Democratic Primary Live Polling Average
       |           

Welcome, Guest. Please login or register.
Did you miss your activation email?
May 01, 2024, 12:47:05 AM
News: Election Simulator 2.0 Released. Senate/Gubernatorial maps, proportional electoral votes, and more - Read more

  Talk Elections
  Election Archive
  Election Archive
  2020 U.S. Presidential Election (Moderators: Likely Voter, YE)
  Democratic Primary Live Polling Average
« previous next »
Pages: [1] 2
Author Topic: Democratic Primary Live Polling Average  (Read 2774 times)
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« on: March 29, 2019, 06:35:57 PM »
« edited: February 03, 2020, 04:11:52 PM by Sorenroy »

Updated February 3rd:



National Polling:
Biden — 27%
Sanders — 22%
Warren — 13%
Bloomberg — 9%
Buttigieg — 7%
Yang — 4%
Klobuchar — 3%
Steyer  — 2%
Gabbard — 1%
Bennet — 1%

Blue = Biden; green = Sanders. White means that the poll leader is polling less than 20% and/or there is a tie in the state (look at the second post for details). Light Grey means the state has had a polling average at one point, but the most recent poll is a month old or older. Number indicates the number of polls conducted in the past month (if less than three).

Polls come from 538's Poll Tracker. If it's not on there, it's not included. If you notice something missing, please let me know. Only one poll per agency is allowed at a time. Hypothetical two-way match-ups and polls including hypothetical candidates conducted after the qualifying deadline for the first primary debate are ignored. If there are multiple polls in a month from a given pollster, I select based on (in order of importance) date concluded, date started, number of candidates, sample size, and percent decided. Graphs include the average of all polls conducted in the past month (by different agencies). The candidates shown were selected if they either received ≥10% before or ≥5% on or after the qualifying deadline for the first primary debate in any state or nationally's average of polls when three or more polls were present. Polls are only included if they were started within the month (ie: monthly polls must be less than a month old). If there are less than three polls conducted in the past month, the graph is removed until there are three.

Suggestions are welcome and appreciated. I am an amateur at this kind of stuff and would love to cut down the amount of time it takes me to put these out.



New Polls:
National - Morning Consult: Biden leads Sanders, Bloomberg, and Warren 28-24-14-14
IA - David Binder Research: Buttigieg leads Sanders, Biden, Warren, and Klobuchar 19-17-15-15-11
NH - University of Massachusetts Lowell: Sanders leads Biden, Warren, and Buttigieg 23-22-19-12

Old Polls:
NH - Monmouth University: Buttigieg leads Biden, Sanders, and Warren 20-19-18-15
CA - Public Policy Institute of California: Sanders leads Biden and Warren 27-24-23
CA - Tulchin Research: Sanders leads Biden and Warren 29-25-12
CA - Capitol Weekly: Sanders leads Warren, Biden, and Buttigieg 24-21-20-11
NM - Emerson College: Sanders leads Biden and Yang 28-27-10
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #1 on: March 29, 2019, 06:37:01 PM »
« Edited: February 03, 2020, 04:08:44 PM by Sorenroy »

National:

Biden — 27%
Sanders — 22%
Warren — 13%
Bloomberg — 9%
Buttigieg — 7%
Yang — 4%
Klobuchar — 3%
Steyer  — 2%
Gabbard — 1%
Bennet — 1%



IA - Iowa (February 3, 2020):

Sanders — 21%
Biden — 20%
Buttigieg — 17%
Warren — 16%
Klobuchar — 10%
Yang — 4%
Steyer — 3%
Gabbard — 1%
Bloomberg — 1%



NH - New Hampshire (February 11, 2020):

Sanders — 25%
Biden —17%
Buttigieg — 14%
Warren — 13%
Klobuchar — 6%
Gabbard — 6%
Yang — 5%
Steyer — 4%
Bloomberg — 1%
Patrick — 1%



NV - Nevada (February 22, 2020):

Biden — 23%
Sanders — 21%
Warren — 12%
Steyer — 9%
Buttigieg — 7%
Yang — 4%
Klobuchar — 3%
Booker — 2%
Gabbard — 1%
Bloomberg — 1%



