For Donald Trump, New York’s big primary win last night was no fluke — and neither was the second-place finish for John Kasich. The primary path now winds through a number of northeastern states, and Ted Cruz can expect to have little traction in any of them. A new Quinnipiac poll shows Trump getting to a near-majority among likely Republican primary voters, riding a wave of populist anger at the GOP’s leadership:

Donald Trump is close to the 50 percent mark among Connecticut likely Republican primary voters, while Ohio Gov. John Kasich tops Sen. Ted Cruz of Texas in the race for second place, according to a Quinnipiac University poll released today. …

Likely Republican primary voters give Trump 48 percent, with 28 percent for Kasich and 19 percent for Cruz. Only 5 percent are undecided, but 25 percent of those who name a candidate say they might change their mind before the April 26 primary.

By 59 – 33 percent, Connecticut Republicans want an outsider for president, rather than someone experienced in politics. Among those who want an outsider 75 percent back Trump.

“Connecticut Republicans have gone for outsider candidates such as Linda McMahon and Tom Foley. They continue that trend with Donald Trump,” said Quinnipiac University Poll Director Douglas Schwartz, PhD. “The conventional wisdom that Sen. Cruz is too conservative for Connecticut looks true, as he comes in a distant third in the Republican primary. Kasich clearly is outpacing Cruz for second, but running well behind Trump.

Next Tuesday appears to be shaping up as a landslide night for Trump. Five states will go to the polls, and Trump has big leads in all five: Delaware, Maryland, Pennsylvania, Rhode Island, and of course Connecticut. Pennsylvania might matter least, since Congressional-district delegates remain unbound, but it will still give Trump plenty of bragging rights about momentum and primacy. In RCP averages, Trump has a 20-point lead in PA, and also leads by 14 in Maryland. Pollsters have not done enough surveys in the other two states to establish any baselines, but so far it seems likely that Trump’s appeal will reach all the way down the Atlantic seaboard.

There hasn’t been much polling in Connecticut either. Emerson released a poll last week that showed essentially the same positioning, putting Trump up 24 points 50/26 over Kasich. In this Q-poll, Trump leads in almost every demo except among young people, where Kasich beats him 39/32. For those who value experience, Kasich leads 61/30 over Cruz, while Trump only gets four percent, but only a third of voters want someone with experience; 59% want an outsider.

So far, the polls in the northeast and last night’s primary landslide in New York show that the #NeverTrump efforts haven’t done much to slow Trump down, at least not in his natural strongholds. Politico’s Adam Schaeffer posits that the attacks might provide an unintended boost to Trump, based on focus-group testing research  [see correction below] as well as results in previous primaries:

The key distinction in what we do compared with focus groups and regular polls is that we do not ask what people think of a message; we observe the impact the message has on vote choice in the treatment group compared with the placebo-control group. Since the respondents are randomly assigned to each group, average support should be about the same in each. Any statistically significant difference between support in the two groups is because of the impact of the ad.

And we found that Cruz’s anti-Trump ad backfires. It doesn’t hurt Donald J. Trump. It helps him.

Our clinical message trial showed Cruz’s anti-Trump actually made voters more likely to vote for Trump, boosting his support by 3 percentage points overall. That’s not a very large increase for the sample as a whole (and not statistically significant). But for blue-collar voters, the attack ad increased support for Trump by 18 percentage points; and it increased support among blue-collar men by more than 33 percentage points. (36 percent of the blue-collar men who watched the coke ad, for example, said they would vote for Trump—compared with 69 percent of the blue collar men who watched the anti-Trump ad.) And in both subgroups, incidentally, the anti-Trump ad caused actually support for Ted Cruz to fall.

While the ad decreased support for Trump among other groups, such as middle-class voters overall, by a few points, that large backlash among blue-collar voters is significant. It is certainly not a result you want to see from broadcast ads that are viewed by a wide range of voters.

If the (perceived, at least) establishment spends $70 million attacking the candidate who successfully positions himself as the anti-establishment choice, should we be surprised when the attacks firm up that candidate’s credibility? This has been a recurring pattern for the past year when it comes to Trump. The successes for Cruz against Trump have come not so much from these personal attacks, but from Cruz directly attacking Trump on policy while offering positive alternatives for voters. Cruz didn’t win Wisconsin because Trump lost support from his base — in fact, Trump largely performed to pre-primary polling in Wisconsin — but because Cruz successfully wooed the rest of the GOP primary voters to his side.

Cruz won’t get much chance next week to dent Trump’s Teflon in the northeast. After that, though, almost all of the remaining primaries will take place in the Midwest and West, where Cruz has had much more success with that kind of campaigning. If he’s not getting kneecapped by outsiders, Cruz might well gain enough delegates to force multiple ballots at the convention. The May 3rd primary in Indiana will be the biggest test of whether Cruz can execute that strategy, and especially a key indicator of whether the expected rout on April 26 has damaged Cruz’ effectiveness and support west of Pennsylvania.

Update: Adam Schaeffer e-mailed me to point out (in a friendly manner) the difference between clinical messaging research and “focus-group testing,” which I inaccurately called it.  I’ve corrected it above.