Poll Analysis: Recall Race Tightens

Share via
Times Poll Staff

The recall election is just over three weeks away and campaigning for the first part of the ballot — to recall Gov. Davis — is now starting to rev up. The “NO” on recall campaign has Dianne Feinstein commercials airing throughout the state in which she talks about the harm it would do to California and its residents. Gray Davis started airing his commercials on Tuesday (Sept. 9) against the recall as well. So far, there have been no specific commercials airing just for the vote “YES” on recall, but individual candidates have been dealing with the recall in their personal messages.

There is unprecedented interest in the recall election — not seen in most other elections. More than six in 10 likely voters say they are “very” interested in following the recall election, compared to just 38% of voters prior to the gubernatorial election in 2002. Who will turn out to vote is the unknown factor in determining whether the recall will pass and if so, who will be the next governor. It is like playing a chess game — you cannot just think of the current situation, but you have to anticipate who will turn out to vote as it gets closer to the election.

According to a new Los Angeles Times Poll, the vote on the recall measure has tightened to within three points. Fifty percent of likely voters would vote to recall Governor Davis, while 47% would vote against the measure. Just three percent are undecided. In a Times’ August poll, there was a five point difference — 50% of likely voters said they would vote to recall the governor, while 45% would vote against it. The survey shows the second part of the ballot, has Lt. Governor Cruz Bustamante and actor/businessman Arnold Schwarzenegger to within 5 points of each other with state Senator Tom McClintock in third place. Among likely voters, 30% would vote for Bustamante, 25% for Schwarzenegger and 18% for McClintock. (Ueberroth received 8% of the vote, but he dropped out of the race while the poll was in the field.) Three percent of likely voters would vote for Arianna Huffington, 2% for Peter Camejo, while 8% would not vote for a candidate and 5% are undecided. As will be discussed later, each candidate has problems that they will have to overcome in order to win on Oct. 7.


Recall of Governor Davis

In order to understand the chances of the recall measure passing or failing, one must identify the voters for each side. For instance, Gov. Davis must look to his core group of voters to help stave off his opponents — union members, minorities, women, Democrats and liberals. On the other hand, the recall measure has to win large percentages of Republicans, conservatives, white and younger voters in order for the ballot measure to pass.

Since the energy crisis, the governor’s job approval rating and favorability ratings have been dismal. His job rating fell lower than any governor had received since the Times asked the question. Although the poll shows the same negative opinions that voters had of Davis in previous polls, this poll shows that he has stopped the bleeding. In the August Times poll, 26% of likely voters approved of the way Davis handled his job as governor, while 72% disapproved. In the current poll, 34% of likely voters rate his job performance positively, while 63% still have negative feelings.

In the August poll, 33% of likely voters had a favorable impression of the governor, while 65% had unfavorable opinions. In the current poll, nearly two out of five likely voters have positive feelings, while 60% still have negative feelings.


Party ID and political ideology: Gov. Davis has to shore up more of the Democrats and especially moderate Democrats. This poll shows that almost four out of five Democrats (79%) likely to vote will not support the recall, while almost a fifth will. Surprisingly, three in 10 moderate Democratic likely voters say they will vote yes to recall the governor. Their share of voting yes on the recall is up 10 points since last month’s poll. On the other hand, virtually all liberal Democratic likely voters say they will vote no on the recall (one in 10 will vote to recall Davis). In reviewing L.A. Times’ exit polls from past elections, Democrats supported the Democratic candidate much more heartily than they are now supporting Davis in the recall election, with one exception. In the 2002 gubernatorial election, Davis captured 81% of the Democratic voters and he barely beat Republican Bill Simon, 47%–42%. In 1998, 90% of the Democrats voted for the governor and he handily beat Republican Dan Lungren by 20 points. In the 1996 and 2000 presidential elections, former President Bill Clinton and Democratic candidate Al Gore each received roughly 90% of the Democratic voters as well and they each won solidly against their Republican opponents. The smaller share of Democrats voting against the recall would suggest a problem spot for the governor.

