election, its registration figures will be lower
and thus its percentage of registered voting will be higher. But
if registration lists are not so purged, as they are not in many states, the
figures for registration will be higher and the turnout based on these inflated
registration figures will be lower. Consider how distorted a turnout percentage
using registration as a base would be in a state such as Alaska, which because
of lack of regular list cleaning and potential flaws with the Census Bureau's
estimates of the state's eligible population, registration figures are regularly
in excess of 100 percent of the eligible vote.
2. The Eligible Vote - The Denominator for Determining Turnout: The
eligible vote in this report is the number of people residing in the United
States who are 18 years of age or over minus the number of non-citizens residing
in the United States who are 18 years of age and over as of April 1. It is an
interpolated figure from the 2000 Census, based on the methodology outlined
below.
For years, CSAE and every other reputable organization working in this
field had used the Census Bureau's estimates of November age-eligible
population (VAP) to determine turnout. That figure came under legitimate
criticism because it included non-citizens; convicted felons (in most states)
and, in some states, ex-felons; and people deemed mentally
incompetent in institutions who could not vote and did not include citizens
residing in other countries, citizens naturalized during the election year and
the citizen portion of the Census' undercount, all of whom could vote but were
not part of the VAP estimate. The Census Bureau has ceased providing its VAP
estimates.
For years also, Dr. Walter Dean Burnham, professor emeritus at the University of
Texas at Austin, has been producing a denominator of age-eligible citizens
(age-eligible population minus age-eligible non-citizens, interpolated by state
and nation from and between decennial Censuses). After some study of this
matter, CSAE has come to believe that this denominator is the best for
determining turnout, subject to the caveat below. It has come to this belief
because of two factors:
1. Available data. One does not determine turnout simply for any given
year but also as an historical comparison with previous years. Data for
several of the issues involving the inadequacy of the age-eligible population (VAP)
figures are either simply not available, not available
in a timely manner, not available over a given period of history or not
allocatable to the states. Data on convicted and incarcerated felons is
only available for a fairly recent time period. State laws on whether
convicted felons and ex-felons can vote are changing and have changed
over time. There is no accurate set of figures on those deemed mentally
incompetent. The number of American citizens residing abroad is
ascertainable but the number of age-eligible has to be estimated and
there are no figures which allow the allocation of these citizens by state.
Naturalization figures come in too late, often a year or two after the election
year, to be usable in any current population accounting. And while any given
Census undercount can be allocated by state, one can only estimate how much of
that undercount is of citizens as opposed to non-citizens.
2. The balance of the figures: In studying this statistical problem, CSAE has
found that the most important issue is that of non-citizens. If one wants to
have a relatively accurate picture of turnout, one must eliminate the
non-citizens from the age-eligible population. On the other hand, the other
adjustments to the denominator would not substantially differ from the
denominator of citizen age-eligible population. In pursuing its inquiry into
this topic, CSAE found that the factors which would lower the denominator -
felons, ex-felons and people deemed mentally incompetent who can't vote -
are roughly equal to two of the factors which would increase the denominator -
citizens living in other countries and naturalization who could vote. If one
added a ballpark figure for the number of citizens in the undercount who could
vote, the factors in those years of an undercount, other than non-citizens,
which would increase the denominator exceeds those which would reduce it.
The one caveat in adopting the Burnham methodology lock, stock and
barrel is that Burnham interpolates from Census to Census. These
Censuses are accurate as of April 1 of each decennial year for all of
the past 50 years. (In prior years, Census results captured the population as of
varying months.) In order to have more accurate figures for November, CSAE has,
using the same methodology, projected citizen
population to November. Thus, CSAE is using in this report and for all
future reports on primaries, the April figure for age-eligible citizen
population, but will use a November figure for the general election.
Methodology
Since the decennial census population figures are accurate as of April 1
in each census year, the VAP Burnham dataset calculates the difference
in the required census figures between a base census year and the same
figures as reported in the following census. To estimate the voting age
population for the years between the censuses, the difference between
them is simply multiplied by the number of months that have passed
beyond April 1 of the base year and then added to the base year figure.
For example, to arrive at the April 1, 1992 voting age population, the
difference between the April 1, 1990 census population and the April 1,
2000 census population is multiplied by 24/120ths (for the 24 out of 120
months between the census counts) and added to the April 1, 1990 figure.
The process for arriving at the CSAE November Eligible figures is the
same, except that the data is projected forward to November instead of
April. To accomplish this, the multiplier is simply changed to the number
of months that have passed since April of the base census year. For instance, to
calculate the November 1996 voting age population, the difference between April
1, 1990 and April 1, 2000 is multiplied by
79/120ths and added to the April 1, 1990 count. The same interpolation
process is applied to the decennial census counts of non-citizens of voting age
in each state. Once estimates of the total voting age population and the
non-citizen voting age population for each state have
been calculated, the non-citizen figure is simply subtracted from the total to
arrive at the appropriate figure.
Since the last decennial census occurred in 2000, it is necessary to project the
figures forward to arrive at the voting age population for 2002 and 2004.
