Treacherous business of primary polling
By MICHAEL McKEON
Primary polling is a particularly treacherous business. By the time this column is published, the March 20, 1990, Illinois primary elections will be over, and columnists and reporters probably will be telling how pollsters were fooled in one or more of the races. Some candidates who had leads in the polls will have lost, and some candidates who were thought to be sure losers will have "come from behind" to score upsets.
One problem with primary polling dominates all others: interviewing people who actually will vote in the election. If people who wind up voting in a primary are not included in poll interviews, the chances of attaining correct results diminish. Two variables make it difficult to guarantee that pollsters are interviewing voters from all the groups who will vote in the primary. One is turnout; the other is core voters. They are interrelated.
How many people will vote in the election, and in what areas are they coming out to vote? What core groups are coming out to vote? When it comes to primaries, both Democrats and Republicans have a core of voters. But who else will be drawn to one or the other party primaries? Will people switch parties? Consider how crucial these variables can be in the results from the Republican primary for governor and the Democratic primary for Cook County board president.
In the Republican primary for governor, Secy. of State Jim Edgar painted his main challenger, Steven Baer, as a fringe candidate with little support among the mainstream Republicans who will vote in the Republican primary. Two weeks before the primary, most political experts predicted Edgar would get at least 85 percent of the vote. The final result depended on turnout and core group voters.
Two issues have separated the candidates: taxes and abortion. Baer is anti-tax and pro-life; Edgar is pro-choice and Baer painted Edgar pro-tax because he supports making the temporary increase in the state income tax permanent. Probably any poll done for this race attempted to track the views on these issues of previous Republican primary voters. Turnout in off-year Republican primaries in Illinois is small, usually drawing from 500,000 to 600,000 voters. If Baer attracted pro-life supporters and anti-tax individuals who don't usually vote in the Republican primary, polls will have underestimated Baer's strength. Conversely, if Baer's pro-life position attracted pro-choice people to vote in the Republican primary election where they usually don't vote, then Edgar's strength will have been underestimated.
What could have played havoc with predictions in this race is if the traditional Republican voters who support Edgar stayed home because they did not see Baer as a challenge and had no other Republican race to attract them to vote in the primary. This would considerably have shrunk Republican turnout, making the effect much greater of any "new" single-issue voters casting ballots in the Republican primary. This is probably the reason the Edgar campaign was spending about $700,000 on television to make sure Edgar supporters would go to the polls.
It was forecast that only about 7 percent of the Illinois adult population would vote in the Republican primary election. That is about the same percentage of the adult population in suburban Cook County predicted to vote in the Democratic primary election for Cook County board president. While most Cook County Democratic primary voters will come from Chicago, those suburban votes when considered with other turnout factors will have been crucial in that four-way contest between Richard
36/ApriI 1990/Illinois Issues
Phelan, an attorney and challenger to the regular organization; Stanley Kusper, well-known as the incumbent Cook Country clerk; Eugene Pincham, a black and former judge; and Ted Lechowicz, a state senator endorsed by the regular pary organization.
Both Kusper and Phelan considered the suburban area as key to being nominated. Man, who had no name recognition then the race began, was trying to reach this small number of voters with television ads. By the time the race is over, he will have spent $1.5 million. His plan was not only to convince people who usually vote in the Democratic primary to support him but to bring in new voters with his reform messages on television. If he was successful, all the polling done early would not have reflected it because the full effect of such television messages comes in the final weeks and days before the election.
Kusper counted on his core: people who have supported him in the past to come out again to support him. Some of his core support in the past has come from traditional Republican voters. If in the final days before the election they thought Kusper was in trouble and needed their vote, did they decide to take a Democratic ballot in order to vote for him?
Judge Pincham was hoping to energize the Harold Washington Coalition in the final days of the campaign. Crucial to him was a higher than expected turnout in black and lakefront areas of the city. Sen. Lechowicz, on the other hand, who was trailing in the polls, was hoping for a very low turnout. Assuming that whatever strength the regular organization still had would dominate that low turnout, he could have come from behind to win.
By now you see the problems in polling primary races. Head to head numbers between candidates can be very misleading indications of what is going on in a primary race. In-depth polling data can generally discover subtle indications and nuances that give political professionals some hint that a trend is beginning to take shape among voters. If discovered too late in a primary campaign, a trend, for example, showing greater turnout — or less - may decide an election dramatically differently than pollsters predicted. By now some subtle trend may have taken its toll.
Michael McKeon is head of McKeon and Associates, a national polling organization.
April 1990/Illinois Issues/37