The Morning Consult numbers in Wisconsin should be a good lesson for people who get jittery about leads.
From August on, Biden has received 50-53% (with one outlier of 48%). Trump has received 41-45%.
Both are undoubtedly within the margins of error. Therefore, if they do turn up randomly, then Biden’s lead can range anywhere from 5 to 12 percentage points. They do, in fact, range from 6 to 11 (except for that one outlier). The largest bounce from one day to the next is three percentage points.
I hope no one believes that 3%, let alone 5%, of Wisconsinites are changing their minds daily. It’s just the proper and expected randomness of good polling.
Only trend lines matter. The average of Biden’s lead in the first five polls ending in August was 10.0%. The average in last five polls ending in August was 9.6%. The average of the first five polls in September was 8.6%. That may be meaningful or may be too small a difference to pay attention to. Even if it is meaningful, a straight line puts him at a 5.8% lead in November. Since he’s bouncing between the same limits as he’s shown for more than a month, I would need to see a break in those limits before I started worrying or celebrating.
There’ve been many articles recently about how extraordinary steady Biden’s lead has been for the last several months. This is another example.
Calm down. Stop the agita about the polls. Turnout will be more important anyway.
Wisconsin ticking up a bit (for Biden) in the most recent poll or two seems to have been a key reason Nate Silver’s model now has him winning 72 of 100 simulations (up from 71 as of yesterday).
Also, as we get closer to Election Day, all other things being equal, Biden will win ever higher percentages of these simulations, as the chance of a major shift (and/or an earth-shattering event) gets less.
(But all other things are unlikely to stay equal, of course.)
I think Ohio is also linked to the “blue wall” states in a way that Arizona, North Carolina, Florida, and Texas aren’t. In other words I can envision scenarios where Biden wins Arizona, Florida, North Carolina, or even Texas while losing some of all of the blue wall. I can’t envision Biden winning Ohio while losing the blue wall.
Just a point of interest here. Each day that Morning Consult reports for Wisconsin here is itself the average of 10 days of polling so there’s really no need to then average consecutive days of those averages together.
In fact taking the five day average of the ten day average introduces a bias toward day d-4 to day d-9 (where d-0 is the most recent day polled) since those six days are the only days to be represented in every ten day average.
It’s not necessarily wrong to do this, but in this particular case it does seem to introduce some drawbacks and no benefit that I can see.
I don’t know that about Morning Consult. That would tend to smooth out the data. Fortunately, it doesn’t materially change my point. Nothing much is happening that wouldn’t be expected by random bounces within the margin or error. And the trend over time is too small to be concerned about.
Each daily data point represents a simple 10-day moving average, and is based on an average of 797 interviews with likely voters in Wisconsin, with a margin of error of 3.5 percent.
But as I said, it’s just a point of interest. It’s a sidebar to, and not an attack on, the point you were making.
Politico is tracking the contested states, for those interested. Not sure why they wouldn’t rate WI as at least leaning Dem given the polling landscape of late, but I suppose they’re being conservative.
Democratic absentee requests are far higher than Republican requests in PA, NC and FL. On the flip side, Republicans are really emphasizing their ground operation this election (Pandemic? What pandemic?), which they are counting on to neutralize the Dems’ advantage come Election Day.
That’s good to know, Nonsuch. Odd, perhaps, because the pandemic is now more persistent in parts of the South and Great Plains that lean Republican. But, as you said, they’re counting on Trump voters to act with a bit less regard for safety.
Back to 538…Biden’s chances are at 75% now. This uptick seems fueled mainly by recent gains in Pennsylvania, Wisconsin, and Florida (slight to moderate Biden leans in each), and also Ohio (Trump lean, but Biden is closing the gap).
Also, MINNESOTA — quite a remarkable comeback for Biden over the last week or two. This morning, 538 added two A+ rated polls there, and if I recall correctly, they both have Biden up by 8%.
I wonder how much of Biden’s uptick is the passage of time. Others have mentioned that the election held today model has Biden at 90%. Presumably this means that the primary model will approach 90% as Election Day approaches, even if the polls remain unchanged. To use a sports analogy, a football team up 20-14 at the end of the 3rd quarter is doing better than that same team up 20-14 at halftime.
I think Trump’s attacks on mail-in voting may end up backfiring spectacularly. He’s put his party in a bad position: If Trump screws up mail-in voting among rural and suburban seniors, Republicans may have no choice but to hope that there is a massive election day turn out. And given the pandemic, that may just…not happen.
Republican Senators and Representatives have been trying to get him to stop the anti-vote-by-mail rhetoric for a couple-three months now without much success. They’re afraid they’ll lose close elections because of it. It’s one of the best things about this election.