Background
This report presents our analysis of an exit poll
conducted in Ukraine during the first round of the
presidential election, October 31, 1999. The exit
poll accurately predicted the outcome of the first
and second round voting; the results of the exit
poll were disseminated to the media in Ukraine on
election night. This report comments on the evolving
political culture in Ukraine, and provides a
detailed look at the characteristics of the voters,
their electoral behavior and their decision making.
Introduction
In the past two years, Ukraine’s electorate
went to the polls three times in national elections
— twice this year in presidential balloting and
once last year in a parliamentary election (October
31 and November 14, 1999, and March 29, 1998
respectively). These elections were held under
Ukraine’s new constitution, giving voters a wide
choice of competing candidates and political
parties, representing a broad ideological spectrum.
This report documents the results of an exit
poll, which was conducted during the first round of
the presidential election on October 31, 1999.
Reference is also made to the exit poll conducted
during the parliamentary election last year, and
comparisons are drawn between these two surveys. And
of course, both exit polls are compared with actual
reported election results. These results provide
important insights into Ukraine’s evolving
political culture, and help put this election into
an historic context.
Exit polls are an important resource both in
understanding and promoting the political
development of a country. They serve three purposes:
1) exit polls provide an independent measure of the
integrity of the voting process; 2) they make the
balloting process more transparent and give news
organizations something to report on election night
when public interest is highest; 3) they provide a
profile of the electorate which is of substantial
political and sociological value.
QEV Analytics, a Washington, DC, based opinion
research and communications firm, designed the
questionnaire, the sample, and the data reporting
system for the 1999 exit poll, prepared the press
releases and analyzed the results. The data
collection (interviewing) was accomplished by a
consortium of three leading Ukrainian public opinion
research firms: KIIS (the Kyiv International
Institute of Sociology), SOCIS (a Kyiv-based polling
firm), and UMS (the Ukrainian Monitoring Service).
The Democratic Initiatives Foundation of Kyiv
coordinated the project overall, secured funding,
and disseminated the results.
For the 1998 exit poll, QEV Analytics also
defined the methodology, analyzed the results,
preparing press releases and a comprehensive report.
The Kyiv polling firm SOCIS did the data collection
that year, and overall project management was the
responsibility of the Democratic Initiatives
Foundation.
The 1999 exit poll was made possible by a grant
from the Charles Stewart Mott Foundation. Funding
for the 1998 exit poll was provided by a grant from
the Eurasia Foundation.
Both the 1998 and 1999 exit polls successfully
predicted the election outcome. Key findings from
the polls were disseminated on both election nights.
The ability to conduct such polls and to disseminate
the results on election night news broadcasts are
testament to the degree of openness and freedom of
expression in Ukrainian society, and the growth of
independent media. The success of these elections in
terms of administration and popular participation
speak to the broad acceptance of competitive
elections and of a multiparty political system. This
positive assessment does not ignore that a
proportion of the electorate wants to revert to
communism; nor does it deny there is a lack of
liberalization in many areas of Ukrainian society.
Lessons from the 1999 Presidential Election Exit
Poll
Ukraine passed on November 14 a landmark event on
the road of her political evolution. As the world
now knows, incumbent President Leonid Kuchma won
reelection: he bested 12 opponents in the
preliminary round with a plurality of 38 percent.
Because he did not receive an outright majority,
Kuchma faced Communist Party candidate Petro
Symonenko in a run off on November 14. Kuchma
decisively defeated Symonenko by the margin of 58
percent to 39 percent.
Our exit poll — in which 6034 Ukrainians were
interviewed as they left 300 polling places
throughout Ukraine — accurately predicted both the
results of round one (see figure 1) and of the
runoff (figure 2). In round one, the exit poll
slightly overstated support for President Kuchma.
Interestingly, an independent parallel vote count
conducted by the Ukrainian non-governmental
organization "Coalition of Voters" found
the same thing: more votes for Kuchma than were
actually reported by the Central Election Commission
(CEC). The exit poll also understated the Moroz
vote, but otherwise came very close to replicating
the actual vote count. The narrowness of the margin
of error between the exit poll and the actual round
one result validates the exit poll methodology.
The exit poll runoff prediction is based on the
question, "if the runoff is between Leonid
Kuchma and Petro Symonenko, for whom would you
vote?" Since the runoff was just two weeks
after round one, there was little opportunity for a
substantial shift in voting intentions between the
conduct of the exit poll and the runoff ballot.
As with the round one vote, the exit poll runoff
question came exceedingly close to the actual
results, overestimating the extent of "none of
the above" ballots a bit, and underestimating
the Symonenko share of the vote. But remember: our
prediction for the runoff is based on those who
voted in round one. Supporters of losing candidates
in round one are more likely than other voters to
say they are going to vote for neither candidate in
the runoff. But a flood of new voters came the polls
for the runoff, and they apparently voted for one of
the two contenders rather than for "none of the
above."
The closeness of the exit poll result and the
actual runoff result is especially interesting in
light of this increased turnout between round one
and the runoff: in round one, 67.4 percent of
eligible voters participated; in the runoff, 74.9
percent participated. The narrow margin of error
implies that those additional voters who came to the
polls for the runoff were not substantially
different from round one voters, meaning they were
not drawn from any particular demographic or
attitudinal segment of the population.
While we had in-hand on the evening of October 31
evidence of Kuchma’s looming large margin of
victory in the runoff on November 14, this data was
not released in Ukraine, owing to prohibitions on
the release of survey data 15 days prior to an
election. Consequently, we knew that media
predictions of a close race, or even of a Symonenko
victory, were wrong. These predictions were based on
a redistribution of the 41 percent of the vote which
went to losing candidates. It was assumed, for
example, that virtually all of the Moroz and
Vitrenko vote would go to Symonenko, while the
Marchuk vote would split between Kuchma and
Symonenko, leaving Symonenko at or above 50 percent.
