By NATE SILVER
Jerry Brown is now a 75 percent favorite to become the next governor of California, according to the FiveThirtyEight forecasting model, after accusations circulated widely that his Republican opponent, Meg Whitman, knowingly employed an illegal immigrant as her housekeeper. Last week, Mr. Brown was a 60 percent favorite, and two weeks ago, he was the underdog in the race.
Although several polls released in the past week show Mr. Brown, a Democrat and former governor, with a small lead, it is unclear how much of the change in his standing is attributable to the reports about Ms. Whitman and the housekeeper, Nicandra Díaz Santillán. Ms. Whitman acknowledged hiring Ms. Díaz Santillán, but she says she dismissed her as soon as her immigration status was disclosed. A poll by Rasmussen Reports registered no change in Ms. Whitman’s unfavorability rating — 50 percent, unchanged from two weeks earlier — even as it found Mr. Brown gaining ground.
Any impact may be hard to measure because it might be felt most among Hispanic voters, and the polls were already in disagreement about how many of their votes Ms. Whitman might get. About 40 percent of Hispanic voters in California prefer to speak Spanish to pollsters, according to the research firm Latino Decisions, and these voters may be underrepresented in some polls.
Still, the allegations are obviously not helpful to Ms. Whitman, whose campaign has reacted with a certain lack of dexterity — with Ms. Whitman, for instance, having volunteered to take a polygraph test to rebut them. Such distractions may be relatively more difficult for a candidate like Ms. Whitman, who is running her first campaign for office, and who is used to writing her own script as the former chief executive of eBay.
History teaches us, nevertheless, that shifts in the polling sometimes reverse themselves — indeed, this happens at least as often as a candidate continues to gain “momentum” from them. So caution is warranted until we get a better sense of whether Mr. Brown’s lead can be maintained.
Several other gubernatorial races have had polling shift in recent days.
In Illinois, several polls show improvement for the Democratic incumbent, Pat Quinn. But the model is being cautious here. Only one poll, from Suffolk University, shows Mr. Quinn with a lead outside the margin of error, while until recently his Republican opponent, Bill Brady, had leads of around 10 points in most surveys. The model, therefore, still gives Mr. Brady a 79 percent chance of winning, although this is down from 91 percent last week.
If the model seems too slow to catch up to Mr. Quinn’s apparent surge in the polls, it’s worth bearing in mind another Midwestern state, Ohio. Some polling in Ohio last week had suggested that the Democratic incumbent, Ted Strickland, was pulling into a rough tie with his Republican opponent, John Kasich, spontaneously overcoming what had been about a 10-point deficit.
The model, however, was cautious about taking those numbers quite at face value. And this week, that looks prudent, since three new polls have come out showing Mr. Kasich still holding an advantage of 6 to 9 points. Obviously, Mr. Strickland could still win — and Mr. Quinn’s chances must be considered improved. But the lesson is not to jump to conclusions. Especially at this stage of the campaign, things like changes in advertising spending patterns can touch off short-lived bounces in the polls.
A race that shows a more unambiguous shift toward the Democrat is the one in Maryland, which had received fairly scant polling. Both a Washington Post poll and a Rasmussen Reports poll now show the Democratic incumbent, Martin O’Malley, with a solid lead there. The model now makes him an 89 percent favorite, up from 67 percent last week.
The Republican candidate, Rick Scott, has gained ground in Florida. The polling had been mixed there — but each of the five polls released in the past week give Mr. Scott a small advantage over the Democrat, Alex Sink. The model assigns Mr. Scott a 60 percent chance of winning, after he was a slight underdog in last week’s forecast; it is plausible that Mr. Scott has been assisted by the Senate race in the state, where the Republican, Marco Rubio, has run a strong campaign and has emerged with a clear lead over two opponents.
Another Republican with an improved forecast is the party’s candidate in New Mexico, Susana Martinez. It is not necessarily that Ms. Martinez has gained any more ground over her Democratic opponent, Lt. Gov. Diane Denish; she has appeared to hold the lead for a while. But with three polls in the past week showing her with a clear advantage — New Mexico had received limited polling before — the model can now be more confident in its forecast.
A few other states are worth commenting on:
* A poll from the University of New Hampshire contradicts two others that had shown a close race there. It gives the Democratic incumbent, John Lynch, a 17-point lead, although with a high number of undecideds. We had not been persuaded that Mr. Lynch was in all that much jeopardy; he has a strong approval rating, and one of the polling firms that was showing a close race, American Research Group, has a dubious track record. But the model now rates his chances of winning back above 90 percent.
* In Connecticut, on the other hand, we were perhaps a bit too confident in assigning the race to the Democratic candidate, Dan Malloy. While Mr. Malloy remains the favorite, fresh polling this week shows him with less than 50 percent of the vote, giving the Republican, Tom Foley, an opportunity to win by closing the race strongly.
* A poll from a Maine firm, Critical Insights, shows the Democrat, Libby Mitchell, with a small plurality of the vote after having trailed the Republican Paul LePage by double digits in their previous survey. But caution here: Mr. LePage has led in several polls, including one released by Ms. Mitchell’s campaign. Maine remains among the more difficult states to forecast, with no fewer than three independent candidates in the race and a high proportion of undecided voters.
* The notion that Carl P. Paladino had a credible chance to upend Andrew Cuomo in New York seems more dubious today, with three new polls showing Mr. Cuomo, the Democrat, with leads of 15 to 24 points. The model now makes Mr. Paladino, whose unfavorability ratings have increased sharply, about a 100-to-1 underdog. You could find better risks to take at that price, like the Buffalo Bills winning the Super Bowl.
* A Texas survey from Public Strategies gives the Republican incumbent, Rick Perry, a 14-point lead, his largest so far. That improves his winning chances to around 90 percent in the model.
* Finally, in Rhode Island, there’s now a clearer consensus that the favorite is the Democrat, Frank T. Caprio, and not the independent, Lincoln Chafee, a former senator. But races with three or more significant candidates are tricky to predict, and it’s plausible that Mr. Chafee could gain ground if the Republican, John Robitaille, who had just 14 percent of the vote in one survey, begins to look nonviable to voters and some of his support migrates to Mr. Chafee.