SC - South Carolina (February 29, 2020):

Biden — 35%
Sanders — 16%
Steyer — 15%
Warren — 10%
Buttigieg — 5%
Yang — 2%
Gabbard — 1%
Klobuchar — 1%
Bloomberg — 1%
Booker — 1%



CA - California (March 3, 2020):

Sanders — 25%
Biden — 20%
Warren — 19%
Buttigieg — 8%
Bloomberg — 5%
Yang — 4%
Klobuchar — 3%
Steyer — 3%
Gabbard — 2%



NC - North Carolina (March 3, 2020):*
Biden — 31%
Sanders — 18%
Warren — 15%
Bloomberg — 8%
Buttigieg — 6%
Yang — 5%
Klobuchar — 3%
Steyer — 3%
Booker — 1%

*Based on 1 Poll.



TX - Texas (March 3, 2020):

Biden — 30%
Sanders — 21%
Warren — 14%
Bloomberg — 11%
Buttigieg — 7%
Klobuchar — 4%
Steyer — 2%
Yang — 2%
Gabbard — 1%
Bennet — 1%



UT - Utah (March 3, 2020):*
Sanders — 27%
Warren — 14%
Biden — 12%
Bloomberg — 10%
Buttigieg — 5%
Yang — 5%
Klobuchar — 3%
Gabbard — 1%
Steyer — 1%

*Based on 1 Poll.



MI - Michigan (March 10, 2020):*
Biden — 27%
Sanders — 22%
Warren — 14%
Bloomberg — 9%
Buttigieg — 6%
Yang — 4%
Delaney — 2%
Klobuchar — 2%
Steyer — 2%
Bennet — 1%
Gabbard — 1%
Patrick — 1%

*Based on 1 Poll.



MO - Missouri (March 10, 2020):*
Biden — 39%
Bloomberg — 14%
Warren — 9%
Klobuchar — 8%
Sanders — 7%
Buttigieg — 6%
Yang — 2%
Steyer — 1%

*Based on 1 Poll.



WA - Washington (March 10, 2020):*
Sanders — 26%
Biden — 21%
Warren — 16%
Bloomberg — 12%
Buttigieg — 8%
Yang — 4%
Klobuchar — 3%
Steyer — 2%

*Based on 1 Poll.



FL - Florida (March 17, 2020):

Biden — 37%
Sanders — 14%
Warren — 10%
Bloomberg — 9%
Klobuchar — 5%
Buttigieg — 4%
Yang — 3%
Steyer — 2%
Booker — 1%



OH - Ohio (March 17, 2020):*
Biden — 32%
Sanders — 21%
Warren — 11%
Bloomberg — 10%
Buttigieg — 6%
Klobuchar — 2%
Steyer — 2%
Yang — 2%
Bennet — 1%
Gabbard — 1%

*Based on 1 Poll.



WI - Wisconsin (April 7, 2020):

Biden — 23%
Sanders — 23%
Warren — 14%
Buttigieg — 11%
Bloomberg — 7%
Klobuchar — 4%
Yang — 4%
Booker — 1%
Gabbard — 1%
Steyer — 1%



NY - New York (April 28, 2020):*
Biden — 30%
Bloomberg — 17%
Sanders — 17%
Warren — 14%
Buttigieg — 7%
Steyer — 4%
Yang — 4%
Bennet — 2%
Klobuchar — 2%
Delaney — 1%
Gabbard — 1%
Patrick — 1%

*Based on 1 Poll.



PA - Pennsylvania (April 28, 2020):*
Biden — 27%
Sanders — 18%
Warren — 13%
Bloomberg — 8%
Buttigieg — 7%
Klobuchar — 4%
Yang — 4%
Gabbard — 1%
Steyer — 1%
Bennet — 1%

*Based on 2 Polls.



NJ - New Jersey (June 2, 2020):*
Biden — 28%
Sanders — 25%
Warren — 15%
Bloomberg — 9%
Buttigieg — 6%
Yang — 6%
Klobuchar — 4%
Gabbard — 3%
Delaney — 2%

*Based on 1 Poll.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #2 on: April 04, 2019, 11:45:45 PM »

Completely revamped the thread. Added in state level tracking graphs, moved them down to the reply, and added a map to the main (first) post, among several other things.

