President Trump's pollster is fighting back against the “Trump is losing his base” narrative — a narrative to which I subscribe. But his arguments leave something to be desired. And in fact, he basically just confirmed that the cracks in the president's base are real.

Here's what Tony Fabrizio tweeted:

First, let's note that Trump's own pollster just shared data showing 1 out of every 4 Republicans (25 percent) disapprove of Trump — a number that has increased by 6 percentage points since June. If that's not losing your base, I'm not sure what is. Does it mean the base has completely deserted him? Of course not. But it shows the steady deterioration from other polls is confirmed by Trump's own pollster. And it sure as heck rebuts Trump's recent claim that his base is “stronger than ever.”

As I noted in my Trump's-base-is-crumbling piece a few weeks back, the 71 percent of Republicans who approved of Trump in a recent Investor's Business Daily poll was already lower than President Barack Obama's base support ever dipped. And while some recent polls show GOP approval of Trump as high as the 80s — CNN recently had him at 83 percent among Republicans, and Quinnipiac University had him at 81 percent — Fabrizio's numbers are actually among the worst we've seen for Trump, to date.

And now the comparison. Fabrizio suggests that, while his own numbers show Trump's support among the base declining slightly, the GOP Congress is losing support even faster. The argument seems to be that congressional Republicans who are criticizing Trump and/or failing to pass his agenda items are actually worse off for it and that Trump remains much stronger than they are.

A new Washington Post-ABC News poll tracks President Trump's response to the events in Charlottesville. (Peter Stevenson/The Washington Post)

But while we have plenty of data on Trump's incremental decline with the base, this is the first publicly available data to suggest the congressional GOP has seen a big decline with base voters in recent weeks. We simply don't have another poll showing GOP approval of Republicans in Congress since the end of June, so we don't know if Fabrizio's numbers are statistical noise.

And it seems possible that it is just that: noise. Previous polls from Quinnipiac University and Marist College show plenty of ups and downs on this measure. The Q poll had 66 percent of Republicans approving of GOP leaders in Congress back in February, only to see it drop to 44 percent in May and then rebound to 57 percent in late June.

Here's how that looks:


So Fabrizio's poll shows the GOP's approval among Republicans voters declining from +36 in June to +8 today; Quinnipiac has shown it dropping from +41 down to -3 and then back up to +22. That's a lot of variation, meaning the two data points in Fabrizio's poll need to be validated by other data (which, again, we don't yet have).

And the decline in Fabrizio's poll appears to owe more to the peak in support than the valley. The 68 percent of Republicans who approved of the congressional GOP in his June poll was the highest in any poll since February, and it came at a time when other pollsters showed considerably lower numbers. The 54 percent who approve of the congressional GOP in his most recent poll is actually pretty on-par with Quinnipiac's regular polling this year. It doesn't suggest a huge drop, when compared to other polling.

The final point is that approval of congressional Republicans is much more amorphous. Many people decry Congress as a whole but approve of their own lawmakers. That's undoubtedly part of the reason for all the ups and down, but it also means that the practical implications are much more limited for specific members.

If Fabrizio is suggesting Republicans need to fall in line behind Trump to please the GOP base, these numbers may be somewhat compelling. But it's unlikely they'll frighten congressional Republicans, and they don't change the fact that Trump's pollster just confirmed his base is weaker than ever.