Voters did not behave according to this script
for five reasons. First there is a pervasive
ideological ambiguity in Ukraine, revealed to us
through a series of focus groups conducted in
September. Ms. Vitrenko may look like a leftist to
Western eyes; but back in Ukraine, 31 percent of her
voters who said they intended to vote in the runoff
supported Kuchma, 23 percent would vote against both
Kuchma and Symonenko, and only 47 percent intended
to support Symonenko (figure 3).
Similarly, we think Kuchma picked up 37 percent
of the Moroz vote (56% went to Symonenko, figure 4)
and over half (53%) of the Marchuk vote (27% went to
Symonenko, figure 5).
We detected in focus groups that none of these
candidates, with the exception of Symonenko, had a
clear, black-and-white ideological identity. And so
the redistribution of the round one vote could not
be accomplished in the straight-line method of
Western journalists.
The second factor in Kuchma’s resounding
reelection victory is that a majority of Ukrainians
voters are right-of-center, supporting
democratization and economic liberalization, and
Kuchma emerged as the candidate of the right. The
sum of right-of-center political parties in last
year’s Rada (Parliament) election received more
votes than the sum of left-of-center parties, but
because of great factionalization on the right, the
left — the Communists in particular — dominate
the Rada (thwarting even the tepid reforms advanced
by the Kuchma administration).
This year, the combined vote for right-of-center
candidates Kuchma, Udovenko, Kostenko, and Marchuk
exceeded the left-of-center votes for Symonenko,
Vitrenko and Moroz.
The largest of the right-of-center reformist
parties, Rukh, was split on the presidential
election between two candidates, the senior Udovenko
and the younger Kostenko. But these two candidates
collectively received a trivial 4 percent, because
President Kuchma emerged as the preferred candidate
of the right. President Kuchma’s best showing was
in the Western region (where he got 63%), a place of
mostly ethnic Ukrainians (as opposed to ethnic
Russians) and traditionally Rukh’s home turf.
The third reason for Kuchma’s victory margin is
that he was competitive with Symonenko in the East
(where the industrial city of Kharkiv is found) and
South (Odessa and Crimea), among majority Russian
populations who are traditionally enthusiasts for
the Communist Party. And this is the most dramatic
evidence available of the political convergence
which is underway in Ukraine.
The evidence from focus groups is that ethnic
Russians in Ukraine (representing 21% of the
electorate) are setting into being Ukrainian. Russia
is widely recognized as a political and economic
failure by the people of Ukraine. So with each
passing year, popular enthusiasm in Ukraine for some
sort of reaffiliation with Russia wanes, even among
ethnic Russians. And reaffiliation — more than the
reconstruction of a state-controlled economy — is
the raison d’être of the Ukrainian left.
Remove popular support for reaffiliation, and you
take away the heretofore strongest appeal of the
Communist Party. So Symonenko received only half of
the votes cast by ethnic Russians, and beat Kuchma
only narrowly (41% to 31%) in the industrial East
— not enough in either case to overcome the Kuchma
lead among ethnic Ukrainians voters.
The end of a serious re-affiliation debate, by
the way, is also likely to help right-of-center
parties like Rukh, which have heretofore been pulled
toward nationalist rhetoric (instead of toward
serious political and economic liberalization
rhetoric) by the Communist threat to Ukrainian
independence.
The final reason for Kuchma’s victory is that
he was not held accountable for economic failure.
Perhaps, had his opponent not been a Communist (who’s
policy prescriptions were clear and controversial),
Kuchma might have been held accountable by the
voters for recent economic performance. But this is
unlikely: our focus group research found a cultural
reluctance of voters to assign Kuchma blame for the
economy, because the economy is beyond the power of
the President to really control, and because there
is no confidence Ukraine can do better than Kuchma’s
management.
So Kuchma faced Symonenko in a fair, mano-a-mano
contest. Symonenko and the Communists were
decisively defeated. This election will come to be
remembered as the passing of the high water mark of
communism in Ukraine. The threat of a Communist
resurgence has been laid to rest, the hearts of the
people do not lie with the Communist Party. So now
the political debate can shift to a new set of
issues and non-Communist members of the Rada ought
to be emboldened to pursue reforms despite Communist
intransigence. All in all, this election was a
landmark event for Ukraine.
Importance of Campaigns
The Exit Poll sought to determine when
individuals made up their minds on how they would
vote. An analysis of this question can indicate the
importance of political campaigns and pre-election
information programs, identify how many voters are
committed, and describe the potential number of
voters receptive to political information.
The 1999 exit poll used a very direct question to
establish voter time line: "When did you decide
for which candidate you would vote?" Six
possible answers were recorded: always knew; more
than three months ago; more than a month ago, but
less than 3 months; more than one week ago, but less
than a month ago; within the past week; decided on
election day. Only a few voters (1%) could not or
would not say when they made their decision.
About half of the voters made up their minds well
in advance of the presidential campaign (figure 7).
A plurality (43%) "always" knew how they
would vote and an additional 16% knew more than
three months before the election. Among the rest,
about one-fifth made the decision a week before the
election or on election day (10% and 8%
respectively). Roughly as many (also one-fifth)
decided a month or two before the election (9%
within a month; 13% more than two months ago).
The time line on the electoral decision differs
slightly among the urban and the rural voters. About
one-fourth in each group made up their minds during
the campaign (a month or a week before the election
or on election day). As for the rest, rural voters
tended to make a decision closer to election day
than did urban voters. Moreover, slightly fewer of
the rural than of the urban voters said they
"always" knew for whom they would vote
(compare 41% rural to 44% urban voters).
