National polls misleading, don’t forecast electoral vote

How could this week’s election in the country of Colombia tell us anything about the U.S. presidential election?

Simple.  Reputable polls there forecast an overwhelmingly favorable vote on the peace deal between the government and rebels.  The agreement lost by a slim margin.

The moral of the story for us is that American media focuses far too much on polling data about the presidential campaign.  In many national polls, both Democrat Hillary Clinton and Republican Donald Trump score in the forty percent range with only a small gap between them.

National polls are wrong for two reasons.  First, major defects have undermined the accuracy of the surveys themselves.  Second, the president is not elected by the popular vote they try to measure, but rather by the Electoral College.

Polls have serious flaws.  Many people refuse to answer.  That throws off the random selection of participants, essential for a valid result by which the survey predicts the actions of all voters.

Pollsters adjust responses to compensate for imbalances between women and men, Republicans and Democrats, old and young, north and south.  Some pollsters do better than others, but the results are never just right.

Of course, the questionnaire itself may be biased and the questions used by one pollster may not be the same as those used by another.  And the quality of the pollsters from one state to another is likely to vary.

Some polls are conducted by recording machines, not real human beings.  So pollsters do not know if the person responding is the person they wanted in the sample.

State-by-state polling reduces the error built into national polling, but polling defects can deny us an accurate picture of the election.

The U.S. uses the secret ballot in elections for good reason.  Voters may not want to reveal how they are voting and may even wish to misdirect others about their intentions.  There’s nothing wrong with providing a less than truthful answer to a pollster, and it is likely that some people do.

Such misdirection may also apply when panels of supposedly undecided voters are assembled to ask candidates questions or rate how they did in debates, often in what are known as “focus groups.”

There’s no way of knowing if those selected are truly undecided or campaign plants.  It’s suspicious when voters, after hearing sharp differences, say they are unmoved by a debate.

Finally, some so-called surveys are conducted among participants who select themselves.  These polls are hardly “scientific” and are meaningless as a gauge of voter sentiment.  They are often used by the candidates themselves.

The result of all these weaknesses in polling, given the great daily attention by the news media, is that voters, who may be influenced by the results, are almost certainly misinformed.

Even worse, the media and many voters are influenced by national poll results, but there’s no nationwide election.  The presidential election is conducted state-by-state.

According to the Constitution, voters don’t cast their ballots for president, but vote for Electors, people who will cast votes for president on behalf of each state.

The Electors form what is known as the Electoral College, reflecting the original constitutional compromise that recognized both population and the individual states.  Each state gets a number of electoral votes equal to its number of members of the U.S. House of Representatives plus two, the number of senators.

In 49 of the jurisdictions – 48 states plus the District of Columbia – all the electoral votes go to their popular vote winners.  It does not matter if the winner had a one-vote margin or a half-million-vote margin.

Maine has four electoral votes.  Two are allocated to the statewide winner.  Each of the other two is selected in one of the two congressional districts.  So the Maine winner may gain three or four of the state’s votes.  Nebraska, with three congressional districts, adopted the Maine plan.

The Electoral College undermines the validity of national polling results.  Clinton may win by a big margin in California and Trump by a big margin in Texas, but their excess popular votes cannot be used elsewhere.

The system works exactly as intended – to help small states count in presidential elections.  It gives Maine almost twice the weight as population alone would give it.

In Colombia and in the recent British vote on Brexit, the polls were wrong.  These were both major, national elections.  The polls may also be wrong here as well.

Most likely, it’s advisable to decide on voting without allowing yourself to be influenced by questionable polls.

Gordon L. Weil

About Gordon L. Weil

Gordon L. Weil formerly wrote for the Washington Post and other newspapers, served on the U.S. Senate and EU staffs, headed Maine state agencies and was a Harpswell selectman.