Conservative Republicans who are likely to vote are overwhelmingly for the recall (90%), as well as more than two-thirds of moderate Republicans. Almost three in 10 moderate Republicans say they will vote against the recall. However this is an eight point decrease from the Poll’s findings last month.

Independents or declined-to-state likely voters (including minor party voters) are splitting their vote on the recall, 50% for the recall and 47% against it. This is a substantial shift since the August poll when independents were voting 64% to 36% in favor the recall.


About as many likely voters who describe themselves as conservative are voting yes to recall Gov. Davis (85%) as likely voters who describe themselves as liberal are voting against the recall (82%). Self-described moderates who are likely to vote are more inclined to vote against the recall (53%) than for it (43%).

Gender, race and age: There is a gender gap on the first part of the recall ballot. Nearly three-fifths of likely male voters will vote to recall Gray Davis, while 54% of likely female voters will vote against the recall. More importantly, Davis has to woo back Democratic men. A quarter of Democratic men will vote yes on the recall vs. 14% for Democratic women.

White voters make up about 70% of likely voters in the poll and this group is voting to recall the governor. Fifty-four percent of white likely voters say they will vote yes to recall Gray Davis, while 43% will vote no. Black likely voters are overwhelmingly against the recall (subgroup too small to break out). In some bad news for Davis — 53% of Latino likely voters are voting yes on the recall and 41% are voting no. This is a significant net negative change from the August poll. In that survey, 45% of Latino likely voters were against the recall and 39% were for it. Latinos make up about 11% of the likely voters in the current poll. (As an aside, in the past Latinos decide late in an election, so it would be important to look for any shifts from this group. This has been evident in voting against Propositions 187, the anti-illegal immigration initiative and 226, the affirmative action initiative.)

The younger voters, 18–44, are solidly behind the recall (63%), while likely voters between the ages 45 and 64 are against the recall (54%) and the elderly (65+) are divided (48% each for and against the recall). The 18–29 year old group is too small to give an actual percentage, but they are overwhelmingly in favor of recalling Gov. Davis and are dividing their vote among the top three candidates. (This younger group historically does not turn out in large numbers, so it will be interesting to note if their enthusiasm for Schwarzenegger will motivate them to vote. Conversely, older voters are the ones who historically vote and this may be some good news for Gov. Davis. However, in the August poll, the elderly, 65+, were against the recall (51%), while 46% were for it.)

Union household: Union voters were the meat of Gray Davis’ vote in 1998 and 2002 and he hopes they will support him in the recall election. Unions have contributed lots of money to the “No on Recall” campaign and endorsed Davis as governor. Hesitantly, they have endorsed Cruz Bustamante as a safeguard to keep a Democrat in office just in case the recall ballot wins. So, it was interesting to note that union members and/or union households are supporting the recall. The union leaders have yet to begin their push with their rank and file and these numbers may turn around. But for Davis, this must not be good news. Fifty-one percent of union households likely to vote are supporting the recall, while 45% of union households are against the recall.

California regions: As you would expect, the Bay Area and Los Angeles county voters are solidly against the recall (70%, 58% respectively), while voters in the rest of southern California and the Central Valley are heavily in favor of the recall (63%, 64% respectively). The rest of northern California voters are split — 50% in favor of the recall and 47% against.


Share of the electorate

Who is more energized to come out to vote? The current registration figures from the Secretary of State show that Democrats make up about 44% of the voters, Republicans make up about 35%, declined-to-state voters are 16% of registration and 5% of the voters belong to minor parties. The Times poll shows that Republicans are somewhat more motivated to come out to vote than Democrats. Forty-five percent of likely voters are Democrats, 16% independents and 39% are Republicans. (In the Times 2002 exit poll for governor, 46% of the voters were Democrats, 40% Republicans and 14% were independents and minor party members.) And conservatives are also more energized. In the 2002 Times gubernatorial exit poll, 35% said they were liberal, 30% moderate and 35% conservative. In the latest poll, 38% of likely voters consider themselves conservative, while 34% say they are liberal and 26% describe themselves as moderates.

