A June 2001 Denver Post voter survey confirmed a prevailing sentiment that Sen. Allard is vulnerable to a Democrat challenge. Tom Strickland, Allard’s 1996 opponent, has resurfaced as the front-runner for the Democratic nomination in 2002. When matched-up in the June survey, Allard and Strickland tied with 42 percent each.

Question: Next let me ask you about some candidates considering running for office. If the November 2002 election were held today and the candidates for U.S. Senate were Democrat Tom Strickland and Republican Wayne Allard, whom would you vote for?

The survey was conducted by Ciruli Associates, June 25-28, 2001, with 452 residents likely to vote in the 2002 general election. The U.S. Senate survey questions were sponsored and published by The Denver Post. This analysis is the sole responsibility of Ciruli Associates.

Allard: Low Identification and Low Negative

Although nearly a fifth of voters (21%) could not rate Allard when asked their impression of him, he received a low negative rating (22%) for an incumbent politician, and a high positive-to-negative favorability (2:6).

Favorability Ratings President and Colorado Political Leaders
LEADER	Favorable	Unfavorable	Don't Know	Ratio
Owens	65%	29%	6%	2:2
Allard	57%	22%	21%	2:6
Bush	57%	38%	6%	1:5
Webb	53%	28%	19%	1:9
Strickland	48%	14%	38%	3:4
Ciruli Associates, N452, June 2001

Question: As I read the following list of political leaders and candidates, please tell me your impression of them—whether you think very favorably of them, somewhat favorably, you think somewhat unfavorably or very unfavorably of them. If you aren’t familiar with them, or don’t have an opinion, just say so: U.S. Senator Wayne Allard, President George W. Bush, Former U.S. Attorney Tom Strickland, Denver Mayor Wellington Webb. [Rotate Names]

Strickland also had a low negative rating and high ratio of positive to negative, but he was unknown by 38 percent of voters. The former U.S. Attorney’s low name identification is a product of his being out of the media spotlight for more than a year, the fact that voters typically do not focus on challengers this far in advance of an election, and the large number of new voters added to the roles since 1996 when Strickland first ran against Allard. More than 8 percent of voters in this survey claimed to have moved to Colorado in the last five years.

An Expensive Sequel

Republicans have controlled the Senate seat currently held by Allard for 22 years. The long reign followed a hard-fought battle in 1978 when first-term Democrat Floyd Haskell lost to Congressman Bill Armstrong. More recently, Bill Armstrong retained the seat with 64 percent over Nancy Dick’s 35 percent in 1984, and Hank Brown easily defeated Josie Heath 56 percent to 42 percent in the 1990 election.

1996 Senate Race Allard vs. Strickland
Candidate	Votes	Percent	Expenditures
Allard	750,325	51%	$2.2 million
Strickland	677,600	46%	$2.9 million
Ciruli Associates, 2001

Strickland lost to Allard by 70,000 votes in a 1996 slugfest. Although the $5 million race was not a spending record (the Tim Wirth vs. Ken Kramer $7.5 million race in 1986 set the Colorado record), this year’s race could become a record-spending campaign if polls remain close leading up to the election. Allard now has $1 million in the bank and President Bush will headline a fund-raiser in mid-August to boost his coffers. Strickland is a prodigious fund-raiser, and both parties and the major Washington political action committees will monitor the race and provide high levels of fund-raising support.

Incumbent Has Advantage

Wayne Allard’s low-key demeanor is deceptive; he has proven a formidable opponent. From his north Front Range and rural political base he built a statewide political machine that defeated former Colorado Attorney General and now Interior Secretary Gale Norton in the 1996 primary, and then Strickland.

In the last five years Allard has gained a reputation for serving his constituents and working for Colorado projects. From the Great Sand Dunes National Monument, to Rocky Flats clean up, to transportation funding, to the Shattuck Superfund site, Allard has been active on topics and concerns in the news. Along with a capable staff, he has assembled a tough team of campaign consultants.

The general rule is that an incumbent senator will be re-elected, unless he has committed a major blunder or the political environment has become especially hostile to his party. The last Colorado incumbent senator to lose was Democrat Floyd Haskell.