Suggestions welcome.
Logged
Senator Incitatus
AMB1996
Junior Chimp
*****
Posts: 6,508
United States


Political Matrix
E: 2.06, S: 5.74

Show only this user's posts in this thread
« Reply #3 on: April 05, 2019, 03:17:23 AM »

It's kind of hard to tell which colors correspond to which candidates. Could you maybe list the key/legend in the order candidates are currently stacked in? Or maybe not list candidates currently under 5%?
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #4 on: April 11, 2019, 11:28:34 AM »

Updated through the 10th. New national, NH, SC, CA, and MA polls.

It's kind of hard to tell which colors correspond to which candidates. Could you maybe list the key/legend in the order candidates are currently stacked in? Or maybe not list candidates currently under 5%?

For my own sanity, I don't think it would be wise to change the order. It messes up the colors (as colors are chosen by order, not name, in Sheets) and I would have to manually go through and edit them every time I wanted to update the chart. And for internal consistency, I would need to do that for every chart currently maintained.

However, changing the cutoff/number of candidates is a much easier task. I was using a minimum of 1% as the cutoff because that is what the Party is using to qualify for debates. However, I get why that would be unhelpful/overwhelming. I have changed it so it only includes people who have polled at 10% at least once in any state or nationally. That brings the number of candidates down to 10 and, I think, will better show how the race played out further in the cycle. Tell me what you think!
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #5 on: April 21, 2019, 01:51:16 PM »
« Edited: May 01, 2019, 04:03:38 PM by Sorenroy »

April 20 Update: New national, Iowa, and Pennsylvania polls.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #6 on: May 01, 2019, 04:05:39 PM »

April 30 Update: New national, Iowa, New Hampshire, Texas, and Wisconsin polls.
Logged
Grassroots
Grassr00ts
Junior Chimp
*****
Posts: 6,740
United States


Political Matrix
E: 1.94, S: 2.09

P P P
Show only this user's posts in this thread
« Reply #7 on: May 01, 2019, 08:43:04 PM »

Thanks you for all this, this is all incredibly useful.

Can the mods sticky this please?
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #8 on: May 08, 2019, 12:42:39 AM »

May 7 Update: New national, Iowa, New Hampshire, South Carolina, and Virginia polls.
Logged
Senator Incitatus
AMB1996
Junior Chimp
*****
Posts: 6,508
United States


Political Matrix
E: 2.06, S: 5.74

Show only this user's posts in this thread
« Reply #9 on: May 08, 2019, 12:23:53 PM »

I think the color/cut-off update is a huge improvement. Thanks!
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #10 on: May 16, 2019, 04:08:14 PM »

May 15 Update: New national, New Hampshire, Nevada, South Carolina, Florida, Pennsylvania, and Indiana polls.

Pennsylvania becomes the fourth state to get 3 polls within 3 months, meaning it now has a chart. However, it looks a little weird with only two days on it and over 5 full months of blank space. Even on the day of the election, there will be roughly 1/3 of the chart still blank (assuming continued polls that keep a graph in place). And the issue with blank spaces on graphs will only be worse for states that still have no graph but may so in the future. As such, I am considering cutting all the charts down in size, likely to only include one year before the first vote is cast.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #11 on: May 29, 2019, 11:44:39 PM »

May 29 Update: New national, Iowa, New Hampshire, South Carolina, Michigan, and Florida polls.

As of this update, all polls (at least for the national category) are post-Biden's announcement. Which brings up something I wanted to address for people looking at these posts. One of the best things about a monthly rolling average is that it is able to have a substantial sample size of polls to work from. It also is a very good way to model how the race is going when numbers are fairly stagnant or move slowly over time. However, in a primary where there are lots of ups and downs, the strength of the model is lessened by the fact that it is so slow to adapt to new numbers. A sudden change (such as a new candidate entering the race or a strong debate performance) may not register immediately or, if that support re-subsides, may not register at all. As of now I am not changing the model. However, if this becomes a real issue (especially when candidates drop out and their support moves suddenly to their competitors) I may implement more control over the numbers (hand select polls with relevant candidates or adding in a simple weighting system to effectively remove older polls).

