UPI/CVoter daily tracking poll

[Wiz in Wis]

It is cool to see how one day in the sample can really reset a trend.

[Wiz in Wis]

UPI/CVoter Through 9/18 (7-day rolling average)Clinton + 1.2%
Clinton 48.0% (+0.6%)
Trump 46.8% (-0.6%)
Other 5.3% (+0.1%)

http://www.upi.com/Top_News/US/2016/09/21/UPICVoter-poll-Hillary-Clinton-holds-118-lead-over-Donald-Trump/2301474469318/

[Wiz in Wis]

UPI/CVoter Through 9/25 (7-day rolling average) Clinton +1.2
Clinton 47.8% (-0.3%)
Trump 47.1% (+1.0%)
Other 5.3% (-1.0%)


UPI/CVoter Through 9/24 (7-day rolling average) Trump +0.2%
Clinton 47.5% (--)
Clinton 46.1% (--)
Other 6.3% (--)

First UPI release in nearly a week. They do not have the PDF attached, so I do not have the other numbers for the days missing in the tracker. If someone comes across it, please let me know.

http://www.upi.com/Top_News/US/2016/09/26/UPICVoter-poll-Hillary-Clinton-Donald-Trump-in-virtual-tie-heading-into-first-debate/8181474896561/

Post has some topline math issues... not to be a dick about it, but thought you should know.

[Wiz in Wis]

Link for updates from last week:

http://projects.fivethirtyeight.com.s3.amazonaws.com/polls/20160925_National_1.pdf

[Wiz in Wis]

UPI/CVoter Through 9/27 (7-day rolling average) Clinton +0.2%
Clinton 47.9% (+0.4%)
Trump 47.7% (--%)
Other 4.5% (-0.4%)

UPI/CVoter Through 9/26 (7-day rolling average) Trump +0.2%
Trump 47.7% (+0.6%)
Clinton 47.5% (+0.1%)
Other 4.9% (-0.2%)

http://www.upi.com/Top_News/US/2016/09/28/UPICVoter-Hillary-Clinton-regains-slight-lead-in-first-post-debate-poll/8801475069632/

Marginal movement post-debate.

It would be unlikely to be that big a shift anyways... but even if we did just extrapolate that shift times 7 days (assuming the shift is constant across all days and doesn't grow/shrink), then we'd expect that Clinton gets a 2.8% bump - all things being equal (which is not at all likely). A small shift towards Clinton fits in with the conventional wisdom and poll structure. Of course, none of this takes into account the days that drop off either. Clinton had a relatively good day 7 days ago, so any gains may be muted by those days dropping off.

[Wiz in Wis]

UPI/CVoter Through 10/2 (7-day rolling average) Trump +2.5%
Trump 49.4% (+0.4%)
Clinton 46.9% (-0.6%)
Other 3.7% (+0.2%)

1285 LV; September 26-October 2, 2016

UPI/CVoter Through 9/28 (7-day rolling average) Trump +1.5%
Trump 49.0% (--%)
Clinton 47.5% (-0.3%)
Other 3.5% (+0.3%)

1288 LV; September 25-October 1, 2016

UPI/CVoter Through 9/30 (7-day rolling average) Trump +1.2%
Trump 49.0% (-0.3%)
Clinton 47.8% (+0.5%)
Other 3.2% (-0.2%)

1223 LV; September 24-30, 2016

If this tracker is to be believed, Trump gained about 2 points post debate.

Sourcing: https://www.documentcloud.org/documents/3118148-US2ndOct-Report.html#document/p1

At minimum, the methodologies of these polls clearly appear to be creating divergent results.

[Wiz in Wis]

Hillary is grabbing this tracker by the p***y!

[Wiz in Wis]

UPI likely underestimates Democratic margin because of this.

"Because the poll is conducted online and individuals self-select to participate, a margin of error cannot be calculated."

Trump supporters are far more likely to take self-participant polls, as evidenced by their overzealous will to spam click on Trump's first debate poll.

Results are still weighted by demo and then they apply an LV screen, so it's probably a minimal effect, if at all.