There  are 15 states with photo ID requirements for voting.  Mr. Obama lost in all  of them.  In places with the weakest controls, specifically counties in  Florida, Ohio, Colorado, and Pennsylvania, he generally drew turnouts in the 90%  or greater range and won by better than 95% of the vote.

Losers  tend to look for external explanations, and a lot of conservatives looking at  numbers like those from Florida's St. Lucie County (where Mr. Obama got 247,713  votes from only 175,554 registered voters) are starting to question the  legitimacy of the electoral results as reported.  That's not good news for  democracy, because the system works only if we trust it -- and having a majority  in the GOP write off a minority who think the results were rigged serves nobody.   Not even Democrats.

So  what we need is an independent means of testing the electoral  result.

The  traditional way of doing this is, of course, to assume legitimacy, then gather  anecdotal evidence of vote-cheating, promote that to sworn testimony through  court proceedings, and hope for a conclusion from the adversarial process this  generates.  That's how we now know, for example, that Stevens was falsely  prosecuted and Coleman beat Franken.

Great.   Except that both Franken and Begich hold office, both voted for ObamaCare,  and both will get generous federal pensions.  Basically, the traditional  approach may be effective, but it's also politically pointless...and  inappropriate in today's context anyway, given that we need something a whole  lot quicker.  We need something, in fact, that can give us a clear result  in time to decide whether there's a case to be made for asking the college of  electors to overturn the nominal result when they vote on December  17.

One  idea that might work would be to compare the results of an honest poll to the  nominal results obtained in the election and then decide what the odds are that  the differences, if any, reflect electoral fraud.

Given  that there have been hundreds of polls, the most public of which roughly  predicted the reported electoral outcomes, this may seem like a dumb idea.   But it may not be so dumb -- and for two main reasons:

First:  internal GOP polling based on face-to-face voter contact consistently  contradicted the D+7 or more results predicted by the major media polls and  subsequently demonstrated in the nominal electoral outcome.

This  shouldn't happen; theory predicts that internal polls based on face-to-face  interviews should produce better results than panel studies or public media  polls -- and until the 2008 presidential election, they generally  did.

Second:  The major media pollsters face non-response problems that render their results  indefensible in terms of normal statistical practices and standards.  In  consequence, those in charge of analyzing and reporting the data typically use  complicated, but ultimately guesstimated, formulas, giving a kind of scientific  patina to what are ultimately intuitive judgments about sub-sample weights --  and the closer the population proportions they're trying to estimate get to  50:50, the more these judgments affect the outcome, and the more risk the people  involved take when they make those judgments.

As  a result, the closer the contest is in reality, the more pressure these guys are  under to follow the leader -- for competing pollsters to recursively adjust  their weightings to invisibly move their results toward a consensus  position.

Notice  that this isn't conspiracy; it's a natural consequence of the costs and  complexities of public polling in today's business environment.  But it is  an exploitable consequence, because someone who wants to drive the public media  polling consensus toward a predetermined outcome need only lean on one of the  market leaders to cause all of them to drift toward the intended  conclusion.

Gallup  published a poll close to what the GOP internals showed, found the DOJ joining  what should have been a nuisance  lawsuit against them, and adjusted its weightings to bring its results into  line.  We know these events happened; we do not know if they're  related.

A  public, national, face-to-face voter poll, using a simple set of questions and  academically defensible statistical methods, would go a long way in clearing up  the questions here.  If the results strongly support the nominal electoral  outcomes, we can be fairly confident, for example, that the election was broadly  fair, that the GOP internal polls were wrong, and that the major media pollsters  behaved honorably and correctly throughout.

If,  on the other hand, our hypothetical national poll produces results that differ  significantly from the nominal election results, it will largely rehabilitate  the GOP internal results, cast significant doubt on the legitimacy of Mr.  Obama's claimed victory, and probably cause at least one of the major media  pollsters to rethink its methods.

Notice,  however, that this type of audit survey is a first cut at the problem -- more to  see whether there is a real problem than to address it.  Ultimately, only  the more traditional methods will let us deal with issues like those raised by  the disenfranchisement of (mostly GOP) service personnel and the enfranchisement  of (mostly Democrat) illegals.