A roughly similar pattern was recorded in the
1998 exit poll. At that time, urbanites were
slightly more likely than rural voters to have
decided on their party vote well in advance of the
campaign (compare 43% of the urban versus 39% of the
rural voters). The differences in the time line of
the urban and the rural voters in the 1999 and the
1998 election, in large measure, can be accounted
for by the educational level and the age profile of
the two populations.
On the time line, the main difference among
educational groups was that the best educated were
much more likely than others to make up their minds
during the campaign. Although pluralities in all
educational groups "always" knew how they
would vote, such a commitment was slightly more
prevalent among the least educated than among voters
with secondary or higher education. However, in all
educational groups, about one-in-ten made their
decision in the week of the election or on election
day. (See figure 8 on next page).
Voters of different age groups also had different
time lines in deciding on their vote. As figure 9
shows, the time line on the electoral decision
varies notably by age, except for the last minute
decision-makers. About one-in-ten in all age groups
decided on election day, with the youngest voters
(under 29 years of age) slightly more likely to have
made a last-minute decision than did those over 60
years of age. About one–fourth of the voters under
39 years of age made up their minds during the
campaign, whereas among older age groups, less than
one-fifth did. In all age groups sizeable
proportions said they were committed to a candidate
long before the campaign started, and this
commitment increased with age.
The exit poll data suggest that young adults are
open to information and make decisions on the basis
of what they hear. The demographic profile of voters
who "always" knew how they would vote in
the presidential election (43%) supports this
conclusion. A vast majority of the voters who
"always" knew were over 40 years of age
(one-half are over 50 and another one-fifth between
40 and 49 years of age). Young adults constituted a
minority of those who "always" knew (of
this group, 16% are between 18 and 29, and 14% are
30-39). This distribution of those who
"always" knew how to vote bodes well for
Ukraine’s political scene. The young are open and
willing to listen, to entertain new policies and to
process new information. The fact that so many of
the eldest age group "always" knew for
whom to vote is probably a legacy of the former
system, where political choices were not a matter of
open political discussion, not a process of freely
selecting from a range of options.
There were some differences in the time line for
the presidential candidates, with one exception.
Each candidate received about one-tenth of his vote
from the last minute decision-makers, those who
decided how to vote on election day. The following
analysis discusses only the appeal of the leading
five candidates – Kuchma, Symonenko, Moroz,
Vitrenko and Marchuk. The discussion is limited to
these five mainly because the small number of votes
for the other candidates precludes drawing
statistically reliable conclusions.
For the two leading contenders, Kuchma and
Symonenko, the time line for selecting a candidate
was roughly comparable (figure 10). About two thirds
of the voters for Kuchma, as well as for Symonenko,
knew well in advance how they would vote: one-half
of all of the voters for these two candidates said
they "always" knew (50% and 52%
respectively, and an additional 16% for Kuchma and
14% for Symonenko knew at least three months in
advance).
The time line for Moroz, Vitrenko and Marchuk
differed notably from that of the two leading
contenders (figure 11). The main difference was in
the far smaller proportion of voters committed to
these candidates far in advance of the election.
Comparing the five leading candidates, Marchuk
picked up more voters during the campaign than
others did. Also, he had the smallest proportion of
those
who were committed well in advance of the
election. Only one-fourth of Marchuk’s voters
"always" knew that they would vote for
him, but most (about a half) decided during the
campaign.
Voters for Moroz and Vitrenko had roughly similar
time lines, with one exception. Vitrenko had more
long-term, committed supporters than did Moroz. Of
the Vitrenko voters, four in ten were
"always" in her camp, whereas among Moroz
voters, only three–in-ten were.
The 1999 and the 1998 exit polls recorded very
similar decision making among age and educational
groups. In 1998, young voters, those under 30 years
of age, tended to make their decision much later
than older adults did. A half (51%) of those over 44
years of age decided well in advance of the
election, whereas only 28% of those under thirty did
so. Moreover, the largest proportion (41%) of young
voters (under 30) decided one month or one week
before the election. Similar to other age groups,
only a small proportion of young voters (7%) decided
at the voting place.
In the 1998 election, voters recorded significant
differences on the time line by party vote. Definite
majorities of voters for the two leading parties,
the Communist Party and Rukh, made up their minds
well in advance of the campaign (68% and 62%
respectively). The decision to vote for the other
parties was made later: only one-third to one-fourth
of the voters for the other parties, i.e., other
than the Communist Party and Rukh, made up their
minds before the campaign began. Most voters for the
other parties tended to make up their minds during
the campaign, ranging from 46% of voters for the
Reform and Order Party to 33% of those who voted for
the Progressive Socialist Party.
A comparison of the voters time line in the 1999
presidential and the 1998 parliamentary election
showed that many in Ukraine, especially young
adults, make up their electoral choice during a
campaign. These data document that a sizeable
proportion of voters are attentive to the messages
of the candidates, be these political parties or
individuals.
The different time-lines for the presidential
candidates may, in large part, reflect name
recognition. However, that is too facile a
rationale, too simplistic an explanation for an
electorate that has demonstrated its political savvy
and critical thinking. Therefore, taken within the
context of Ukraine, the exit poll data on time lines
suggest that campaigns had a significant role on the
electoral decision of voters.
Empirical data are not available to confirm how
important campaigns are to those who do and do not
vote. Nor is there data that would permit examining
if a campaign reinforced voting decisions. Findings
on the time-line cannot be used to evaluate
campaigns or messages of particular candidates (or
of political parties). The data record when voters
made up their minds, without corroborating evidence
of what they used to make up their minds. However,
what the exit polls unambiguously showed was that a
large numbers of voters did make up their minds
during the campaign and that campaigns are
particularly important for capturing the votes of
young adults.
