Princeton Election Consortium

A first draft of electoral history. Since 2004

Aug 06: Biden 372 EV (D+6.3% from toss-up), Senate 51 D, 49 R (D+3.9%), House control D+3.6%
Moneyball states: Senate MT KS ME, Legislatures KS TX NC

Pardon our mess…

June 9th, 2016, 4:23am by Sam Wang

A fair amount of construction is going on around here. There may be a few bumps along the way. For example, right now we’re tweaking the banner (until now it was updated manually).

Starting this summer, I am on the lookout for a new partner in running PEC, as Mark Tengi moves on from Princeton. Ideally, the person is part of the Princeton community (student or otherwise) and conversant or willing in Python, WordPress backend, and Linux. Write me!

Tags: 2016 Election

33 Comments so far ↓

  • Scott

    Sam – I’m not a Python guy, but I’d like to offer some help with content if you’re interested. I’m suggesting publishing new content in between your brilliant posts, which by necessity come sometimes days apart. I imagine the blog gets significant daily traffic – and the readers might appreciate some new content when they arrive. I’m already a moderator on the PEC Facebook page, since I helped set that up for you in 2012.



    • bks

      Please do not lower the signal-to-noise ratio of this site.

    • truedson

      The comments from informed readers is plenty in my opinion. The last thing this site needs is bloated info to get page views. That is my gripe with TPM where I use to be a paid member. Josh Marshall is a wonderful commentator. But unfortunately the site is overstuffed now in order to get people to come back and find ‘new’ material.

    • Matt McIrvin

      I actually really liked the extra content on 538 during the 2008 election cycle, when they had people reporting on election efforts in different states all over the country–but now that it’s a big-media site it’s become completely unfocused.

    • Truthy

      100% totally against anyone but Dr. Wang posting. Any insights you have can be placed in the comments and readers desperate for more have a million and one sites to read the latest pundit analysis.

  • Brian

    As long as you keep the banner in the format of an HTML “list” I’m happy. That’s what the Chrome extension is expecting.

    • Rob

      Maybe this is a change in the way the banner information s organized, but the Chrome extension seems not to be collecting updated information. Today, June 22, it is showing information for June 14.


    • Sam Wang

      The banner is organized differently, but in principle it should still be scrapable…

    • Mark Tengi

      Rob — I did inadvertently break the extension because I didn’t know it existed! Sam wanted to reorganize the banner so I changed the format and now I guess the extension can’t parse it.

      Brian — in order to allow for future flexibility, would it be OK if I give you s with unique IDs with the relevant information instead of using a list?

  • Greg Hullender

    A good way to track this or any blog that is updated infrequently is to use, which makes use of the syndication mechanism. Whatever changes you make, please don’t mess that up! :-)

  • Steven

    Am I correct that the Obama Net Job Approval/Disapproval graph just changed significantly? I had been looking at it fairly closely and I recall it showing a sharp uptick for Obama right around the March appointment of Merrick Garland. (It was fairly consistently at -4 and then moved to a fairly consistent +2.) Now it appears as a more-or-less steady increase since January.

    • Sam Wang

      That, we haven’t changed. Our display is a bit crude: we downsample the daily median by some factor – 3 I think. Therefore it shows only every third day of data going back in time. That can lead to odd changes in appearance from day to day.

      When the dust clears, I will clean up all the documentation.

  • 538 Refugee

    Does this need to be ONE person? I have Linux server experience.

  • Sean Moss

    I have linux, python, and wordpress experience but am not in Princeton. Would love to help out if you can’t find anyone that’s a better fit!

  • Ted Nadeau

    I live here in Princeton & would be eager to help / meet. I’ve done some work in : Python & Linux but not much in WordPress but have done additional relevant work in Data Visualization (Tufte), Reporting (lots), Politics, social networks, etc. Please contact me & let me know next steps. – Ted

