Metro Giant
Navigation Bar
Navigation Bar

  • "Levey Live" archives

  •   Q&A with Rich Morin

    Tuesday, February 16, 1999

    Rich Morin
    Rich Morin

    "Levey Live," appears each Tuesday from 12 to 1 p.m. Eastern time. It is a live, moderated discussion offering users the chance to ask questions directly of the people who make the news and the people who report it. Your host is Washington Post columnist Bob Levey.

    Bob Levey
    Bob Levey
    Todd Cross/The Post

    Bob's guest this afternoon was Washington Post polling director Rich Morin. Before joining The Post in 1987, Morin spent five years as survey and research editor at the Miami Herald and was an assistant city editor, investigative reporter and urban affairs reporter. He also worked as an investigative reporter and night police reporter for the Arizona Republic.

    Morin's professional interests include the use of quantitative and qualitative research techniques in reporting. You can read Morin's columns and learn more about polling in general with a visit to's Politics section.

    Here is a transcript of today's discussion:

    Warrenton, VA: Since the polls indicate that the public agrees that Clinton is guilty of the crimes, but that he should remain in office, what does that tell us in your opinion? That our society is going to hell or it only matters in a recession? Or perhaps both!?!

    Rich Morin: Good question. I think most Americans agree that Clinton's behavior was wrong but not ghastly enough to warrant overturning the 1996 election. Essentially, the public wants to protect the presidency, not necessarily this president.

    Bob Levey: What about the phenomenon of pollsters who work for a candidate, and who always seem to serve up results that the candidate wants to hear?

    Rich Morin: That pollster would be looking for work, very quickly. Most politicians want their pollsters to tell them the bad news--but don't want them to repeat it to reporters.

    Springfield, VA: In conversations with friends, family and co-workers in the DC area and around the country, I find that most people I come in contact with (liberal and conservative) do not reflect the polls in the Clinton affair. Do you believe that polls really reflect the mood of the nation or do people have a tendency to follow the "trend" in responding to polls that is not their true belief?

    Rich Morin: I believe that people tell the truth to pollsters most of the time. I have been pleasantly surprised how stable public opinion has been on the Clinton scandal, which I take to be a sign that we have been measuring something that's real. The "trend" that you suggest people are following is, I believe, a real reflection of public sentiment.

    Bob Levey: You always say that your polls are based on "random telephone surveys." How does it actually work? Do you pick up the phone and dial the first ten digits that jump into your head? What if you reach an answering machine? What if you reach a child? What if someone hangs up on you?

    Rich Morin: We use a technique called random digit dialing to insure that everyone with a household telephone has a chance to be contacted in our surveys. What we do is append a computer-generated random four-digit number to working area codes and telephone prefixes (the first three digits of your phone number). That means we can get people with unlisted phones in our sample. Then we ask to speak to the person in the household who last celebrated a birthday. If we just talked to the person who answered the phone, we would disproportionately interview women, since women are more likely than men to answer the phone in most households.

    Arlington VA: For diehards such as I, the aftermath of the impeachment has left me with a true disdain for the Republican party and their conduct during the process. Now, I am left to deal with my residual anger on my own, and I would like to address those feelings in a constructive manner.

    My chief political goal, as a wife and a mother of a young teen, with no ties to any organized political identities, is to eradicate the nation's House of those who unleashed this plague upon us. Through the efforts of James Carville and Stan Greenberg, and their Democracy Corp., I will be an avid participant in channeling funds and launching awareness campaigns to defeat the Republican incumbents in close districts.

    My question is, has anyone done a survey and/or focus group on the effectiveness of such an organized, grass-roots endeavor? How well-received will this attempt be? [edited for space]

    Rich Morin: Polling suggests that people resent "outsiders" coming into their communities and attempting to influence local races. That means truly grassroots efforts by voters in their own districts would be helpful but attempts by national organizations might backfire. So whether you're a Republican or Democrat, the advice is, stay in your own back yard.

    Washington, DC: If I am to believe that polls are accurate enough to, say, predict who is going to win the next presidential election, then won't knowing this either (a) make me give up on voting against a sure thing, or (b) vote to help tilt the balance against a candidate I dislike?

    And by the way, who IS going to win the race for presidency? Or is it too early to tell?