To accomplish this, the difference between the 1990 and 2000 decennial censuses
is used to establish a rate of growth. This
rate of growth is then used to project forward based on the number of
months passed since April 1990 out of the 120 months between the censuses.
For instance, to obtain the voting age population for April 2004, the difference
between April 1, 1990 and April 1, 2000 is multiplied by 168/120 and added to
the April 1, 1990 total.
3. The Vote - Primaries: The vote totals in this report are final and
unofficial (except Arizona, Delaware, Missouri, New Hampshire, Oklahoma,
Virginia and Wisconsin, which are official) and provided by the chief election
officer in each state. Final and official figures will be very close to those
reported here, save for California which is still counting more than 1 million
absentee ballots. The vote for President
was the highest in all states, save two - California where the vote for
ballot propositions on bond issues, spending limits and others exceeded
the vote for President by about four percentage points; and in Ohio
where the votes cast for Sen. George Voinovich's renomination exceeded the votes
cast for President Bush by one percentage point. Because this is a report on
Presidential primary turnout, the figures contained in this report are only with
respect to that issue. Later reports will provide data on full turnout,
including primaries for other offices and,
where relevant, ballot propositions
In this report, CSAE has produced two frames of analysis of the primary
turnout - individual state charts and a collective set of charts of all the
states which have held primaries. The collective charts are divided
- including charts of all states which held primaries this year whether or not
they had primaries in previous years, and of those states which held primaries
this year's for which there is continuous data going back several years. The
charts which are labeled total turnout are states which held primaries in both
Democratic and Republican parties.
NOTE: The determination of primary turnout is not a perfect science. The tables
in back reflect the figures for the votes which actually have
been counted so far. The estimated numbers in the text of this report
with respect to total votes in the primaries represent CSAE's best estimate of
the votes which are likely to have been cast after all of these states issue
their final figures and California's absentee ballots are counted. And long
range comparisons with previous high turnout years
represent CSAE's best judgement of how to compare turnout in the present with
the more limited number of primaries which were held in previous years.
4. The Vote - Caucuses: CSAE does not track caucus voting totals. But it did
ascertain that while the raw number of citizens participating in the
Iowa caucuses was similar to the number which participated in 1988 (the
high point for Iowa caucus turnout), the population of the state had grown
sufficiently to make the percentage turnout lower than in 1988.
5. Individual vs. Grouped Primary Comparisons: In this part of CSAE's
report, it attempted to compare turnout rates between those primaries
which were the only primaries on a given date and those which were
grouped with other states to give some indication of the relative impact
on turnout of retail (grassroots intensive) campaigns versus the media
campaigns which are necessitated when primaries are grouped. All
comparisons are of primaries which were held before the races were
decided in each year in each party.
6. Acknowledgments: Primary research for this report was done by Mark P. Harvey,
CSAE research associate, who is also responsible for creating the denominator
database for the analysis of November turnout.
Organizing the analysis for this report was made profoundly easier by a
custom database program developed by Samuel Schreiber, CSAE research associate
emeritus. CSAE would also like to express its profound gratitude to Dr. Walter
Dean Burnham, professor emeritus at the University of Texas at Austin, for
sharing his database, helping to
devise CSAE's new November denominator for the analysis of registration and
turnout and for his continuing help to CSAE's work.
7. Culpability: The analysis contained in this report has been done by
Curtis Gans, CSAE's director, who is solely responsible for any and all
errors contained within.
SUMMARY CHARTS
1. Turnout Comparisons: Comparisons between Presidential primary
turnout, as a percentage of eligible voters, in 2004 versus previous
years. 2004 primary turnout is estimated due to an indeterminate number
of absentee ballots, exceeding one million, still to be counted in California.
Overall turnout is for those states with both Republican and
Democratic primaries.
YEAR OVERALL DEMOCRATIC REPUBLICAN
2004 20.8 (est.) 11.4 (est.) 6.6
(est.)
2000 26.4 10.1 14.9
1996 20.4 9.0 11.2
1992 24.9 12.7 10.1
1988 27.3 15.8 10.0
1984 23.9 15.4 7.2
1980 30.7 13.9 14.9
1976 32.0 19.3 12.6
1972 34.7 23.1 11.3
1968 27.7 24.3 11.2
1964 36.6 21.5 15.2
2. Individual Versus Grouped Primaries: Presidential primary turnout as
a percentage of eligible vote in individual as compared to grouped
primaries since 1988 during the period before the nominee was decided in those
primaries which were contested within a give party (Total turnout is for those
states which held nominating primaries in both parties):
YEAR PARTY INDIVUDUAL GROUPED
2004 Democratic 16.2 10.3
2000 Total 29.6 24.8
2000 Democratic 11.3 10.7
2000 Republican 19.1 14.3
1996 Republican 12.3 8.5
1992 Democratic 8.0 12.2
1988 Total 28.7 23.0
1988 Democratic 16.6 15.3
1988 Republican 10.0 7.7
|
