One way you know that CBS News’s new poll of the presidential race is bad news for former president Donald Trump is that Trump’s campaign is trying to spin it as unreliable.

In a “confidential memorandum” sent to reporters on Sunday, a data consultant for the campaign argued that the poll, conducted by YouGov, had been “manipulated” to show an advantage for Vice President Harris — a shift since the last CBS-YouGov poll released last month. It was a return to the days of poll “unskewing,” a central component of Republican optimism about Mitt Romney’s chances in the 2012 presidential election … which he then lost.

The objections from the Trump campaign can be dismissed most easily by considering the motivations of the two parties involved. YouGov is in the business of providing a wide range of businesses and media outlets with reliable evaluations of public opinion, the sort of effort that rewards accuracy over the long term. (YouGov is the fourth-best pollster in 538′s ratings.) Trump’s campaign, on the other hand, is in the business of presenting him as a political juggernaut with an unmatched level of support.

In other words, why the pollster would tweak a midyear presidential poll to benefit the Democrat is not clear. Why the Trump campaign would like to suggest that the poll was tweaked is obvious.

Even given its motivations, the response from Trump’s campaign to this one poll was odd. The poll, as the memo notes, shows a subtle shift in the race, one that puts Harris in front of Trump but still well within the margin of error. There are other indicators that the race is shifting in Harris’s direction (polling analyst Nate Silver’s average has Harris leading nationally and gaining in swing states), but modestly. This isn’t surprising, given that changes in political support are generally gradual. It’s possible that as the campaign progresses, Harris will open up a wider lead — or Trump will regain one. For now, though, the shift is small.

That said, the CBS poll does suggest that a wider Harris lead is quite possible. CBS has published three polls evaluating the presidential contest over the past month. The first asked respondents to choose between Trump and President Biden, his opponent at the time. The second asked for a choice between both Biden and Harris. The third, released this weekend, had only Harris. This lets us see both how the Trump-Harris race compares with the Trump-Biden one — and how Harris’s position has changed since she became the presumptive nominee.

The biggest changes are among independents and Black respondents. In the CBS poll released in the middle of last month, Trump led Harris among independents by 11 points; they’re now tied. Among Black respondents, Harris led by 55 points. Now she leads by 63 — much closer to the Biden-Trump split shown in 2020 exit polls, though still less heavily Democratic.

The bigger shift is seen when respondents were asked to select from a wider field of candidates. Here we see bigger shifts over the past month, in large part because respondents considering third-party candidates shifted their support to Harris. Notice, below, that the change in support for Trump is relatively subtle month-to-month. The change in support for his Democratic opponent, though, is often significant.

That’s in part because younger voters and non-White voters — two central components of the Democratic base — are more likely to back Harris in a larger field than they were to back Biden. In CBS’s new poll, both groups also indicated that the shift of the nomination to Harris made them more motivated to vote in November. Such questions should be taken with a grain of salt, but this comports with polling over the course of the year showing broad apathy about Biden’s reelection bid.

We can see how this works when looking at the trend in other questions included in CBS’s poll. Last month, half of Biden supporters said their vote was a function of wanting to stop Trump, not of support for the president. Now, about 4 in 10 indicate that their vote is against Trump — meaning that more respondents plan to cast an affirmative vote for Harris.

These numbers are consistent with YouGov’s broader finding that Harris has had a jump in favorability among poll respondents, itself a not-uncommon shift for a new candidate. Trump and his allies are trying to portray the enthusiasm for Harris’s campaign as inauthentic, but they might recall that before Trump’s campaign launch in 2015, his favorability among Republicans was well below water.

YouGov’s poll has also consistently asked respondents who have a preference between the two major-party candidates whether they would ever consider voting for the other party’s candidate. Overall, fewer Trump supporters say they would never support Harris than said they would never support Biden last month. That’s particularly pronounced among younger respondents, women and Hispanic respondents. (There were not enough Black Trump supporters to include in the chart below.)

Some of this is a function of the newness of Harris’s campaign. As her campaign is rolled out, it’s likely that opposition from many of those who support Trump will firm up. It’s possible, too, that she’ll peel away some of that support or, at least, soften the support in a way that dampens turnout in November.

It’s early. That “confidential” Trump campaign memo is correct in noting that the shift in the CBS poll is subtle, though its disparagement of the pollster is not. What this and other polling shows, though, is a shifted presidential contest, one that has moved from mirroring 2016 to one that is potentially, instead, mirroring 2020.

And in political polling, the direction of movement can be more important than scale.