Big Lizards:Blog:Entry “How to Watch the Election and Know What's Going On”

« ...And the Elite Media Are NOT the "Fat Ladies" We Mean | Main | Another Caution: Ignore Projections Based Upon Exit Polling »

October 31, 2008

How to Watch the Election and Know What's Going On

Hatched by Movie Badger

Note from the Mgt: Movie Badger is a new occasional contributor to Big Lizards. It is not any of the normal contributors. It is involved in the entertainment industry, thus prefers anonymity to avoid job complications... since it emphatically is not a liberal.

Watching election night coverage can be confusing and frustrating, because reporters rarely give you a complete picture of what's going on. They want you to keep watching and think every development is of utmost importance; clarity often gets sacrificed for this end.

This is your guide of what to look for and how to determine the results of the election before the news is willing to tell you.

First, I urge you to download the Election Tracker I put together:

Right-click the link;

Select save link or save target from the context menu that opens.

This Microsoft Excel file will combine results with expected results from safe states, and give you predicted vote totals. There are lots of websites that let you do the same thing with maps, but they'll all be slammed on election night; this tracker will actually be saved on your local system, so there will be no internet-induced delays. (Plus my interface is quicker.)

If you have trouble getting the tracker to work, please leave a comment and I will try to help.

Will the election be close?

Polls range from showing a statistical dead heat to Obama with an insurmountable lead. Which ones are right, if any?

The answer is that nobody really knows. Built into every poll are the pollsters' own assumptions about voter turn-out. The pollsters all predict huge Democratic turn-out -- far more than in previous elections; but these assumptions are little more than educated guesses. They may turn out to be right, since the pollsters are the experts about this sort of thing. But there's no science behind them, and they could just as easily turn out to be wrong. Treat these predictions as having the same degree of veracity as an expert sportswriter's opinion on who's going to win the Superbowl.

There are some indications that these turn-out guesses could be wrong: In early voting, the split has been pretty even between Democrats and Republicans, and both campaigns (who have access to much more accurate polls than the public ones) are acting as if the election is close. But then, early voters aren't necessarily representative of voters in general; and strategically, it makes sense for both campaigns to act as if the election will be close, even if they think this only has a small likelihood of being the case. If the election's a blowout, nothing they do will matter at this point; so they might as well focus exclusively on the possibility that it won't be.

This analysis will allow you to determine relatively quickly whether the election's a blowout or not. If it is, you'll be able to turn off the TV at 8:30 Eastern Time, knowing that Obama's the winner. If it's close, you'll know what to look for to figure out the winner before anyone else.

Safe states

Each nominee has a bunch of safe states they are almost certain to win. The media may try to hype results from these as if they're news, but they're not. You can just check them off the list. On the other hand, if a nominee loses any of his safe states, that's huge news; that can only happen if he is losing horribly overall. If any of these states switch sides, the election is over, and you can turn off the coverage.

McCain will almost certainly win the following 21 states:

Alabama, Alaska, Arizona, Arkansas, Georgia, Idaho, Kansas, Kentucky, Louisiana, Mississippi, Montana, Nebraska, North Dakota, Oklahoma, South Carolina, South Dakota, Tennessee, Texas, Utah, West Virginia, Wyoming.

Obama will almost certainly win the following 16 states:

California, Connecticut, DC, Delaware, Hawaii, Illinois, Iowa, Maine*, Maryland, Massachusetts, New Jersey, New York, Oregon, Rhode Island, Vermont, Washington.

With these safe states, Obama has 197 electoral votes, while McCain has 163. (It takes 270 to win, but the nominees can theoretically tie at 269 each; see below for that possibility.)

In addition to safe states that each nominee will certainly win, each nominee has states that he must win; if he loses one of these, you can count the election over. (Unless the nominees start trading must-win states; but with a few exceptions, that's not very likely. Overall opinion will lean one way or the other, which ought to have a similar impact on each state's race.)

I'll take these by region, in order of when the polls close.

The East

Obama must win: Pennsylvania
McCain must win: Florida, North Carolina, Virginia*
Toss-ups: New Hampshire, one electoral vote of Maine.

If both nominees win their must-win states, then it's a close election. In that case, New Hampshire will be a good bellwether: Its four electoral votes may not determine the election, but should give you a good idea of which way the toss-up states are leaning. And there are a lot of scenarios in which those four votes could make the difference.

Maine is a safe state for Obama; but Maine and Nebraska have a different system for allocating electoral votes than all other states (which are winner-take-all). In Maine and Nebraska, two electors are chosen by the statewide vote total, but the rest are allocated district by district. This won't matter for Nebraska, which will be solidly McCain; but Maine may not be solidly Obama: One of its two districts -- therefore one of its electors -- might go for McCain. Like New Hampshire, this is a bellwether that has a slight chance of determining the election.