As always, suggestions and criticism is appreciated, especially if you have ideas for how to address shortcomings that may arise in the future.
Logged
Grassroots
Grassr00ts
Junior Chimp
*****
Posts: 6,740
United States


Political Matrix
E: 1.94, S: 2.09

P P P
Show only this user's posts in this thread
« Reply #12 on: May 30, 2019, 10:56:12 AM »

Loving this, keep it up!
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #13 on: June 12, 2019, 09:39:08 AM »

June 11 Update: New national, Iowa, Massachusetts, and Texas polls.

Additionally, Texas gains a graph and Michigan ages out one poll (bringing it down to one). Also, since I'm not posting this till the morning of June 12th and a poll has been published today, I figured I'd make a quick reminder letting y'all know that I only include polls that are published from finished days. That way I don't include one from this morning only to leave out one from this afternoon. It's mostly for my own sake, but if a update says June 11th, then it will include polls published only up to June 11th 11:59 PM EST.

Also, I overrode my previous update post by accident. It included a few corrections to messups I had made, a loss of Florida's graph do to lack of recent polling, and a slight change to methodology to exclude early-cycle hypothetical two way matchups.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #14 on: June 23, 2019, 02:05:51 PM »

June 22 Update: New national, Iowa, New Hampshire, Nevada, South Carolina, Minnesota, Texas, Virginia, Michigan, Ohio, Arizona, Florida, Wisconsin, and Pennsylvania polls.

Happy 100 national polls everyone! I decided to put off the update a few days to get the 100th poll in. Lots of stuff to see, including new three new states on the map (Minnesota, Ohio, and Arizona), several new graphs in the second post, and two age-outs for the states of Alabama and Oregon. Additionally, Warren finds herself on the map for the first time and Sanders falls off. I may end up keeping the releases on the weekend as it is a little easier on myself and it seems like more agencies are releasing later during the regular week.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #15 on: June 26, 2019, 07:13:31 PM »

June 26 Update: New national, Iowa, New Hampshire, South Carolina, Mississippi, and Florida polls.

This is a special Pre-Debate update. Normally, I wait 'till after midnight so I know that all of the polls are in for the day. However, this is a special update that includes polls done today (June 26th). Don't expect a lot of these, just wanted to put out newer numbers before our first real event of the cycle (besides announcements).
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #16 on: July 04, 2019, 07:15:54 PM »

July 4th Update: New national, Iowa, and Maine polls.

Happy fourth everyone! I decided since no-one is going to be putting out polls today and since it is an important day in American history/politics, I would put out an update for the state of the race 241 years after the beginning of the USA. Also, it is an interesting showcase of why going back a month for polls might not be the greatest idea. We had a focusing moment in the debate just about one week ago, but to have it reflect in the graphs and numbers presented here, the same pollsters who polled before the debate would have to poll again to update their numbers. As of now, only 9 of the 17 polls included to calculate the national numbers were started June 27th or later. As such, almost half of the numbers are pre-debate and reflect a different state of the race.

Additionally, one of the new polls in the numbers is from YouGov Blue. There is a very good argument to make that it really should not be given special treatment, since including it and YouGov does mess up the numbers and HarrisX/The Hill and HarrisX/Rasmussen.com are not differentiated in the model. However, as I have mentioned before, this is a project based on the 538 polling page and as long as it lists them under different names, I will count them as such.

Finally, while Maine was the only state added to the map, old polls (≥3 months) did get removed. As such, Nevada, California, and Pennsylvania all lose their graphs and Massachusetts is reduced to only a single poll.


Polls done post-debate only cover the national race and Iowa. Those are included below if you are interested.


National:
Biden — 24%
Harris — 15%
Sanders — 15%
Warren — 15%
Buttigieg — 6%
O'Rourke — 2%
Booker — 2%
Castro — 2%
Yang — 1%
Klobuchar — 1%
Gabbard — 1%
Gillibrand — 1%



Iowa (Feb 3, 2020):*
Biden — 21%
Harris — 17%
Warren — 17%
Sanders — 11%
Buttigieg — 8%
Klobuchar — 3%
Booker — 2%
Bennet — 1%
Castro — 1%
Delaney — 1%
Gabbard — 1%
O'Rourke — 1%
Yang — 1%
Bullock — 1%
de Blasio — 1%
Gillibrand — 1%
Williamson — 1%

*Based on 2 Polls.