    Rich Morin: Yipes, I'm a pollster, not a prophet. The 2000 race aside, I think that polls have very little influence on voting behavior. People who would be dissuaded from voting because their candidate leads in the polls are the kind of people who simply don't vote. Besides, polls are not always accurate. So don't give up on voting, regardless of what the pre-election polls say.

    Arlington VA: My husband works for a political polling firm, and he's constantly referring to "weighting" the polls. Once he gave me an example of how this is done, but it seems to me that, if there is a lack of integrity and/or severe desire to contort the numbers in a favorable light, there does exist an opportunity for foul play to some degree. Could you elaborate on this method?

    Rich Morin: Weighting refers to the minor statistical adjustments that pollsters make to their samples to make them conform to census figures. For example, my national sample of 1,000 randomly selected adults may be 50 percent female and 50 percent male, whereas the census says the country actually is 52 percent female and 48 percent male. I would "weight" (adjust) my sample to census figures.
    Yes, there is always the possibility that renegade pollsters can manipulate the data via weighting and by other means. But they wouldn't last long doing media polls--their editors would quickly ask why their results were so different from those of their competitors.

    Washington, DC: Does you work in polling give you any thoughts about whether Americans are swinging to the right or left? And is the Christian Right growing in power, diminishing or holding the same pattern?

    Rich Morin: America is lurching to the middle. Both the extreme right and the extreme left (if you can still find any hard-core lefties!) are in disfavor with voters. Moderation in all things is the rule in politics these days. The proportion of Americans who identify with the Religious Right remains unchanged. These voters will continue to exert tremendous influence in the Republican Party, and remain a key group to watch in the 2000 presidential race.

    Washington, DC: I have seen polls stating that Americans in significant numbers believe the following: that we are being regularly visited by extraterrestrials, that ghosts walk among us, as do angels and that Elvis is actually in hiding. One poll, quoted in USA Today, said that two percent of the respondents also believed that the Magic 8 Ball could predict the future.

    My question: Are Americans getting dumber as we speak or are these polls skewed in some way?

    Rich Morin: Americans are right about extraterrestrials, wrong about ghosts and angels and who knows about Elvis? Actually, I love these contrarian results. Just goes to show that America is a truly wacky country--and that in the population are more than enough people who believe just about anything.
    Also, don't discount our sense of humor: If somebody asked me a silly poll question about Elvis, I just might want to give them a silly answer.

    Bob Levey: Half an hour remaining with our guest, The Washington Post's director of polling, Richard Morin.

    Bob Levey: In a Washington Post article last month, you noted that different words in a polling question can lead to widely different results. So how do you formulate questions, knowing that there not be any such thing as a neutral way to say something?

    Rich Morin: Polling is a mix of art and science. The science is well established. The "art" part is what give pollsters fits. Question wording is a bigger source of error in polls than the more frequently described "margin of sampling error." In order for a poll question to be unbiased, it must offer respondents both alternatives, asking, for example, whether they "approve" or "disapprove" of the job Bill Clinton is doing as president. In tests of question wording we've done at The Post, simply asking people whether they "approve" of the job Clinton is doing produces about a 5 percentage-point increase in his job approval rating over what it would be if you offered respondents the "approve" or "disapprove" alternatives. In polling as in many other things, god is in the details.

    Falls Church VA: "The margin of sampling error for the overall results is plus or minus 3 percentage points."

    I would like for you to address the margin of error that exists with every poll. The difference between +/- 4% or +/- 3% seems significant, yet all polls are taken at face value (unless you read the oftentimes "fine print" disclaiming the margin of error). [edited for space]

    Rich Morin: Very good question. Let me first disclose a prejudice. I believe that more polls are undone by "margin of thinking error" than by margin of sampling error. In other words, I think we pay too much attention to sampling error when reporting our results.
    But to your question. A poll result is an estimate, the best estimate based on our survey. To report every survey result not as a "point" estimate--56 percent--but rather as a range--"between 53 percent and 59 percent"--would be awkward (but perhaps more accurate). I do, however, agree with the thrust of your question: We do report results in a way that suggests they are more accurate than they probably are.

    Bob Levey: So many people say, "I don't trust polls because I've never been polled." But doesn't statistical theory account for this?