Latinos are about the same share as they were in the 2002 Times exit poll. In that election, 10% of the voters were Latino. In this poll, 11% are Latino. (In the 2000 exit poll for president, 13% were Latino.) Some political consultants think the Latino vote will be closer to the 2000 presidential election. African American voters in the governor’s election in 2002 were 4% of the voters, while in this poll 8% are African American. Three-quarters of the electorate was white in the November ’02 election; in the current poll they represent about 70% of the electorate.

Davis’ job approval and favorability ratings

Gov. Davis’ job rating is still negative, but it has improved slightly. A third of likely voters approve of the way the governor is handling his job, while 63% disapprove. In the August poll, 26% approved of his job performance and 72% disapproved. Democrats and liberals give him positive marks on his job performance as do voters living in the Bay Area. However, among moderate Democrats, 54% disapprove of his handling his job as governor vs. 41% who approve. Latinos have a very negative opinion of Davis’ job performance. Just 27% of Latino likely voters approve of his job performance, while 72% disapprove. Among whites, it is 30% to 67%.

Davis’ favorable ratings have gone up slightly, although still negative. Now, almost two out of five likely voters have a favorable impression of the governor, while 60% have an unfavorable. In last month’s survey, a third had a favorable opinion, while 65% had an unfavorable one. Moderate Democrats are split — 49% each favorable and unfavorable (which may be a problem for Davis). Yet, three-quarters of liberal Democratic voters give him a positive rating. More than three-fifths of independent voters have a negative opinion of the governor. Another group of likely voters that have a positive impression of Davis is voters in the Bay Area (63%). In Los Angeles County, voters are divided — 49% favorable and 50% unfavorable.

When the poll asked which candidate has the best experience for the job, 35% of likely voters say that Davis is the one. A quarter of likely voters say McClintock is the candidate with the best experience, followed by Bustamante at 20%. Just 1% of voters say Schwarzenegger has that attribute. Along with experience for the job, 44% of likely voters also say that Davis increased money for California’s schools and deserves some of the credit for improvement in student test scores. Four out of ten voters disagreed with that assessment. But nearly two out of five voters agreed with the statement that Davis spends too much time raising campaign funds and not enough time working on the state’s problems. Democrats and liberals are split on this issue, but all other groups agree with this sentiment.

The candidates

Lt. Governor Cruz Bustamante, the only major Democrat on the second part of the recall ballot, continues to maintain a slim lead over a large field of candidates who would like to replace Gray Davis as governor should the recall succeed. Bustamante garnered 30% among likely voters compared to his closest rivals, both Republicans— actor/businessman Arnold Schwarzenegger with 25% and state Senator Tom McClintock with 18% of the vote in the latest Los Angeles Times Poll. The survey found Bustamante’s support has declined, McClintock’s has risen and Schwarzenegger’s remained virtually unchanged since a Times Poll taken three weeks ago. Bustamante’s lead has slipped from the thirteen points over Schwarzenegger found in that previous survey to a slim five points now (which is just inside the margin of error.) The lt. governor has been the subject of scrutiny in the press and targeted by his rivals on the subject of his campaign contributions from the Native American owners of gambling casinos and his support of drivers licenses for illegal immigrants.


The outcome of this race remains far from certain, of course. It has yet to enter the critical period just before the election during which the candidates and their supporters aim resources at the groups they hope will turn out and push them ahead to victory. Much depends on who actually shows up to vote on election day.

