Second time: Do you have any evidence that Nate Silver is giving some preference to Clinton that the numbers don't support? Do you have any evidence that Silver is doing anything other than aggregating polling data weighted for the past performance of each poll (and size of sample, whether they called cell phones, etc.)?
I do not believe that Nate Silver is deliberately skewing his data analysis to favor anyone. What I do believe is that Nate's process applied to politics inherently favors establishment candidates and campaigns that employ establishment practices in traditional election settings. My argument isn't focused on whether or not Nate Silver is a closet Clintonista sabotaging his own livelihood to make HRC look better than she is.
My argument is based on the proposal that Silver's establishment political metrics are fine at producing establishment consistent data and analyses. Ultimately, however, the question is whether or not there is enough non-traditional, non-establishment political activity occurring to have a significant impact upon the Democratic primary and possibly the general election as well. We will know for certain once the only polls that actually matter begin actually occurring. If I'm mistaken, then it should be fairly clear by the end of the first week in March, if not the last week in Feb.
Is there any properly conducted poll anywhere that shows Clinton still losing ground to Sanders?
Generally, I'm more focused on longer term trends than a single poll is usually capable of capturing, and until the last month, I'm really not too concerned about the daily, or even weekly, wiggles.
That said...
Over the last month, I see the following which I would consider, lacking evidence to the contrary, to be properly conducted polls:
Over the first two weeks I see what look to be:
an Ipsos/Reuters Poll indicating Clinton 57 Sanders 28
Oct. 31 - Nov. 4
another IPSOS/Reuters Clinton 52 Sanders 35
November 7–11
A Public Policy Polling survey indicting Clinton 67 Sanders 25
November 12–14, 2015
an average difference of (29 + 17 + 42)/3 =~29 points difference
Over last 2 weeks:
Public Policy Polling Clinton 59% Sanders 31%
November 16–17, 2015
Ipos/Reuters Clinton 52% Sanders 31%
November 14–18, 2015
ABC News/Washington Post Clinton 60% Sanders 34%
November 16–19, 2015
An average difference of (28 + 21 + 26)/3 = 25
https://en.wikipedia.org/wiki/Natio..._Democratic_Party_2016_presidential_primaries
Hillary had a (fully and admittedly, within MoE) ~4 percentage point larger lead in the earlier part of this month than she does currently.
But the only polls that matter don't start for another couple months or so.