Fr. Frank's perspective on polls, state initiatives and last minute duties right before Election Day 

Perspective on the Polls

Finally, let me share with you some wisdom about opinion polls.

Throughout the entire Presidential election year of 2004, there were 239 national opinion polls about the race, and in the year 2000 there were less than half that number. In 2008 there were about 800.

This explosion of polls brought about more confusion than clarity. For example, nine or ten polls were taken one day and showed margins of 3 to 10 points difference between the candidates. The next day, the polls taken showed margins from 2 to 15 points, and on the third day, the range was from 3 to 7 points.

People wonder, “Who’s right? Why are the polls so different from each other, even on the same day?”

First of all, the numbers you hear represent all different kinds of polls. There are national polls and state polls. Then there are polls of polls (for example, some numbers represent the average of all the polls taken in, for instance, the last two weeks).

Now if the race is going to be determined by a handful of key states, changes in other states will be reflected in national polls but won’t matter for the actual outcome of the election. Moreover, if there is a change in voter attitudes taking place, it will take more time for such a change to show up on the state level than in the national polls – and there are fewer of the former than the latter. Finally, if a poll is a two-week average of polls, then changes taking place today will not be as obvious as they would be in a poll that tracks day to day changes.

The sheer multiplicity of polls, moreover, also means that there is quite a disparity of methodology and expertise. There’s a big difference between a polling company with decades of experience and a university group that decided to start polling this year. And if a pollster is polling in a particular state, the poll will be better depending on the expertise the pollster has about the electorate in that particular state.

Among polling companies, moreover, one must consider the quality of the sample size, the methodology used, the presumptions about the makeup of the electorate, the internal adjustments made in the numbers, and so forth. When you hear poll results, in other words, you have to ask, “Who exactly is being counted here, and what presumptions are being made?” Are the people who were polled registered to vote? Are they “likely” voters, and how has that been defined? From what party does the polling sample come?

Generally, a poll will take 40% of its sample from each major party and then 20% independents. But we have seen some polls representing one party with 50% of its sample, and the other party 30%. Those results will be skewed. It’s very helpful for a political interest group to create a poll like this for fundraising purposes, so they can send out a letter saying, “See, we’re up by ten points; send us more money so we can stay up!” Meanwhile, the donor hasn’t read any fine print about the poll’s methodology or sample group.

And presumptions about who will be “likely” voters can be tricky. Will the registration of many new young voters, for instance, translate into those young people actually showing up at the voting booth? That remains to be seen. A poll is an educated guess about how much of that will happen.

Polls often miss the dynamics involved in turnout, such as the motivation that various groups have, or the impact that the weather on Election Day will have. The good news for pro-life people is that the difference between a likely pro-life voter and other likely voters is that the pro-life voter is more likely to vote. Past election analysis shows that the pro-life movement’s voters are more motivated than those in the pro-abortion movement. We are better at getting our voters to vote. We care more about issues. We care more about showing up at the polls. This is not something a poll will measure in advance.

In the 27 states for example that have had marriage amendments, every single poll has understated the final turnout by as much as six to eight points. In states that looked like the marriage issue might be fairly close ended up not being close at all. Even in states like Oregon where it was expected to lose, it won fairly handily.

And, of course, a key number to pay attention to when you hear polls is the number of undecided voters. By definition, they can go either way, and haven’t gone there yet. Elections are decided by the undecided, who decide on Election Day itself.

We tend to “anoint” polls, but we shouldn’t. We tend to think that they are far more scientific than they actually are and we give them far more credence than they actually deserve. But they are filled with presumptions, projections, guesses, and margins of error, if not bad sampling, erroneous methodology, and outright bias.

Again, only one thing matters in elections, and that’s turnout.

When all the attention is on the Presidential race, we have to remind ourselves to pay close attention also to other races, on every level. A President is stronger or weaker depending on the makeup of Congress. A President can sign a bill into law only if Congress puts it on his desk first. A President can veto a bill, but a strong Congress can override that veto. A President can nominate a judge, but the Senate can either confirm or reject that nominee. Congress matters, and every Congressional race deserves our attention.

Races at all levels of government matter, and voters need to do their homework to come to know the candidates.

State measures

Voters also need to pay attention to various ballot initiatives, State Constitutional Amendments, and other measures on which they will be asked to vote. Voters nationwide should talk to those they know in these states, educate them about these initiatives, and urge them to vote appropriately.