On the face of it, answering the question as to what happened in the GOP primary in Wisconsin seems like a no-brainer. As Edward Morrissey writes, Trump shot himself in the foot—dissing popular governor Scott Walker, and flubbing abortion questions—and ended up losing by 13 points, 35 to Cruz’s 48.

To shore up this argument about a Trump reversal in Wisconsin, Morrissey cites a Wisconsin poll from late January and one from late February, the first of which had Trump leading by 6 and the second by 10. So the narrative seems to make sense—that is, until you actually look a bit deeper, when you find that something additional might have been going on.

That “something” was named Marco Rubio, who dropped out between those two polls and the primary. His leaving the scene meant that it narrowed down to a Cruz/Trump race with Kasich far behind, instead of the three-person Trump/Rubio/Cruz race it had been (there were other candidates too, back in January and February, but they were far back in the pack, and at that point Kasich had been almost invisible in the state). On the day of the primary, Trump actually received a higher percentage of votes than the scores he had gotten in either of those two earlier polls, the first of which had him at 24% and the second at 30%. In the primary he got a 35% share of the vote, which represents a significant increase, not a decrease.

But that January poll with Trump at 24 had also featured a higher combined Rubio/Cruz total of 34. In the February one had where Trump had led with 30, Rubio and Cruz also had a higher combined total, 39.

Therefore it appears that only a small portion of the voters who had supported Rubio (and the other candidates who dropped out during that period) ended up joining the Trump camp by primary time. They went in significantly greater numbers to Cruz and Kasich, and that seems to have accounted for Trump’s loss as well as for the magnitude of his loss. This doesn’t invalidate what Morrissey wrote in terms of the possible reasons that Trump failed to pick up the votes of the dropout candidates and close the deal; he certainly might be correct about that.

If you look at this poll from about a week before the primary—when Rubio had left the scene and it was just Cruz, Trump, and Kasich—you’ll find that although Trump had achieved the 35% that he actually received during the primary, Cruz was only one point higher than Trump, and Kasich was at 19% (that’s 5 points more than he got in the primary). So the movement between then and the primary was all in the Cruz and Kasich dimensions, up for Cruz and down for Kasich. And this poll, taken about two weeks before the primary, shows something similar: Cruz 36.2, Trump 31.4, and Kasich almost at 20.8.

In that poll Trump was within the margin of error of his final share of the primary vote, Kasich considerably further off, and Cruz the farthest of all, with Cruz gaining considerable votes by primary time and Kasich losing some.

It is risky to extrapolate too much from Wisconsin to other states; New York, for example, is shaping up to be a very different story. But in recent months there does seem to have been considerably less movement in Trump’s totals in each state than in the vote share of the other candidates. For Trump that’s true in both the positive and the negative sense: lately he doesn’t seem to have lost many of his supporters, but he doesn’t seem to have won all that many new ones, either.

[Neo-neocon is a writer with degrees in law and family therapy, who blogs at neo-neocon.]