The 1978 Haskell race, while unique, offers a few important lessons for Allard. Haskell, who had a similar low-key style, was swept away in the growing national anti-tax trend that presaged Ronald Reagan’s presidential victory two years later. In the June 2001 Colorado poll, 75 percent of voters agreed: “Washington politicians have done very little to help the economy or lower gas or fuel prices.” Although the economy and fuel prices may not be major issues next year, national issues are important in U.S. Senate races and can overcome good local records. If President Bush is unpopular or the country is in a funk, incumbent politicians, especially Republicans, could be in trouble.

In addition, Strickland could benefit from the current focus on such issues as the environment and health care, which typically help Democrats. In 1996, Clinton (with his ethical problems) lost Colorado to Bob Dole by 3 percentage points amid a national election focused on character issues. At the same time, Allard’s effective attack on Strickland’s character caused damage his campaign could not overcome.

Finally, this will be the “new” Strickland—fresh from a successful, if short, term as U.S. Attorney in Colorado.

Despite Strickland’s assets, the early view is that if you can’t beat a mild-mannered, good- natured yet little-known rural congressman, you’re not likely to defeat a U.S. Senate incumbent. Strickland blamed his loss in 1996 on a 106,000 Republican-registration advantage. Unfortunately for Strickland, it still exists—the latest registration now gives Republicans a 168,000-voter edge over Democrats. Advantage in this race goes to Allard, but expect an expensive war.

Methodological Note

One of the most difficult aspects of election polling is determining who is likely to vote. It is especially problematic for a poll querying voter preference for an election 16 months away (November 2002). It is expected that about 55 percent of registered voters will turn out for the 2002 general election. Census data indicates that about 75 percent of Colorado adult residents are registered to vote. Hence, only about 40 percent of people who could be contacted in a random digit dialing phone survey are likely to vote in November 2002.

Every pollster uses sampling techniques honed after years of field testing. But a recent study was conducted that provides some additional insights into targeting likely voters in phone surveys.

The PEW Research Center for the People and the Press (www.People-Press.org), the nation’s leading and best-funded think tank using survey research, issued a report at the American Association of Public Opinion Research’s May 2001 conference on polling methods that can be used to better determine likely voters. This Colorado survey adopted two of their recommendations.

After a question to determine the person’s registration, PEW recommended a follow-up question to double check the current status of registration, which is the most common error people make about their registration. Question: Are you certain you are registered to vote or is there a chance your registration has lapsed because you moved or some other reason?

The other question concerns determining future voter interest. There are several versions of questions used to probe a resident’s intention to vote in a future election. The version PEW recommended uses a scale to measure voter intention. Question: Thinking about the general election next year in November 2002 for Governor, Senator, Congress and other offices, I’d like to rate your chances of voting on a scale of 10 to 1 where 10 represents a person who will definitely vote and 1 represents a person who definitely will not vote. Where on this scale of 10 to 1 would you place yourself?

PEW research showed the application of these questions increased the statistical probability the respondent was a registered likely voter. However, the issue highlights why polls are grainy snapshots, not clear predictions.

SURVEY FACTS

Telephone survey of 452 Colorado adult registered voters likely to vote in the November 2002 general election. Selection by random digit dialing with a random sample of statewide area telephone exchanges gave all Colorado residents telephone numbers, listed and unlisted, an equal chance of being included. The most recent birthday method was used to randomly select members of the household to interview. Respondents were screened to identify likely registered voters. Statistical range of accuracy in 19 out of 20 cases is ±4.6 percentage points for a sample size of 452. Sample tolerances for subgroups are larger. For example, the confidence interval for a subgroup of 250 respondents is ±6.2 percentage points. Due to rounding, not all totals equal 100 percent. Survey results can be affected by other factors such as question wording and order. For additional copies of this report contact Ciruli Associates.

.

Permission to quote or reprint is granted provided the source, Ciruli Associates, is credited. Ciruli Associates • 1129 1/2 Pennsylvania St. • Denver, CO 80203 • PH (303) 399-3173 • FAX (303) 399-3147. For additional information on Colorado politics, log on to Ciruli Associates’ website: www.ciruli.com.

[top] [poll archive] [home]