  • Jay Sheckley

    Hope you get the help you need, Sam.
    While you’re redesigning/discussing the site’s direction: I’d like to see the header “Probability of Clinton November win” promote a link to the history and snapshot, or explanations linking to em, with the current EV MM history chart boldly at top right where the Obama approval chart stands now.
    I’ve been telling people on FB and JoeMyGod to watch your site. Because: truth. People on JoeMyGod upvoted the suggestion. An intelligent Facebook friend (to me, a real friend) responded that PEC makes her feel stupid. I replied I too don’t follow all the algorithms, but seeing em discussed by folks that get em builds confidence. Plus I found the site not only spot on during the 2008 and 2012 elections, but the only media that consistently showed the election as a steady (rather than terrifying) process. She called me names and defriended me. ::sigh:: You know how human humans are, Sam. This is all prologue to saying I wish PEC were easier for newbies. I know you care about this because I can see so much for youve done for them. But can they see it?
    Perhaps an actual button to click at top that says: NEW HERE? Leading to FAQs. Because people don’t KNOW theyre asking about the metamargin, yeah? Please forgive me, Sam, for seeming to quibble. I like and am grateful for all you do here and am thrilled to see the Median EV estimator coming into focus. You have completely changed my election experience. But the average Jane seems to live in terror. How can we ease the hysteria and get out the vote?

    • Amitabh Lath

      Jay, like your friend sometimes I also read a PEC post and feel all at sea. I usually put up a “I’m not getting it, walk me through this please” post and then Sam and the other posters will explain (and sometimes Sam will even post the MATLAB code) and then I’ll put up a “I’m still foggy on X and Y” and there will be yet more posts.

      This happened (is happening) most recently with the Realignment metrics Sam busted out in the previous post. The whole concept is yet to settle in my mind, but the PEC community has been amazingly helpful.

      There are some remarkably good explainers on this site. True, it’s tough slogging because sometimes I don’t even know what it is I don’t understand, but the end result for me has been a better understanding of the nature of measurements and uncertainties in an area completely outside my area.

    • Sam Wang

      Okay, I guess I do have to re-explain things every four years, plus describe new things. It is basically a matter of resuscitating code and essays from 2012. This will come in time.

      This year has been more work than anticipated – the need to model and track primaries caught me by surprise. And it wasn’t automated. It got tedious starting in late March, when it became statistically overwhelmingly likely that Trump and Clinton would be the nominees. Oddly, that was when traffic went up. You guys.

    • Jay Sheckley

      Looking good, Sam!
      Amitabh: Yes yes and yes.

    • DaveM

      I think traffic goes up precisely ~when~ statistical likelihoods start to become overwhelming. That’s because the gulf between that statistical reality and the punditocratic narrative becomes so large that we kids who see the emperor’s not wearing any clothes take comfort in the company of fellow clear-eyed observers.

  • Ed Wittens Cat

    Just a heads up, Dr. Wang.
    The liberal-bias/unskewing poll analysts are back.

    • Olav Grinde

      That website uses American Thinker in the singular. I wonder who they have in mind…

    • Dan Engelberg

      I linked to the site Mr. Cat. But it seems to have been hijacked by right-wing ideologues. Can you fix the link please.

  • Joseph

    Love those interactive maps, Professor! Assuming Hillary takes all the pinks, light blues, and blues, Hilary wins all but 18 states! Man, what an impact THAT would have on the down-ticket!

  • Michael thomas

    Currently, the Presidential histogram is displayed on the Senate page.

  • Hans Messersmith

    Dr. Wang, it seems that the final meta-margin/EV time series charts aren’t available on the site. I can find the final 2008 chart in a few places, but not the 2012. For example, only the map is available here:

    Is this chart still available?

  • Patricia Goff

    once again, I have to thank you, Dr. Wang, for your continued outstanding work in gathering all the crucial data about this election process and creating thoroughly fine-tuned analyses of what it all means. I am decidedly a neophyte in statistical analysis yet intelligent enough to understand as well as digest how the critical components fit together and what these figures reasonably suggest. More than ever, I am ‘glued’ to both absorbing and personally weighing the day-to-day ‘news’ and, having PEC to factor into my perspective is absolutely crucial.
    Perhaps, when you’re able to obtain the additional qualified staffing you’re needing, you might consider creating a backend ‘Understanding How Statistics Help Create a More Accurate View’ section for those of us needing to augment our understanding of how all of this comes together and what the stats/charts suggest. I’m certain I’m not alone in wanting to better educate myself to more clearly understand the extremely valuable work product you are offer here.
    Best of luck in bringing a strong assistant onboard to further improve your very important site. It truly helps to better inform all of us — at a time when every bit of data is important in voters understanding the forces affecting our not too distant lives!

Leave a Comment