Please note that the most recent national Morning Consult and YouGov Blue numbers have polling done on the 27th. While both firms seem to say that that day's polling was done after the debate, it does seem like something that should be mentioned in regard to their numbers.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #17 on: July 30, 2019, 07:15:18 PM »

Special Pre-Debate July 30th Update:

Unfortunately, with the Gallery functionality disabled on the Forum currently it has been a little harder to cycle in a dozen or so new weekly images. That being said, I am still keeping the documents with all the numbers up to date. As such, since we have a new debate, I wanted to get out the new national numbers, even if they are unaccompanied by an image.

National Polling:
Biden — 29%
Sanders — 16%
Warren — 14%
Harris — 11%
Buttigieg — 6%
O'Rourke — 3%
Yang — 2%
Booker — 2%
Castro — 1%
Klobuchar — 1%
Gabbard — 1%
de Blasio — 1%
Gillibrand — 1%
Hickenlooper — 1%
Steyer — 1%
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #18 on: August 29, 2019, 03:02:09 AM »

August 28 Update

It's finally done. There is way too much to be able to list out what all has changed in the over a month and a half since the last update. Gallery functionality and then computer issues meant holding off on the update until I got enough time (i.e.: a solid four hours) to actually go through and put new numbers into all the spreadsheets. Lots of new states are on the map, mostly from the South and all for Biden. As for the white states, if you don't feel like "Ctrl+F"ing for state names, Biden leads in New Hampshire and Washington while Harris leads in California. Finally, while many new states were added, there were also several age-outs of older polls. The only place where this actually removed a state was Indiana, which joins Oregon in the pool of states that have been polled before but have had no public polls started in the past three months. There are also, from memory, about two other states that face the same fate if no new polls come out for them in the next week or two.

Anyway, the impetus of this update was to snapshot where we were at on the qualifying deadline for the September debate(s).


P.S.: Fun fact regarding the EV text, it doesn't show up when you put down the candidate percentage as 10-20. If you were wondering, Washington only has 1 poll right now, not 3.
Logged
eric82oslo
Junior Chimp
*****
Posts: 5,501
Norway


Political Matrix
E: -6.00, S: -5.65

Show only this user's posts in this thread
« Reply #19 on: August 29, 2019, 01:51:46 PM »

National:

Biden — 29%
Sanders — 17%
Warren — 16%
Harris — 8%
Buttigieg — 5%
Booker — 3%
O'Rourke — 2%
Yang — 2%
Castro — 1%
Gabbard — 1%
Klobuchar — 1%

I suggest you substitute Obama and Clinton with Yang and Castro in the chart.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #20 on: September 07, 2019, 11:41:12 AM »

National:
-snip-
Biden — 29%
Sanders — 17%
Warren — 16%
Harris — 8%
Buttigieg — 5%
Booker — 3%
O'Rourke — 2%
Yang — 2%
Castro — 1%
Gabbard — 1%
Klobuchar — 1%

I suggest you substitute Obama and Clinton with Yang and Castro in the chart.

There's definitely a good reason to do that, as that would probably better show the current state of the race. However, the goal of the chart is to show people who really did, at some point, have a chance at winning delegates. As such, the requirement has been 10% in any poll in any state. Yang and Castro have never polled that high.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #21 on: September 09, 2019, 12:18:32 AM »

September 8 Update: New national, Iowa, New Hampshire, Nevada, South Carolina, Massachusetts, Texas, and Wisconsin polls.

This update represents Warren's overtaking of Sanders for second place nationally. This is the first time Sanders has not been in second place in the national polling average ever. Ironically, it also represents Warren's fall in the individual state-level polls. The new Iowa and Wisconsin polls see her lose two of her three states to Biden while her third, Minnesota, ages out. To further that irony, it also sees Sanders regain a state (New Hampshire). As of now, Biden leads in 22 of the 24 states polled within the last three months (Harris currently leads California). In addition to Minnesota, Virginia's solitary poll also aged out, as did one of North Carolina's two.