PRESIDENTIAL Candidates and Political Parties
The 1999 exit poll sought to document if and how
political values were reflected in the presidential
election. Specifically, did a voter’s overall
ideological orientation affect his or her decision
on which candidate to support? Questions on
political values were considered essential to shed
light on Ukraine’s evolving political culture, in
particular the relationship between a voter’s
ideological preference and his or her support for a
candidate. Such findings could document the
relevance of political and economic values in
Ukraine’s political process.
The need to include questions on political values
was also dictated by the fact that demographic data
generated in the 1999 exit poll were considered to
be insufficient to explain the preferences of
voters. Many analytical studies have demonstrated
that political and economic opinions are less
dependent on age, education or sex than on
ideological orientation. In other words, attitudinal
factors tend to be more relevant than demographic
attributes in explaining public opinion on political
and economic issues. The importance of the
attitudinal factor applies to societies at all
levels of development and has been demonstrated in
analytical studies of opinions and attitudes in
Ukraine.
The nature of the exit poll required that
interviews be completed within no more than a few
minutes. Therefore, any measure on political values
necessitated placing no more than two short
questions. The decision was to use one question,
specifically, "For which political party did
you vote in the 1998 parliamentary election."
To assist in the recall on this question,
respondents were given a list containing the seven
leading parties of Ukraine.
A very small proportion – about one-in-ten –
said they had not voted in the 1998 election.
Additionally, one-fifth (22%) could not remember for
which party they had voted. This distribution of
responses is not surprising, since recall after one
year is problematic, especially since there were 30
political parties competing in the 1998
parliamentary election.
Kuchma’s victory reflected his appeal to voters
who support a wide range of parties in the center
and the center-right. Analysis of opinion data has
repeatedly documented that more of the public in
Ukraine espouse centrist and center-right values
than leftist or Communist ideologies. This
distribution of the electorate was also documented
in an analysis of the 1998 exit poll. Using votes
for all political parties in the 1998 election
(i.e., not using the 4% threshold that was applied
to parties), showed that half of the voters in
Ukraine (51%) supported centrist and center-right
parties, whereas 44% were on the left.
As figure 12 shows, in the first round of the
1999 presidential election, Kuchma received half or
more of the votes from supporters of the People’s
Democratic Party (People’s Dem.), the United
Social Democratic Party (Soc. Dem.), and Rukh. He
also received a plurality of votes from supporters
of Hromada and the Green party. Only small
proportions of supporters of the leftist parties
voted for Kuchma: one-in-ten of those who in 1998
voted for the Communist party and one-in-five for
the Bloc of Socialist and Peasant
Symonenko’s strength lies with the Communist
Party of Ukraine. He received 68% of the Communist
vote, the largest proportion that a candidate
received from supporters of any one party. This high
proportion of votes for Symonenko, albeit from his
own party, confirms the frequently voiced view that
the Communist party is highly organized and its
members disciplined and committed. What is equally
informative is that Symonenko rarely received more
than 10% from supporters of any of the other
parties. Generally, he received votes in single
digits and only 14% from the leftist Bloc of
Socialist and Peasant parties.
Moroz received a plurality (43%) from his party,
the Bloc of the Socialist and the Peasant parties
(figure 13). His support from the Communist party
was small, only 10% voted for him, a low level of
support that is understandable since Moroz was
competing with the head of the Communist party,
Symonenko. Moroz received better support from voters
who had supported centrist and center-right parties:
he captured 15% of the Hromada supporters and as
large a proportion of the Social Democrats. The
lowest support for Moroz came from Rukh supporters.
In the presidential election, Vitrenko
demonstrated her ability to reach individuals across
the political spectrum. She received roughly the
same level of support from the country’s two
leading parliamentary parties, the Communist party
(7%) and Rukh (8%); votes from supporters of the
other parties ranged between 9% and 16%. (Figure 13)
Marchuk’s appeal along the political spectrum
resembled that of Vitrenko. About one-in-ten in each
of the seven leading political parties voted for
Marchuk. The one exception were the 1998 voters for
the Communist party: Marchuk received only 2% of the
Communist vote.
Candidate Support by Demographic Groups
Not only did Kuchma win the election on November
14 and had a substantial lead on October 31, in both
these elections, majorities or pluralities in most
oblasts voted for him. Although his appeal varied
from region to region and within regions among
oblasts, his vote came from across the country. In
addition to the regional differences, support for
the presidential candidates differed among
demographic groups. Some of the findings in this
section report only on the standing of the five
leading candidates — Kuchma, Symonenko, Moroz,
Vitrenko and Marchuk. As previously noted the focus
on the five is because no statistically valid
conclusions can be drawn on the basis of the small
number of cases for each of the other candidates.
In the 1999 exit poll (round 1), respondents were
asked for whom they voted. To expedite the
completion of the interviews, as well as to
assist recall, respondents were shown a copy of
the ballot (see Appendix). The non response rate to
this question was very low: only 2% (or 128
individuals of the sample of 6034) refused to say or
could not remember how they voted.
There are very small gender differences in the
appeal of the five candidates (figure 15). In other
words, men were as likely as women to vote for the
leading five candidates, including for the female
candidate Vitrenko. There was one exception,
slightly more men than women voted for Marchuk.
As figure 16 shows, the appeal of the candidates
varied by place of residence. Kuchma, overall,
received a larger proportion of votes from the rural
than the urban voters. In contrast, Symonenko
attracted a larger proportion of urban than rural
voters. But, urban residents were as likely as rural
residents to vote for Moroz, Vitreko, and Marchuk.