In addition, voters are still making up their minds. While most voters were able to name a favorite candidate, many said they are still considering their options. Just over three in 10 of Bustamante’s voters are still considering their vote, along with 42% of Schwarzenegger’s voters and more than half (54%) of McClintock’s. Bustamante’s support is most firm, with seven in 10 who said they will definitely vote for their candidate. Fifty-eight percent of Schwarzenegger’s voters and 46% of McClintock’s voters are sure of their support.

Arianna Huffington, running an outsider’s campaign, garnered 3% of the vote in this survey and 2% voted for Green Party candidate Peter Camejo. Former Olympics advisor and Baseball Commissioner Peter Ueberroth retired from the race on Tuesday while the survey was underway, and garnered about 8% of the vote over the entire period. This survey does not indicate to which candidates Ueberroth’s vote will go, and the former candidate has yet to endorse one of his erstwhile rivals, so that remains one of the big unanswered questions of this race.

Schwarzenegger (23%) and McClintock (20%) are seen by likely voters as able to provide the strongest leadership for the state, followed by Davis at 18% and Bustamante at 14%. Not surprisingly, since it is his issue, Schwarzenegger (42%) was seen as the candidate who can best reduce the influence of special interests in Sacramento, with Bustamante (7%) , McClintock (8%) and Davis (5%) trailing behind.

Lt. Gov. Cruz Bustamante

Up until recently (and during the entire course of this survey), Bustamante’s campaign has walked the fine line of opposition to the recall and support for his own candidacy. His campaign mantra has been “No on the recall, yes on Bustamante” and he has focused on campaigning against the recall itself which, if not passed, would make his and all other candidacies moot. About half of Democratic likely voters would vote no on the recall and yes for Bustamante when marking their ballot. The survey found little evidence that a desire to elect Bustamante has driven up the “no” vote on the recall — only five percent of likely voters say they are both voting yes on the recall and voting for Bustamante. About one in ten said they plan to vote no on the recall and not cast a vote to replace Davis at all. Voters in the survey were split (43% to 39%) over whether Bustamante is in touch with the needs of average Californians as well as whether the lt. governor has been too much a part of Gray Davis’ administration to solve the problems in Sacramento.

Bustamante’s rivals have attacked Bustamante’s acceptance of campaign contributions from Native American-owned gambling casinos, saying that these donations would present a conflict of interest if Bustamante were elected governor and thus in a position to negotiate contracts with the tribes. This line of attack may be paying off, as negative campaigns often do. The general goodwill toward Bustamante found in an LA Times survey taken just three weeks ago — as evidenced by his 48% favorable to 29% unfavorable approval rating at that time — has eroded to the point where half of likely voters now have an unfavorable view of him, while his favorable rating has fallen to 41% in the current survey. Bustamante’s campaign recently vowed to focus more on their candidate and less on the recall.


A major problem for Bustamante is the erosion of support among Latinos (65% of whom voted for Davis in 2002) and Democrats in a race in which the Latino lt. governor is the only major Democrat on the ballot. Just under half (47%) of Latinos picked Bustamante, compared to 51% in the previous poll. While over half of Latino likely voters have a positive impression of the lt. governor, a large minority — 42% — do not.

Bustamante would, if the recall passes and he garners the most votes, become the first Latino governor of California since 1875, and a quarter of Latino voters responded positively to that idea, saying they are more likely to vote for Bustamante because of it. However, almost three-fourths said that it would make no difference to their vote. Overall, one in ten likely voters statewide said they were more likely to vote for Bustamante for that reason, 4% said they were less likely, and 85% said it made no difference. Four out five Democrats overall remained unmoved by that argument, while 17% said they would be more likely to vote for him because of it.

California Democrats aren’t exactly lining up behind their party’s major candidate either. While the vast majority — just under eight in ten — of likely Democrats are against the recall, and 57% say they would vote for Bustamante (down from 65% three weeks ago), more than one in ten say they are not planning to cast a vote for a candidate at all and another four percent just aren’t sure. It is largely the liberals who are driving the vote for Bustamante among likely voters in his party, showing that there is more ground to be gained in the middle than on the left where he has been positioning himself of late.