Two small corrections were also made this update. The first saw Arizona age out of the map, something that should have been done in the previous update. The second shifted a poll in the national polling graph that listed the ending date for a specific Morning Consult poll as June 24th rather than June 26th. This may be an example of human error, but more likely it is as a result of 538 retroactively updating their numbers for developing polls without listing such updates. The specific poll in question was commissioned by 538 and thus had numbers being released as they came in.

One other quick note: I normally try to release these updates every week as Saturday updates (and sometimes additional special updates for important midweek events). However, this Sunday at midnight saw the lifting of a release embargo for an A+ rated pollster. As such, I decided to shift back the update by a day to accommodate the extra poll (and some pleasantly unexpected extra additions as well).
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #22 on: September 09, 2019, 11:18:52 PM »

This is a little bit different, but now that national polling has been going on more consistently, I've been looking at a few screeners to see how the race is going without lots of bad, old, partisan, or small polls crowding out the good ones. I decided to use the harsh screening method I put forward in my 2018 polled and unpolled races thread, as well as leave the chart blank immediately when three polls by different pollsters weren't done in the past month, rather than stretch it out to three months.



Biden — 30%
Warren — 17%
Sanders — 16%
Harris — 7%
Buttigieg — 5%
Yang — 3%
O'Rourke — 2%
Booker — 2%
Gabbard — 1%
Castro — 1%
Klobuchar — 1%
de Blasio — 1%
Delaney — 1%
Ryan — 1%
Williamson — 1%


I'm gonna be keeping with the more inclusive model for updates, but it is interesting to see how little the two graphs differ. I think part of that is just cause the outliers in both models are just absorbed into the sample of multiple polls.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #23 on: September 14, 2019, 12:45:57 PM »

September 12 Update: New national, New Hampshire, and Texas polls.

Normally I try to do updates for the day of/after the updates name. Due to the lateness of Thursday's debate and my busyness on Friday I was not able to. However, I did want to put up an update with the polls that were published before the debate so there is some sort of reference to look back on once new polls start coming in. The new polls show Biden retaking the lead in New Hampshire and Sanders taking back second place in national polling. Even as Biden leads all but one state, his national lead falls below 10 points for the first time since he announced that he was running.

In addition to the new polls, Florida lost it's graph as one of it's three polls aged out. (Actually, all of the polls of that state are so old that they will be gone by the 22nd.) One of the Texas polls released this week also showed Castro at 12%, which added him to all the graphs. This also shifted the colors, so make sure to take a look at the color codes if you are confused by the change.

This update also comes with an overhaul of the national level polling graph. When going back through to make the graph of only good polls I noticed a larger set of errors that meant deleting everything and starting over. The only real changes were the addition of a Quinnipiac University poll in June and the consolidation of four March HarrisX one day samples (originally released on 538 as separate polls) into two.
Logged
Sorenroy
Jr. Member
***
Posts: 1,701
United States


Political Matrix
E: -5.55, S: -5.91

P P
Show only this user's posts in this thread
« Reply #24 on: September 20, 2019, 03:01:43 PM »

Another experimental thing, this one looking at leads over time (experimenting with the =large function). I found that including the normal range of candidates made the chart to busy, so I decided to only include those who ever averaged 10% in the given model. The first one is national. As you can see, with Sanders being in second place for almost the entire graph, there is an almost semetric relationship betweeen his and Biden's numbers (until Warren starts edging Sanders for second place).




As a bonus, here's California. It's got a good number of polls and has had a couple of different polling average leaders. For me, it makes a more interesting graph, where you don't just see one person dominating far above everyone else.



Because there is a farily good chance that this will be a one off test, I decided to make each chart independent of one-another. As such, the time scale and the included candidates are different between the two.


The regular/full update will come out for Saturday, so expect that in a few days.
Logged
Pages: [1] 2  
« previous next »
Jump to:  


Login with username, password and session length

Terms of Service - DMCA Agent and Policy - Privacy Policy and Cookies

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Page created in 0.16 seconds with 12 queries.