I put an asterisk by Virginia because it's not quite a must-win for McCain; if McCain loses Virginia but wins New Hampshire, there are still some realistic but less likely paths to victory for him. He'd either have to pick off one of Obama's must-win states, or win every toss-up. On the other hand, Obama winning Virginia and McCain winning Pennsylvania is one of the realistic swaps of must-wins; in that case, Obama's the one who's in a lot of trouble. He would have to either pick off another McCain must-win or else sweep all the toss-ups. (If McCain loses Virginia, Pennsylvania, and New Hampshire, the election is over.)

The center

Obama must win: Michigan, Minnesota, Wisconsin
McCain must win: Indiana, Missouri, Ohio

No toss-ups or complications here. Just a simple opportunity to call the election over if one of the nominees loses a must-win state.

Mountain and Pacific states

There are no must-win swing states in these areas. Only safe states and toss-ups.

Toss-ups: Colorado, Nevada, New Mexico

Putting the math together

If each nominee wins his safe and must-win states, McCain would be leading by 260 electoral votes to Obama's 254, with 24 votes up for grabs: 4 for New Hampshire, 1 for the swing district of Maine, 9 for Colorado, 5 for New Mexico, and 5 for Nevada. As the toss-up states are announced, you can add up these numbers. McCain would need 10 points worth of toss-ups to win, while Obama would need 16.

Adjust these as necessary for any wackiness. For example, if they swap Pennsylvania and Virginia, that nets McCain 8 votes; he would only need 2 more from the toss-ups, whereas Obama would need all 24 available votes. By contrast if McCain loses Virginia and Obama holds Pennsylvania, McCain will need to win 23 of the 24 toss-up votes available, while Obama would just need to win 3.

What if there's a tie?

If it's tied 269 to 269, the first thing to worry about is a faithless elector: If someone votes differently from how his state voted, and if that ends up being determinative, it will cause a constitutional crisis that will make the 2000 debacle look like peanuts. Electors are generally chosen because they're party faithfuls -- but even the most partisan of partisan whores can be bribed or blackmailed.

Assuming we get past that minefield, and every elector votes the way he's supposed to, the election will be decided by the newly elected House of Representatives, with each state getting one vote; Democrats currently have a slim majority of delegations, which they will probably, but not definitely, hold in the election... so a tie means that Obama will probably but not definitely win. The vice president is chosen by the Senate using ordinary voting rules, and the Democrats will certainly keep their majority; so Joe Biden would definitely be selected VP. There is an outside chance that we could end up with McCain as President and Biden as VP, which would be silly.

How to know the results early

The media will know more than they're telling you. They wait until polls are closed before giving out results, and will err on the side of caution to avoid a repeat of the 2000 Florida debacle. [Or not; they've been awfully much in the tank this year. -- Dafydd.] But you should be able to get some idea of what's going on just by looking at the news anchors. If you see a lot of happy faces, that's good news for Obama. A lot of sad faces in the media is good news for McCain. (Reverse that if you're watching Fox News.)

Also, as one nominee nears the 270 electoral-vote threshold, the media will start getting more and more reluctant to call states that put a nominee over the top. But if you look at different channels, you'll see that they are calling different states. If any one channel is confident enough to call a state, they're probably right -- unless there's a freak mishap like in Florida in 2000 when some flunkie in charge of compiling exit poll data mistyped the results.) If you add up the states that different channels are calling, you may find that one of the nominees has enough to win. In 2004, I used this method to figure out the winner a half-hour before any channel was willing to declare it.

Lastly, use the tracker I made. If it's a really close election, they may not be willing to call it until the last polls close in Hawaii and the Aleutian Islands. But except for Nevada, everything in the Pacific time zone and points west is safe for one nominee or the other, so by that point the tracker should be making it obvious who's going to win.

Hatched by Movie Badger on this day, October 31, 2008, at the time of 1:18 PM

Trackback Pings

TrackBack URL for this hissing: http://biglizards.net/mt3.36/earendiltrack.cgi/3314

Comments

The following hissed in response by: Dishman

My source whispered to me today that the final tallies were McCain 51.04%, Obama 47.53%. The word "were" seems a bit strange, but there ya go.

We'll know by the last two digits of each if my source is actually credible.

I also posted this at DeanEsmay.com.

The above hissed in response by: Dishman at October 31, 2008 3:19 PM

The following hissed in response by: thehix

Great tool I look forward to using it on election night.

The above hissed in response by: thehix at October 31, 2008 4:28 PM

The following hissed in response by: thehix

Great tool and helpful hints for following the electoral college totals on Election Night. Can't wait to use it. I expect many surprises on Tuesday - I have NO faith in the polls this year.

The above hissed in response by: thehix at October 31, 2008 4:59 PM

The following hissed in response by: Jaded

What a EXCELLENT tool and thank you. I will spread this around for people to use....but DEEP in my heart I believe this country will do the RIGHT thing and elect McCain/Palin......I know the Country First Americans did not let me down in 04.

The above hissed in response by: Jaded at October 31, 2008 6:36 PM

The following hissed in response by: Dick E

Welcome aboard, Movie Badger. I gather you don’t mind being objectified by being referred to as “it”, rather than the far more chic “(s)he”.