These data suggest that Symonenko’s voters
reflected traditional supporters of communism,
namely workers in urban areas. What it also showed
was that rural residents, despite the 70 years of
collectivization, are hesitant (if not opposed) to
re-introducing communism and reverting to the road
of socialism.
As figure 17 shows above, support for Kuchma
declined with age. He was most popular among the
youngest age group (capturing 45% of this age group
in the first round). In contrast, Symonenko received
more votes from older than younger adults, with one
third of the over sixty group voting for him.
Marchuk received roughly the same proportion of
votes from across all age groups (figure 18).
Support for Moroz and Vitrenko varied by age;
Vitrenko had a slightly larger proportion of votes
from the middle age group, (40 to 49) and the appeal
of Moroz fluctuated among age groups.
Among educational groups, Kuchma had slightly
greater appeal to the more educated (figure 19). In
contrast, support for Symonenko declined with
education and was highest among the lesser educated.
The other leftist candidate, Moroz, received roughly
the same proportion of votes from all educational
groups. Vitrenko was least popular among those with
primary and university education, whereas the better
educated were more likely to vote for Marchuk than
the less educated.
The educational profile of voters for the five
leading candidates did not present any surprises.
The data did confirm the conventional wisdom that
education increases support for liberal policies,
but it also showed that not all educated voters
support a liberal candidate.
The group who voted for none of the candidates
(an option on the ballot) constitute 4% of the
voters nationwide. This group is too small (in the
sample) to analyze, except for these very broad and
tentative conclusions:
Men were as likely to fall in this groups as
women;
Younger adults (under 40 years of age) were more
likely to vote against all candidates than older
adults more of the ethnically Russians tended to
vote against all candidates than did ethnically
Ukrainians.
On balance, the findings on the profile of voters
for the leading five candidates documented the
importance and relevance of education and economic
well-being to the democratization process. The lower
support of Kuchma from the older age groups and from
the least educated is probably reflecting the fears
and the concerns of these groups. The eldest and the
least educated groups probably view with
apprehension economic and political liberalization.
Moreover, as attitudinal studies document, young
adults (under 30 years of age) are more likely to
favor economic reforms and are less concerned about
change in society than are older adults.
In addition to the differences among age and
educational groups, voter support for the
presidential candidates differed along ethnic lines
(figure 20). Symonenko appealed more to Ukraine’s
largest national minority, the ethnic Russians, than
he did to ethnic Ukrainians. In the first round,
about one-fifth of ethnic Ukrainians voted for
Symonenko, whereas among ethnic Russians over
one-third did. In contrast, Kuchma received a larger
proportion of votes from ethnic Ukrainians than from
ethnic Russians. Kuchma’s appeal to these two
ethic groups differed minimally, whereas Symonenko
was twice as likely to get a vote from the
ethnically Russian group than from the ethnically
Ukrainian group (compare: Kuchma, 39% ethnic
Ukrainian to 32% of the ethnic Russian voters;
Symonenko, 19% ethnic Ukrainian to 36% ethnic
Russian).
Vitrenko was the one candidate that drew roughly
the same proportion of voters across all ethnic
groups. Moroz and Marchuk, however, proportionately
had slightly more votes from ethnic Ukrainians than
from ethnic Russians. Figure 21 provides an overview
of the votes by ethnicity, including the
comparatively larger support of Symonenko among
ethnic Russians.
As already discussed, on November 14, Kuchma
received a majority or a plurality of votes in most
oblasts. His appeal varied within and across
regions.
In the first round, Kuchma had the largest number
of votes in 15 of the country’s 24 oblasts and in
Crimea. He had a particularly strong showing in the
western region, where he captured from three-fourths
to a half of the votes (highest in the Ternopil
oblast, 72%). In the other oblasts where he led,
Kuchma received at least one-fourth, but generally
around one-third of the vote. He had lowest
proportions of votes in Vinnytsia (18%) and Poltava
(20%), the two oblasts captured by Moroz.
Symonenko led in 6 oblasts in the first round.
His strength was in the southern and eastern regions
and he had lowest standing in the western region and
in Kyiv, both the city and the oblast.
The two candidates who faced each other in the
second round, Kuchma and Symonenko, did not capture
three oblasts in the first round: Vinnytsia, Poltava,
and Sumy. Moroz led in two oblasts (Vinnytsia and
Poltava). Vitrenko had roughly similar proportions
of votes in all oblasts, with a high in the East
Central and a low in the western region. Marchuk,
however, had a slightly better showing in the
western region and in Kyiv, the city and the oblast,
than in other parts of Ukraine.
The first round of the 1999 election showed the
two leading leftist candidates, Moroz and Symonenko,
have somewhat different constituencies and drew
their strength from different regions. Moroz was
stronger in the west central regions, whereas
Symonenko’s main support came from the eastern and
southern regions.
Comparing the votes by oblast for the first and
the second round, what was striking was the
proportion of votes that Kuchma received in the
western oblasts. Kuchma received an exceptionally
high proportion of votes in the western region, near
unanimity (over 90%) in Ivano-Frankivsk, Ternopil,
and Lviv. These high levels of votes for Kuchma
undoubtedly contributed to his winning the election,
although these votes alone would not have given him
the election.
In the second round, Kuchma received a majority
of votes in two oblasts that in the first round gave
the lead to Symonenko: Donetsk, where Kuchma won by
a sizeable margin (Kuchma 52.9% to Symonenko
41.23%); Kharkiv, where Kuchma led by a very few
votes (Kuchma 46.46% to Symonenko 46.46%). Kuchma
also received more votes than Symonenko in Sumy, the
oblast that in the first round gave Vitrenko the
lead. (For votes by oblasts, see Appendix.)