    Rich Morin: People always ask me this question, and my answer is: Stick close to the phone, we'll get to you eventually. There are about 180 million adults in this country, so the chances of being interviewed is very, very small. Last year, The Post alone interviewed about 25,000 people for its surveys, so the chances that you'll be contacted sometime are good.

    Washington DC: This may seem like an 'obvious' question but I'm seriously interested in what the inherent point/purpose of conducting polls is? What do we hope to gain from a poll?

    Rich Morin: That's an important question. I believe public opinion polls are vitally important in a democracy because they offer the people's voice on important issues. I'm very skeptical of politicians and pundits who claim to speak on behalf of the people. Polls allow the people to speak for themselves. That's why The Washington Post sponsors surveys--to free ourselves from conventional wisdom and find out what people really are thinking.

    Washington, DC: What is your work day like? Are you constantly involved in conducting one poll or another or are you also involved in other projects? (You also write a column for the Post, right?)

    Rich Morin: My day, like yours, is hectic. I write two columns for The Post (one on public opinion, another on new research in the social sciences for the Outlook section). But my real job is directing public opinion polls. At any given moment, we likely will have a poll "in the field" and another one in the works. I also serve as a sounding board to newsroom colleagues who want to know if a poll they've just received is reliable or newsworthy. I also try to get out and about the newsroom to find out what people here think the next big story will be, and try to have a poll ready when that story breaks.

    Springfield, VA: In reference to your earlier response on "outsiders", what do the polls suggest regarding Hillary Clinton's possible attempt to run for Senator in New York in 2000? I would not like it if a candidate moved into my state just to run for the Senate (or any other office).

    Rich Morin: Hillary Clinton is far more popular than her husband, according to our latest poll. Whether that popularity translates into a Senate seat remains to be seen. The biggest obstacle she must overcome is the perception that she's an outsider. But New York may be the perfect place for her--more tolerant of her outsider status than other states, and more receptive to her political views.

    Kensington, MD: I keep reading that nobody knows how many people are really using the Internet and the Web. Why can't polls give us a better idea of the actual numbers of Net users in the U.S.?

    Rich Morin: Actually, there have been a number of recent surveys on internet usage. They show that about half the country has signed onto the 'net sometime in the past year. Surprisingly, 'net users are looking more and more like the country as a whole: The proportion of women, African Americans and older people who are signing on has increased dramatically in just the past two years.

    Bob Levey: You recently reported that 53 percent of people you polled do not know how their representative voted on impeachment. If so many people are so deeply asleep, what does polling really tell us?

    Rich Morin: The simple answer may be the correct one: Many Americans just didn't care deeply about the scandal and tuned out the impeachment and trial.

    Bob Levey: During the impeachment circus, Rep. Lindsey Graham raised an interesting question: What would the Founding Fathers have said about polls? He never really answered it, so it's your turn, Rich.... Would George Washington have decided whether to wage war against the British because of polls? Would Congress have moved from Philadelphia to D.C. if polls hated the idea? And so on.....

    Rich Morin: I'd like to think that George would be pleased, and I don't think he would have used polls to wage war. The short history of polling suggests that politicians who use polls to make decisions are making a huge mistake. Public opinion is reactive: People respond to events, policies, events. So my guess is that the Founding Fathers would have acted--and then polled--and our history would not be any different.
    Actually, there were things that pretended to be polls hundreds of years ago. One early version was called a "Whiskey poll." Politicians would announce that every voter who dropped by a certain place at a certain time would get a free drink of whiskey. At the appointed time, the voters who showed up were asked who they planned to vote for, and were given a drink. We've improved our methods somewhat since then. Polls are better today, though perhaps not as much fun.

    Bob Levey: That's it for today. Many thanks to our guest, Richard Morin. Be sure to join us on Feb. 19 from 1 to 2 p.m. Eastern time for the Friday edition of our show. It's called "Levey Live: Speaking Freely." The format is anything-goes. Next Tuesday, from noon to 1 p.m. Eastern time, my guest will be the new president of The University of Maryland, C. D. (Dan) Mote.

    © 1999 The Washington Post Company

    Navigation Bar
    Navigation Bar
    yellow pages
    Marketplace NextCard Internet Visa - Apply Now
    Archives Search Help! Home Politics Classifieds Sports Style News