Moderate and conservative likely Democratic voters are more lukewarm to Bustamante’s candidacy at this point than those to their left. More than two-thirds (68%) of liberal Democrats, down from 74% in the last survey and only 42% of non-liberal Democrats (down from 57% in the last survey) would vote for him. While 78% of liberal Democrats had a positive impression of the candidate compared to 10% who do not, that proportion drops to 52% among non-liberal Democrats vs. two in five who see him in a negative light.

Another area of vote slippage which Bustamante must look to reverse in the coming weeks if he is to halt his slide in the overall vote is taking place among union households. While union leaders have endorsed Bustamante and are urging their members to vote against the recall, likely voters who live in union households are split 31% for the Democrat and 32% for Schwarzenegger at this time. In the Aug. survey he led against Schwarzenegger by 20 percentage points among this group. Nearly three out of five union households voted for Davis in 2002.

Businessman Arnold Schwarzenegger

With 23% of the vote three weeks ago and 25% in this poll, Schwarzenegger’s candidacy has shifted in composition but not moved much in the overall vote. He leads among Republican likelies by 45% to McClintock’s 31% and Bustamante’s 4%, but conservatives, half of whom see the former actor as being to the ideological left of them, are nearly evenly split 38% to 39% for McClintock. Among all Republicans, Schwarzenegger has gone from 39% in last month’s poll to 45% today.


Schwarzenegger’s share of the vote among moderates, Latinos, and union households has increased as well, all groups which supported Davis in the last two elections. Schwarzenegger has been actively courting the Latino vote since the beginning of his campaign, emphasizing his immigrant status as a point of shared experience. While Latinos are more likely than other state voters to say they are turned off by Schwarzenegger’s association with former Governor Wilson and more than half said they are less likely to vote for him because of it, support for Schwarzenegger has more than doubled among Latino likely voters over the past three weeks. Almost three out of every ten Latino likely voters say they would vote for Schwarzenegger, up from 12% three weeks ago. Still, this is a potential source of trouble for him, since Pete Wilson is widely associated with Proposition 187 which is unpopular in the Latino community and elsewhere.

Overall, half of all likely voters said it made no difference to their vote to know of the candidate’s association with Pete Wilson. Those who objected did so strongly, however, with nearly three in ten saying that they were much less likely to vote for Schwarzenegger because of it, along with another 7% who said they were somewhat less likely. Thirteen percent said they were more likely to vote for the candidate because of it.

Only 5% of liberal Democrats would vote for Schwarzenegger, but he has the vote of 16% of moderates in that party. Fourteen percent of moderate Democrats said they supported Ueberroth, so again it will be interesting to see where those votes go now that the candidate has withdrawn.

The famous bodybuilder and actor has a strong following among Republican women, half of whom said they would vote for him, compared to 40% of Republican men.

Likely voters split 45% to 44% over whether Schwarzenegger is an outsider who can bring needed reform to Sacramento, but about two in three agreed that he has tried to avoid taking public positions on important issues.

Schwarzenegger has agreed to participate in only one debate during the campaign, preferring to rely on advertising and media attention. It doesn’t seem that this strategy has hurt him too badly so far. More than half of likely voters said that Schwarzenegger’s avoidance of the other debates will make no difference to them when it comes time to vote, although just over two in five said it makes them less likely to vote for him.