In the second round, Symonenko beat Kuchma in 9
oblasts, in some by a definite majority and in
others by a very small margin. In the second round,
in addition to losing Donetsk and Kharkiv (where he
led in the first round), Symonenko received a
majority of votes in Vinnytsia and Poltava, the
strongholds of Moroz. The western oblasts gave only
a handful of votes to Symonenko (5% in the Lviv,
Ivano-Frankivsk, and Ternopil oblasts).
Profile of Voters in 1999 Election
A political culture in a liberal democracy
depends on broad participation of its citizen in the
political process. In the 1999 election in Ukraine:
Were men and women as likely to vote? Was voter
turnout comparable among age groups? Were cities
under or over represented in the election? An
analysis of the 1999 poll offers a detailed look at
who came out to vote in the presidential election
As figure 22 shows, men were more likely than
women to vote in the 1999 election. Also, rural
areas had a slightly higher turnout than urban
centers. Voter turnout seems to have been inversely
related to city size, meaning that voters in small
cities turned out in larger proportions than did
voters in large cities. Illustrative of this pattern
is the electorate in Kyiv. For both the first and
the second round voter turnout was higher in the
oblast than in the city (compare: first round 61.61%
city to 70% oblast; second round 65.94% city to
69.11% oblast).
Voter turnout increased with age (figure 23) and
was below the national average for the 18-39 age
group. Turnout was lowest among those under 30 years
of age, increased slightly among those 30 to 39,
notably among the 40-49 age group and skyrocketed
among the 50-59 age group. Largest participation in
the election was by voters over 50 years of age.
This age group is least supportive of economic
liberalization and the emergence of a market system.
Understandably so, since many are pensioners (or
eligible for pensions) and look at the economic
changes with some trepidation. They are the least
likely to benefit from the advantages that economic
and political liberalization bring and the most
likely to be adversely affected by the changes in
society.
Voter turnout in the first round varied among
ethnic groups (figure 24). Among ethnic Russians,
voter turnout was lower than the national figure and
below that of ethnic Ukrainians. Turnout among other
national minorities was very low, less than half of
this group voted in the first round.
The demographic profile of voters in the 1999
first round presidential election was very similar
to that of the 1998 parliamentary election. In both
elections, men were more likely to vote than women,
turnout increased with age, and it was below the
national average for minorities, including ethnic
Russians. However, in 1999 a smaller number of
ethnic Russians came out to vote than did in the
1998 parliamentary election.
In the 1999 election, a larger proportion of
urban residents voted than did in 1998. In the 1998
parliamentary election, 67 percent of the urban and
81 percent of the rural population voted, while in
the first round of 1999, 66 percent of the urban and
69 percent of the rural residents did. This shift
was maintained into the second round. Illustrative
of the narrowing gap in voter turnout of urban and
rural residents are Kyiv and Crimea:
In the second round, about two-thirds voted in
the city and the oblast (65.94% city, 69.11%
oblast), while in 1998 the difference between urban
and rural turnout was 13 points (59% city, 72%
oblast).
The shift in Crimea is even more dramatic. In the
1999 second round, voter turnout in the capital,
Simferopol, and the autonomous republic differed
minimally (61.54% Simferopol, 60.21% autonomous
republic), while in the 1998 parliamentary election
the difference was large (51% Simferopol, 65%
republic).
The demographic profile of voters in the first
round, overall, applies to that in the second round.
The increase in voter turnout nationwide was not
that substantial to affect the overall profile of
voters. Moreover, the differences in voter turnout
among urban and rural residents will minimally
affect the voter composition in terms of gender, but
may have a slight impact on age. Increased voting
among urban residents in the second round probably
was due to an increase of voting among the younger
adults, those under 40 years of age and especially
among the youngest, those 18-29.
Voter Turnout
1999 & 1998 Elections
On November 14, 1999, Ukraine’s electorate
turned out in vast numbers: 74.9% of the voting age
population (28,212,484) participated, many more than
did in the first round (67.4%) or in the 1998
Parliamentary election (71.6%).
As figure 25 shows, turnout in all three
elections varied significantly from oblast to
oblast. What stands out in the second round is that
in each oblast majorities voted, ranging from around
60 percent to over 80 percent. It was highest in the
west and central areas and decreased as one moved
east and south. In the west and the east-central
regions at least 7 out of 10 voters participated in
the second round, whereas in the southern and the
eastern regions about 6 in 10 did, except for
Donetsk (75%) and Luhansk (69%).
In the second round presidential election, the
highest turnout was the Ternopil (88.2%), the
Ivano-Frankivsk (86.6%) and the Lviv (85.6%) oblast,
lowest turnout was in the Odessa oblast (59.5%) and
in Crimea (61.54%). These same areas recorded the
highest and the lowest turnout in the first round:
79.0% in the Ternopil and 55.4% in the Odessa
oblast.
Among the small proportion of citizens who live
and vote outside Ukraine, voter participation was
far below the national average. However, it also
increased from the first to the second round, from
19.8% to 24.8%. Most voted for Kuchma by a margin of
more than five to one in the second round (79.7% for
Kuchma to 15% for Symonenko) and by a smaller margin
in the first round (Kuchma 57.1%, Marchuk 14.6%,
Symonenko 5.1%).
The pattern of higher voter turnout in the west
and lower turnout in the south and east was typical
of all three elections held under the new
constitution, the 1998 parliamentary, and the 1999
first and second round presidential elections.
As figure 26 shows, compared with the 1998
parliamentary election, voter turnout decreased in
the first round and swung back in the second round,
exceeding the turnout in the 1998 election. These
national shifts obscure the notable differences
between regions and the even more dramatic changes
in some oblasts. Voter turnout remained relatively
constant in the Kyiv region, in the city as well as
the oblast, with a slight decline from the high
turnout recorded in the oblast in 1998 (see Figure
25 & 26).
