For most of this century, polls have played an increasingly important role in political decision making and news coverage. Post Polling Director Rich Morin's explained the science and art of poll-taking as today's guest on Direct Access. The transcript is below.
Welcome to washingtonpost.com's Direct Access. We will be live for the next half hour with Post Polling Director Rich Morin. Welcome all.
Edmond, Okla.: How can a sampling of 500-1,000 people accurate reflect the opinion of 250 million Americans?
Rich Morin: Like a professor of mine once said, it's math-a-magic! But here's the real answer. Just as you can sip a teaspoon of tomato soup and know what the whole pot tastes like, you can interview relatively small numbers of people to characterize the attitudes of the whole population. The key is obtaining a random sample of the population. Generally speaking, what pollsters must do is insure that every adult in the population has an equal chance of being selected in your sample. One of polling's dirty little secrets is that most of us do telephone polls, which means that people who don't have telephones (roughly 3-5 percent of the population) have no chance of being represented in our samples. Good enough for election polls, but unsatisfactory if we're interested in measuring the attitudes of the poor.
Herndon, Va.: Like millions of others, I'm fascinated by the ability of scientific polling to give us quick access to the thinking of the general public. However, I'm also a believer in the sanctity of free choice. If you can predict how I'll vote or what I'll think in advance of the fact, what has happened to my freedom?
Rich Morin: I like to think polls give you more voice, not less, in the political process. Before polls, we were at the mercy of "talking guts" politicians and pundits who said their "gut" told them what the people were thinking. Now we have reasonably reliable ways of accurately measuring what people really think, which often is very different than what the self-appointed experts claim it is.
Also, polls are just snapshots in time. Polls are great at characterizing current attitudes and beliefs; their track record as crystal balls is far less enviable. As events change, attitudes change as well and you're free to change your mind. This process of attitude formation and change keeps me in business.
Held over by popular demand! Rich Morin has agreed to stay online with us until 1 p.m.
Cleveland, Ohio: How come you never hear the income, educational level, and whether the person polled is a registered voter when you hear polling results? You often see race and sex. Is it because they are not asked or are the stats just not broken out for purposes of time? or PC'ness?
Rich Morin: We pollsters can be as PC as the next, er, person. Most polls do ask questions about income and education level, and many stories do break down results by income, education, marital status and other important variables.
The big problem is time: Here at the Post, we often get survey results back at noon and have the completed story to our editors at 5:30. That doesn't give much time to spend much time with looking inside the survey results.
Space is another problem. We often just don't have enough room to include much beyond the overall results in our stories. I'm amazed that my friends who do polls for the television networks spend so much money (a 1,000-sample poll costs upwards of $25,000 or more depending on length) for a scattering of results that might fill 20 seconds of air time.
Bottom line: We media pollsters remain data rich, and analysis poor.
Philadelphia, Pa.: In this morning's story, the writers use this wording to tackle the relationship of the economy to the President's status: 'Despite recent turmoil in the stock market and the international economy, three out of four Americans approve of Clinton's handling of the economy.'
My question is couldn't it be the other way around, in other words, that BECAUSE there is so much instability elsewhere there is a strong inclination not to court the inevitable disruption of a protracted impeachment process? Furthermore, is it possible that the greater the instability elsewhere, the firmer the support for Clinton to continue?
Rich Morin: Excellent point. There's more evidence for your hypothesis is elsewhere in the new Post-ABC News poll. Among those who don't want Clinton impeached, half say it's because they don't think he did anything serious enough to warrant impeachment. But the other half say they don't want to put the country through the trauma of the impeachment process, irregardless of what he's done.
Paeonian Springs, Va.: How difficult is it for a pollster to separate the outcomes that they would prefer (or that their sponsors would prefer) from the outcomes that represent a valid cross-section of the population? Doesn't the former motivation often drive some pollsters to influence the outcome by leading questions or by questioning individuals who seem to respond a certain way?
Rich Morin: Pollsters like to talk about "margin of sampling error." I believe that more polls are undone by margin of thinking error, and that has everything to do with question wording, order effects and a host of other gremlins that undo even the most rigorous pollster.
Yes, I am certain some private pollsters do shade the results to please a client, either by errors of commission writing biased questions or by errors of omission, such as merely reporting the results that a client wants to hear and ignoring those that send a disturbing message.
But I'll also guess that such pollsters wouldn't be in business long. Most candidates and businesses who spend the money to do a real poll want to hear the truth so they can make informed decisions. If you want to be flattered, call mom.
Media pollsters, even if we weren't naturally such an honest lot, have even less incentive to fudge the numbers. There are so many good polls out that those bad numbers would soon be spotted, and our employers would be asking us some tough questions. That's why I'm comforted, not dismayed, by the increasing number of polls.
Belmond, Iowa: Do you think that pollsters are obscuring the real issues that voters want addressed during campaigns? They always seem to try and influence the debate during campaigns with their polling on "side" issues. What do you think?
Rich Morin: I think we work hard to identify the issues important to voters, and by and large I think we do a good job. The problem is that what may be important to people may not be important to political candidates, who may choose to emphasize an issue that gives them strategic advantage over an opponent over one that is viewed as more important by more people.
Also, I look at my polls and other media polls. We're consistently showing that people care most about education, the economy, values and morals all mainstream issues.
Boston, Mass.: Who are these people being polled? Is it true that the polls are taken during the hours of 8 a.m. and 5:00 p.m.? If this is true, are only soccer moms and the chronically unemployed being polled?
Rich Morin: Yipes! Anyone who polls only between 8 a.m. and 5 p.m. is looking for trouble and may soon join the ranks of the chronically unemployed. We poll in the evenings between 5:30 and 9:30 on weekdays, and in the afternoons and evenings on weekends so as to get a random sample of the population, including soccer dads and workaholics.
That's all for today's Direct Access. Thanks to our guest Rich Morin and to our audience for their questions. Tune in to washingtonpost.com's next online chat with Child behaviorist Judith Rich Harris.
© Copyright 1998 The Washington Post Company