About a third of likely voters overall consider themselves liberals, while 38% say they are conservative and 26% see themselves as moderate. While Bustamante and McClintock are seen by approximately half of voters as being respectively to the left and to the right of them politically, Schwarzenegger seems to have positioned himself squarely in the middle. Even though Schwarzenegger has avoided taking stances on many issues so far in the campaign, only one in 10 cannot place him in the spectrum of political ideology compared to themselves. Just over three in ten likely voters say he is more liberal than they are, a third see him as more conservative and 26% say he is just the same. Seven in ten liberals see him as being to the right of them while 24% think he is either the same, or more liberal than they are. Six in 10 conservatives see him as being to the left of where they stand, while nearly a third see him as being the same or more conservative than themselves. Moderates aren’t as certain. One in five moderate likely voters think he is to their left, 25% think he is to their right, 38% see him as the same as they are and 17% just aren’t sure.

Schwarzenegger’s support is strongest in the southern California region outside Los Angeles where 31% would vote for him compared to 25% for Bustamante and 21% for McClintock; and even stronger in the Central Valley where he leads the field 39% to McClintock’s 21% and Bustamante’s 20%. If he can continue the trends of increasing support among moderates, Latinos and union households, and can increase his share of the conservative vote, the state could see this candidate moving upward in the polls.

State Senator Tom McClintock

The vote for conservative GOP state senator Tom McClintock has risen from a 12% share in the August survey to 18% today. His support is strongest among conservatives, among whom he is tied 38% to 39% with Schwarzenegger, and among conservative Republicans where he only trails the actor and businessman by 5 percentage points.

McClintock does not have the name recognition of either of the other candidates, but overall he is seen in a positive light and as people have gotten to know him, his positive rating has increased. More than two in five likely voters didn’t know enough about him to say if their opinion of him was positive or negative in the Times Poll taken three weeks ago but in the current survey that percentage dropped to 31% while his favorable rating has risen from 34% then, to 46% today. Nearly four out of five conservatives view McClintock favorably, while 87% of conservative Republicans think well of him, too. More than a third of moderate Democrats and 42% of moderate likely voters in general view him favorably. Thirteen percent of Latinos would vote for McClintock.

More than half (54%) of all likely voters agreed when read a statement that asserted McClintock is straightforward and says what he believes even if it unpopular. A plurality also said that the conservative candidate is too politically conservative to have a realistic chance of being elected. Forty-four percent agreed with that statement, while 25% disagreed.

Just under three in four self-described liberal likely voters, 50% of moderates and 26% of conservatives see McClintock as being right of them politically. Fifty-three percent of conservatives, and 56% of conservative Republicans say that he is about they same as they are in the ideological spectrum. Conservatives made up 35% of the electorate in the governor election in November 2002, according to the Los Angeles Times exit poll.


Some Republicans have put pressure on McClintock to step aside so that Schwarzenegger might take the lead over Bustamante, but the conservative candidate has been adamant that he will not drop out of the race. This survey shows that McClintock’s support grows as his name recognition has increased and he is giving the actor a run for his money among the powerful conservative base of the Republican party in this state.

How the Poll Was Conducted

The Times Poll contacted 2,249 California adults, including 1,553 registered voters and 922 voters considered likely to turn out on October 7th. The survey was conducted by telephone Sept. 6–10. An additional random sample* of the state was also conducted over the same period and added to the sample which brought the total number of Latino likely voters in the survey to 121. The entire sample of adults was weighted slightly to conform with census figures for sex, race, age and education as well as to registration figures provided by the secretary of state. The margin of sampling error for likely voters is 3 percentage points in either direction; among Latinos it is 9 percentage points. Surveys conducted during busy campaign seasons are also subject to influence by news events; for example, Peter Ueberroth retired from the race on Tues. while this survey was being conducted. Poll results can also be affected by factors such as question wording and the order in which questions are presented. Telephone numbers for Times Poll samples are randomly selected from the set of all telephone exchanges in the state, which allows all Calif. residents equal chance of being contacted without regard to listed or unlisted numbers or their presence on registered voter lists. Interviews were conducted in English and Spanish. All racial and ethnic groups are proportionally represented in this survey even when there may not be enough in the sample to be specifically mentioned.

* Field work for the random Latino sample was conducted by Datascension of Brea, Calif.