Comparing the three elections, the western
oblasts had the largest decline of voters in the
first round and dramatic increases in the second
round. Two oblasts in the west recorded the largest
decline nationwide: the Zakarpattia (-11.3) and
Chernivtsi (-9.3) oblasts. In the Rivne (-7.69),
Ternopil (-5.5) and Ivano-Frankivsk (-5.4) oblats,
the decline was not as high, but still well above
the national average. Only in the Volyn oblast
(-3.3) turnout declined less than the national
average and the Lviv oblast turnout increased
slightly (+1.95). In the western oblasts, shifts in
voter turnout for the second round were as dramatic,
with large increases of turnout in Zakarpattia, Lviv
and Ivano-Frankivsk (figure 25).
Only one eastern oblast recorded as dramatic
shifts. In the Donetsk oblast the increase in voter
turnout in the second round was as high as in any
oblast (+13.9), whereas in the first round the
oblast recorded a slight increase. In two eastern
oblasts, Kharkiv and Luhansk, turnout remained
relatively constant (with a very small decline in
the first round and a small increases in the second
round).
In the southern and west-central regions, shifts
in voter turnout in the oblasts were roughly
similar. A decline at about the national average in
the first round and a slight
increase in the second round. The two exceptions
to this pattern were the oblasts of Odessa and
Crimea. In the Odessa oblast, the decline in voter
turnout for the first round was as great as in any
oblast of Ukraine (minus 11.68). The Autonomous
Republic of Crimea also had an above average decline
in the first round (minus 8.48), and a small
increase in the second round.
The decline in voter turnout in the first round
(from the level in 1998), cannot be attributed to
such external forces as the weather, as it was a
pleasant and a sunny day. The upswing of the votes
in the second round suggests that the public may
have understood the importance of the contest. They
may have seen the need to take a position as to the
overall direction of the country and to decide
whether they want to be governed by incumbent
president Kuchma, a declared centrist, espousing
economic and political liberalization, and committed
to maintaining Ukraine’s independence and its
relationship with the west. Or, do they want to be
ruled by Petro Symonenko, a communist who pledged to
return Ukraine to the road of socialism, reestablish
the social order of communism, and realign Ukraine
with Belarus and Russia.
Figure 27 summarizes the swing in voter turnout,
using the 1998 parliamentary election as the base.
The changes in the first and second round are
recorded for each of the oblasts.
Methodology
1999 & 1998 Exit Polls
The 1999 Exit Poll in Ukraine was a collegial
effort of Ukrainian and American professionals.
Funding for the exit poll was provided by a grant
from the Charles Stewart Mott Foundation, an
independent private organization.
The overall management of the exit poll project
was the responsibility of Ilko Kucheriv and the
Democratic Initiatives Foundation (Kyiv).
The sample and the selection of respondents was
designed by Steven Wagner, President of QEV
Analytics, a Washington DC polling firm. The
questionnaire, designed by QEV Analytics, was
finalized in consultation with the Democratic
Initiatives Foundation.
Fieldwork was a collaborative effort of Ukraine's
three leading polling firms – KIIS (the Kiev
International Institute of Sociology), SOCIS, and
UMS (the Ukrainian Monitoring Service) of the
Ukrainian Institute of Social Studies. Dr. Vladimir
Paniotto of KIIS coordinated the work of the three
firms; Dr. Mykola Churilov led the work at SOCIS and
Dr. Olga Balakireva and Oleksander Yaremenko at UMS.
The questionnaire for the 1999 exit poll
contained 10 questions. Six measured opinions on the
election, specifically: for which candidate
individual voted; when the choice on the candidate
was made; if he/she would vote in a second tour and
for whom -- Kuchma, Symonenko, or Vitrenko; and for
which political party did individual vote in the
1998 parliamentary election. This last question was
to empirically establish what is the relationship
between political parties and candidates for
president. In other words, who was the presidential
choice of those who in 1998 supported a centrist
party and of those who voted for the communists?
Four questions recorded personal attributes: sex,
age, level of education, and ethnic identity.
To maintain full documentation on the
interviewing process and the sample, interviewers
recorded the number of the polling place, the
oblast, and the size of the settlement (city size
and rural settlement). To expedite interviewing,
when asked for who they voted for in the
presidential election, or the parliamentary election
respondents were given a copy of the ballot. For the
presidential, the list contained 15 candidates, with
two names crossed-off since the candidates had
withdrawn from the race; the last entry on the
ballot was the negative vote -- "do not support
any candidate for president of Ukraine." For
the question on the 1998 the parliamentary election,
respondents were given a list of the leading 7
parties (the Communist, the People’s Rukh, the
People’s Democratic, Bloc of Ukrainian Socialist
and Ukrainian Peasant Parties, the Social
Democratic, Hromada, and the Green Party). (See
pages 26 to 27 for the English text of the exit poll
question)
The sample design, developed by Mr. Steven
Wagner, was based on the voter turnout in the 1998
parliamentary election. For each oblast and for
Crimea, interviews were allocated proportional to
the number of votes cast in the 1998 election. The
voting places for the interviews were randomly
selected, except for rural areas where travel to and
from the sampling point was a factor. To retain the
principle of a random sample, at each voting place,
interviewers selected every n-th person. The three
firms who were selected to conduct the fieldwork –
KIIS, SOCIS, and UMS, implemented the sample design.
All interviews for the exit poll were done on
October 31, 1999, the day of Ukraine’s
presidential election. When polling places opened,
interviewers arrived at 300 randomly selected
polling places, scattered throughout Ukraine, in all
oblasts and in Crimea. Professional interviewers
randomly selected 6,034 voters as they exited the
polling place. This sample of 6,034 voters
represents individuals who came out to vote in the
first round of the presidential election.
