This appears to be the Perpetual Media Question for 2017: Has Donald Trump suffered a body blow to his base support? The latest Quinnipiac poll gets described by Politico as “brutal for Trump,” showing erosion in some of his key demographics, in polling taken prior to the firing of James Comey. Quinnipiac itself calls the downward turn in Trump’s fortunes as “near record,” but that should come with an important caveat:

American voters, who gave President Donald Trump a slight approval bump after the missile strike in Syria, today give him a near-record negative 36 – 58 percent job approval rating, according to a Quinnipiac University national poll released today. Critical are big losses among white voters with no college degree, white men and independent voters.

Today’s job approval rating compares to a negative 40 – 56 percent approval rating in an April 19 survey by the independent Quinnipiac (KWIN-uh-pe-ack) University and a negative 35 – 57 percent score April 4, his lowest approval rating since he became president.

To understand the topline numbers in context, however, the chart published by Quinnipiac comes in very handy:

Bear in mind that the margin of error for this poll series is ± 3 points. In that context, the yellow line of support looks … rather consistent over the last four months. The biggest difference is that the 20% undecided in January have mostly migrated to disapproval. That’s not to say that Quinnipiac is incorrect in calling this result a “near-record negative” for the gap, which it clearly is, but the bigger story is more that Trump’s not winning any converts rather than sinking in support over the last four months. That’s not really brutal, but it’s not exactly sparkling-shiny for the White House, either.

What about the demographics? Those have necessarily larger MoEs, although Quinnipiac does not provide those calculations. Among white men, Trump went from a 53/41 to 48/46 three weeks ago, which looks like not much change within the MoE. The shift on education seems more pronounced, with non-college whites going from 57/38 to 47/46, but the sample size for this subdemo might be small enough for this to be statistically less significant too. Independents did shift more negatively too, from 38/56 to 29/63. There is almost no change outside the MoE for overall opinions on his personal qualities, either; the most dramatic of these was on honesty, where he went from 37/58 to 33/61, right around the MoE. That strongly suggests that the MoE on those demographics is wide enough to chalk this movement up to statistical noise.

For what it’s worth, the Q-poll is actually an outlier among recent surveys. The RealClearPolitics aggregation puts the average for Trump’s approval rating at 42/53. Over the past month, Quinnipiac’s 36/58 is the lowest; the only other poll putting Trump’s approval below 40 is the IBD/TIPP poll, which usually produces friendlier results for Trump. The RCP aggregate support level has also not changed much over the last three months, varying between 40% and 46%.

Under normal circumstances, we could expect the next iteration of the Q-poll to show adjustments within the MoE that would make things look slightly better. However, the next poll will undoubtedly show some impact from the firing of James Comey, which will likely not be positive. Trump defenders will no doubt blame the media coverage of the decision, but they’d be better advised to focus on the disorganized and contradictory messages coming out of the White House for any loss of confidence and support that results. Unless Trump hits a home run with an independent and credible replacement, this one’s going to sting for a while, at least on the negatives.