What’s up with the RCP average?

Without getting conspiratorial, the RCP poll numbers from which the final RCP average (Clinton +2.9%) is derived give rise to two observations:

First, two distinctly separate sets of polls make up the RCP final average, and the picture of the race going into Election Day depends entirely on which of the two one consults.

The mainstream media-sponsored polls (Bloomberg, Clinton+3; ABC, Clinton+3; CBS, Clinton+3; NBC, Clinton+5) show a small Clinton lead, still within or close to the margin of error;

But the three leading independent polls show nearly the reverse, though Trump’s average lead is slightly smaller than Clinton’s above (IBT/TIPP, Trump+2; LAT/USC, Trump+5; Rasmussen, Clinton+2).

So what is it?  Trump ahead by about 2, or Clinton ahead by about 4?

This isn’t a small difference in the world of professional polling.

Either legitimate methodological or sampling differences explain the large difference, or someone is trying to create and support a self-fulfilling narrative.

Second, a phenomenon noted throughout the last month of the campaign appears again in the final polls making up the Election Day RCP average.

Certain polls have regularly appeared in the RCP average during the last month – e.g., ABC/Washington Post, IBD/TIPP, LAT/USC, with less frequent appearances by Fox.

But today, and also at various points during the final month, the Clinton “lead” in the RCP average appears to have been rescued by the sudden appearance of the cavalry in the form of a sporadically produced newcomer or two – a newcomer that always seem to show up at the right time, and to be cited at RCP just as the Clinton lead would have sagged or disappeared if based only on the “regulars.”

For example, today RCP throws into the broth numbers from such trusted and reliable sources as Bloomberg, CBS, and NBC, whose substantive coverage of the race has been such a model of even-handedness, and whose numbers alone prop up the alleged Clinton lead of 2.9%.

Sean Trende no doubt has a cogent and convincing explanation for these observations/cavils.  I’d like to hear it.

The RCP “average,” with all its strange newcomers, may prove correct.  Or not.

On the other hand, the very different picture shown by independents’ numbers may prove correct, or at least more correct.

After all this is over, perhaps Mr. Trende could explain once again, for the benefit of those who like me are somewhere between confusion and suspicion, just how polls are either included in the RCP average (e.g., NBC today) or excluded (e.g., Rasmussen today).  

In the meantime, a note to Trump supporters: Absolutely no one knows now what the actual electorate will look like when the dust settles, or how it will vote.  Remain resolute, and vote today for Trump.