Interviewing proceeded without incidents and
voters willingly and openly responded to the
questions. The predictive value of an exit poll –
its success -- depends on the truthfulness of the
answers. The accuracy of predicting election
outcomes of the 1999 exit poll, as well as that of
1998, attests to the open and free atmosphere in
Ukraine and the absence of fear from publicly
voicing political choices.
On election day, the answers of the sample of
6,034 voters were delivered to Kyiv via voice by
telephone. This was the only available electronic
link between Kyiv and the field and restricted the
amount of information that could be delivered on
election day from the field to Kyiv. Therefore,
field managers or interviewers tabulated the
responses and reported aggregate figures on select
items in the questionnaire. Results were phoned into
Kyiv twice during the day – in mid afternoon and
immediately after the polls closed. The use of three
polling firms offered some security -- if the
voice-telephone network of one firm failed, data
would be coming in from the other firms. The
telephone network did work and the aggregated
figures were reported in a timely fashion.
In Kyiv, data were received, recorded, and
entered into a computer by each of the three firms
responsible for fieldwork – KIIS, SOCIS, and UMS.
QEV Analytics aggregated the data for Ukraine and by
the six geographic regions. The division of Ukraine
into six regions reflects the analytical studies of
KIIS, the Kyiv International Institute of Sociology.
Although the six regional groups are geographically
contiguous, they were formed on the basis of
political attitudes and behavior.
On election night, only two questions from the
exit poll were released — the number of votes for
each of the candidates and the intention to vote in
the second tour. Two factors account for releasing
only this limited amount of information: the lack of
a telecommunications network (as discussed above)
and the legal constraints. The laws of Ukraine
forbid the release of any data predicting a vote
fifteen days before an election. Therefore, the
questions on voting in the second round were not be
disseminated on election day.
When polls closed on the evening of October 31,
preliminary results of the exit poll were released,
reporting that incumbent President Kuchma had a
clear lead, but insufficient to win and would face
Symomenko in the second round. At midnight on
election day, final results from the exit poll were
released, providing a detailed breakdown of votes
for all thirteen candidates and reporting the votes
by six geographic regions, as well as for the total
of Ukraine. QEV Analytics analysts compiled the
incoming data and drafted the press releases, which
the Democratic Initiatives Foundation disseminated
electronically to domestic and international media.
The day after the election, questionnaires were
delivered from the field to the Kyiv offices of the
three firms doing the fieldwork. The responses were
coded, entered into a computer, and a data file
created. QEV Analytics used the data file to prepare
this report.
To ensure that the exit poll sample reflects as
accurately as possible the profile of the voting
public, data were weighted, and, thus, removing
fieldwork biases, such as respondent selection,
non-completion of interviews, refusals, and the
like. Weights were developed separately for six
regions – Kyiv city and oblast, the western,
west-central, central, east central and southern.
These regions, developed by KIIS, are based on
attitudinal survey studies and group oblasts with
comparable political attitudes and behavior. Weights
were calculated on the basis of official election
results using the total number of votes cast in each
region and the number of votes cast for the top five
contenders (Kuchma, Symonenko, Vitrenko, Moroz, and
Marchuk), others, and those who were against all
candidates. Weighting minimally affected the
results, as would be expected, since the exit poll
accurately predicted the vote. However, weighting
ensured the representativeness of the samples and
that the finding – the responses of the samples
(of the 6034 voters in 1999 and 10000 voters in
1998) could confidently be projected unto all of the
electorate who participated in the electorate.
The 1998 exit poll, the fist ever in Ukraine, was
fielded on March 29, the day of Ukraine’s
parliamentary election. On that day, at 400 randomly
selected polling places interviews were conducted
with a sample of 10,000 voters as they were leaving
the voting place. Respondents were asked for which
political party they voted, when they made the
choice of the party, if the election was fair and
honest, and what expectations they had for the
immediate future. To expedite interviewing,
respondents were given a copy of the ballot listing
thirty political parties and the option of not
supporting any party.
In 1998, the exit poll was the main feature of
the "Election Night Show" of Studio 1+1.
The show was an innovation for Ukraine and the
managers took a risk featuring the exit poll.
Although political polls are a part of Ukraine’s
civic culture since its independence, interviews
conducted right outside the polling place was a new
experience in 1998 for voters as well as election
officials. In planning the 1998 poll, there was
concern that voters may be reluctant to be
interviewed and that election officials may
interfere in such activity close to the polling
place. But fieldwork proceeded without incident,
both in 1998 and 1999. When plans were made by
Studio 1+1 for the 1998 "Election Night
Show," there was no assurance that the exit
poll would generate predictive data. The success of
Studio 1+1 election night coverage created broad
interest in the 1999 exit poll. Results from the
1999 poll were made available electronically to all
news media in Ukraine and the findings were
discussed on all of Ukraine’s leading television
programs.
The success of the 1999 and 1998 exit polls are a
reflection of the openness of the environment in
Ukraine. When planning the first ever 1998 exit
poll, there was concern about the feasibility of
conducting the interviews. Although since
independence polling has been an integral part of
Ukraine’s civic culture, the conduct of interviews
outside the polling place was a new experience for
the voters as well as election officials. The 1998
and 1999 exit poll interviews were completed without
incidents.
Equally important to the success of an exit poll
is a dynamic and free communications environment.
The news media accepted alternate information
sources, a nongovernmental entity (the Democratic
Initiatives Foundation), as well as the private
sector. The media introduced innovative ways to
reach audiences and cover political events.
Appendix