With  that in mind, let's look at the mechanics of actually doing it -- but bear in  mind, please, that there are many different ways of doing this, and what I'm  suggesting here is intended to be illustrative, not  prescriptive.

First  we need to establish the universe: the target group whose population proportions  we want to estimate.  Since we're interested only in people whose votes  were counted and we often don't know who they are, we'll start with the list of  all registered voters -- a list we make by combining lists from all  jurisdictions and not eliminating duplicates.

Next  we need to establish the question.  Since we want to know what percentage  of registered voters voted for each presidential slate, the two core questions  are:

1.  Did you vote?

2.  For which presidential slate?

If  the nominal election results are broadly correct:

1. all  of our respondents should be reachable at the addresses listed for  them,
2. about  58% of our sample should report having voted;
3. about  30% of our sample should claim to have voted for Mr. Obama,  and
4. about  28% of our sample should claim to have voted for Mr. Romney.

Any  significant differences from that distribution will indicate fraud.  Normal  statistical methods can then be used to quantify both the likelihood and the  significance of the fraud.

Next  we need a sampling methodology -- in this case, we'll number our combined lists  from one to whatever and use the computer's pseudo-random number capability to  pick our interviewees.

The  big issue is sample size.  The determinants for this are:

1. Population  or universe size: with well over 100 million voters, the statistical rules for  infinite populations sampled without replacement  apply.
   
2. How  confident do we want to be that the population proportion estimates we produce  are in the same ballpark as the answer we would get if we talked to everybody  and tabulated those results?
   The traditional confidence level targeted by  pollsters is 95% -- meaning that if you did the sampling 20 times, you'd expect  the people drawn to correctly reflect population means 19  times.
   
3. How  close do we need to get to the real proportion?  I.e., how big a ballpark  can we live with?

The  traditional polling answer is plus or minus 5% -- largely, incidentally, because  pre-computer-age pollsters rejoiced in the happy coincidence that this precision  combined with a 95% confidence level let clients jump to the conclusion that  their ability to add the two numbers to get 100% meant they understood something  and produced an easily memorable, easily squared z-statistic that just happened  to work out to an even number.

In  our case, the basic 95% confidence that our estimate is within 5% of the real  number, given no more than 3% non-response, requires a sample of about 400;  raising that to gain 99% confidence that our estimate is correct to within 3%  requires a sample of about 1,900; and going for 99% confidence that our  estimates are within 1% of reality would require a sample size of about  17,000.

So  what we'll do is draw 17,000 names, draw 1,900 names from that list, and finally  produce an initial sub-sampling of 400.  We'll then interview those first  400, decide whether it's worth continuing, and, if so, continue to at least  1,840 successful interviews from the list of 1,900 names.  Finally, we will  re-evaluate again before either stopping or proceeding to the full  17,000.

Our  next decisions involve interview methods.  Since we cannot tolerate much  non-response and do not want the interviewer to bias the result, we're going  to:

1. cold  call on the doorstep,
2. dispatch  two interviewers on each call, and
3. ask  only two questions:

• did  you vote? and
• if  so, for which presidential slate?

Teams  will not ask the respondent to say which slate was preferred but will, instead,  hand the respondent two cards or other tokens with instructions to keep one  while placing the other in a box or other container proffered by the  interviewer.

This  process will be extremely expensive -- far more so than the telephone interviews  conducted by the major media pollsters.  As a very rough first guesstimate:  getting the infrastructure in place quickly enough for the result to be  meaningful and then carrying out the first 400 interviews will run upwards of  $400,000, with total costs rising to the $2 million range if it is necessary to  interview the full 17,000 sample.

What  we're really proposing here is a first audit of the election result, and  regardless of outcome, its primary value is in reducing  uncertainty.

Right  now, saying Obama cheated is about as credible as saying he didn't, so a  positive result will go a long way to debunking various destructive conspiracy  theories and thus contribute to the smooth functioning of American democracy.   Conversely, a negative result will form a strong basis for multiple legal  and political actions aimed at delegitimizing an illegitimate  president.

Bottom  line?  Knowing is better than not knowing.  So who's got three million  bucks?