But there's no science behind them, and they could just as easily turn out to be wrong. Treat these predictions as having the same degree of veracity as an expert sportswriter's opinion on who's going to win the Superbowl.

You should have a chat about this with Dafydd.

…both campaigns (who have access to much more accurate polls than the public ones)…

How do you know the campaigns have more accurate polls? (Specifics not required if it will pierce your anonymity.) I’m sure the campaigns have more control over the methodology of their internal polls, and more access to the details concerning assumptions and adjustments, but better accuracy?

Anyway, thanks for the tracker.

The above hissed in response by: Dick E at October 31, 2008 10:40 PM

The following hissed in response by: Dafydd ab Hugh

Dick E:

How do you know the campaigns have more accurate polls? ... I’m sure the campaigns have more control over the methodology of their internal polls, and more access to the details concerning assumptions and adjustments, but better accuracy?

The methadology, assumptions, and adjustment are the accuracy! The rest is just sophomore-level mathematics that anybody could do.

Public-opinion polling is not a science; it's an art that utilizes statistical principles. You ask how I know which polls are good and which bad; you may as well ask how I know which authors are first rate, which are honest hacks, and which are charlatans: I can point to specific characteristics of individual authors and discuss why that makes them good or bad... but I cannot give you a general heuristic you can follow to decide in each future case.

It's a fuzzy algorithm: For one thing, you must have a long enough memory to recall which polls got it right and which were way off in the polls they took shortly before the election; a poll that consistently exaggerates Democratic support and minimizes Republican support (compared to the actual vote) -- such as the Star Tribune poll in Minnesota -- is simply not reliable.

(If you remember even somewhat well, you can usually find the poll itself with some lengthy Googling; that way you can test your recollection against contemporaneous accounts.)

You're correct that you cannot say much about June polls based entirely on a November election. But when one poll is a consistent outlier, and then proves woefully inaccurate right before the election, it's reasonable to conclude that it's more likely that poll was wrong then and wrong now... than it is to conclude that the poll was right then, right in its last poll -- but the electorate lurched 15 degrees to starboard between Sunday and Tuesday.

Remember, I'm a math guy; I've been fascinated (or obsessed) with polls since I was in junior high, and I have a pretty good memory. And I'm compulsive enough to spend three hours of a fine Saturday afternoon hunting for specific polls from specific past elections, just to satisfy my curiosity.

Public polling is not a science, but neither is it a random function. It's an art -- and there are good artists and bad.

Dafydd

The above hissed in response by: Dafydd ab Hugh at October 31, 2008 11:20 PM

The following hissed in response by: Dick E

Dafydd-

Well, if we’re going to hijack Movie Badger’s thread…

The methadology, assumptions, and adjustment are the accuracy! The rest is just sophomore-level mathematics that anybody could do.

Obviously. But the campaigns’ polls are only more accurate if they have BETTER methodology, assumptions, and adjustments than others. How do we know that?

I guess I now understand something of how you decide which polls you will trust. But my question from the other thread still stands: What are the rest of us supposed to do? I mean the 99.999%+ who don’t assiduously track polls after the fact for accuracy.

The above hissed in response by: Dick E at November 1, 2008 12:03 AM

The following hissed in response by: Dafydd ab Hugh

Dick E:

Obviously. But the campaigns’ polls are only more accurate if they have BETTER methodology, assumptions, and adjustments than others. How do we know that?

Generally, public polls are commissioned by some academic or journalistic entity with an axe to grind; that is, the people paying for the poll have a political opinion, and too often it creeps into the methodology or turnout assumptions, queering the poll. The researchers know which side of the bed has the butter.

Contrariwise, working polls (not "release polls") used by campaigns have a mandate to be as brutally honest as possible -- the candidate and the campaign mangler need to know exactly where they stand -- and they generally include a range of methodologies and assumptions with corresponding results for each.

In the simplest case, were I a candidate, I would want a poll with, say, five different turnout models, ranging from great-for-me to rotten-for-me. I would try to find a campaign strategy that helped on as many different models as possible.

The best public pollsters try hard not to let their personal feelings interfere; for example, the Battleground poll is conducted by a Democratic pollster and a Republican pollster working together, so they have to agree on a methodology and turnout model.

But even then, they don't try as hard as when they're being paid gazillions of dollars by campaigns to come up with really useful poll results -- upon pain of losing a huge contract if they're wrong.

It's sort of like the fact that tools designed for do-it-yourselfers and sold through Sears are not going to be as good as more expensive tools designed for workmen (mechanics, electricians, appliance repairmen, etc) and sold through more specialized stores. It's not that Sears Craftsman couldn't make better tools if they wanted... it's that it wouldn't be cost-effective, given their target customer base.

If the NYT/CBS poll turns out to be skewed to the left, who's going to care? For that matter, who among the target audience would even notice?

Dafydd

The above hissed in response by: Dafydd ab Hugh at November 1, 2008 4:20 AM

The following hissed in response by: Xpressions

The mainstream media illuminati will make their own predictions from here on out, like they have, and America will believe them and not go out to vote.

The above hissed in response by: Xpressions at November 14, 2008 1:16 PM