Frequently Asked Questions (Updated 2008)

Answers, as Stated by John Zogby


Who gets called for these polls? The people in my circle and I have never been called!

How can you get accuracy with such a small number of people polled?

What is a "margin of error"?

What political party is Zogby International affiliated with?

When you poll the U.S. where exactly do you call?

If I give you my phone number will you put me on your "calling list"?

What makes Zogby International's answers so different from other polls?

The so-called "Bradley Effect", the idea that some white voters are reluctant to say they support a white candidate over a black one, was first raised after the Democratic primary in New Hampshire. According to your experience in polling the primary election, to what extent does the race factor affect the reliability of polling result? Can you give me a few examples?

If you think that there is such danger that the reliability of polling result would be undermined, do you have any precautions to ensure a reliable result?

Compared with previous presidential election, what are the main characteristics of this year's poll trend?

Based on the polls on candidates and election-related issues so far, can you tell us who has better chance to win in November? What would be the major factors (or voter groups) affecting the ultimate outcome?

John Zogby: "Phone numbers are chosen purely by random, ensuring that every household in the US (or wherever we are polling) have as much chance of being selected as any other. With tens of millions of adults in the US, it is still rather easy to be missed - but stay in there and maybe stay home more often. You probably have more of a chance of being called than having a visit by Ed McMahon."

How can polls be so accurate when you only ask such a small number of people?

John Zogby: "It's pure probability and statistics. The same theory is involved as when you take a blood test and the clinician draws only a small sample rather than draining all the blood out of your body."

What is a "margin of error"?

John Zogby: "The margin of sampling error means that if we do the same poll 100 times, in 95 cases out of 100, we will get the same results plus or minus a certain percentage. Generally, a sample of 400 gives you an MOE of +/-5%, 600 gets you +/-4%, 1000 gets +/-3%. We need to be sure our sampling is random and scientific. It all relates to probability and statistics. If there are a million marbles in a jar, some black and some white - how best do I determine the number of each color, short of counting every marble? If I draw 1000 out at random, chances are I will get the same numbers of each 95 times out of 100 within a margin of error of +/-3%. However, there are other sources of survey error - e.g. how questions are phrased, etc."

I've heard so many conflicting stories - is Zogby International a Republican pollster, a Democratic pollster or an Independent pollster?

John Zogby: "We are independent and nonpartisan. I am personally a Democrat, but the firm does a lot of work for media (like Reuters America, New York Post, St. Louis Post Dispatch, etc.) and we work for both parties."

Where do you poll for Zogby America? What regions, area codes or time zones?

John Zogby: "If we are polling the US, we poll from a sample drawn from all households with telephones in 48 states. We, like others, do not poll Hawaii or Alaska because of time differences and because Republican Alaska cancels out Democratic Hawaii. As well, out of a sample of 1000 likely voters there would only be a total of 1 from both states (combined)."

Can I be put on a list so that you can call me for a future poll?

John Zogby: "I'm sorry but our polls must be random. As soon as we "place" people in our pool of respondents we run the risk of skewing results."

I read a lot of polls and yours is so different from the others - what makes your answers so different (and accurate)?

John Zogby: "We poll only likely voters who are different from just all adults. In addition, we poll all day long - 9am to 9pm local time (to the region we're calling). Finally, we apply weighting for party identification to ensure that there is no built-in Democratic bias in our sampling."

The so-called "Bradley Effect", the idea that some white voters are reluctant to say they support a white candidate over a black one, was first raised after the Democratic primary in New Hampshire. According to your experience in polling the primary election, to what extent does the race factor affect the reliability of polling result? Can you give me a few examples?

John Zogby:"I believe that the Bradley effect may in fact be a thing of the past. Number one, Harold Ford, Jr., an African American who ran for the Senate in 2006 in Tennessee, actually got more votes on Election Day than he had received in pre-election polls. Secondly, during the primary season, while polls may have been right or wrong, they generally seemed to get Barack Obama's support. Where most of the polls went wrong were last minute breaks for Hillary Clinton. Thirdly, as we began to look into the race issue in earnest, starting in Ohio, we've found that people have been forthright in saying that they will not vote for an African American candidate. On a professional level, it means that we can do good research, but on a personal level I wish they would go back in the closet."

If you think that there is such danger that the reliability of polling result would be undermined, do you have any precautions to ensure a reliable result?

John Zogby:"I believe the general election will be easier to poll than the primary election. First of all they normally are easier anyway. There is a more predictable model in general elections and the reliability of polls in general elections in the past have been quite high. Primary electorates revolve around get out the vote operations and the momentum of candidates, especially this past year when the schedule was so compressed that previous historic models weren't sufficient."

Compared with previous presidential election, what are the main characteristics of this year's poll trend?

John Zogby:"Compared with previous presidential elections this is a more typical election that revolves around centrist voters, compared to 2004 which was much more hyper-partisan with a very small percentage of undecideds who remained undecided until the last minute. This election will be fought in the middle as is normal. We're also looking at heightened turnout from African Americans, Hispanics, and young voters."

Based on the polls on candidates and election-related issues so far, can you tell us who has better chance to win in November? What would be the major factors (or voter groups) affecting the ultimate outcome?

John Zogby:"In this instance, polls don't predict, they are like taking a photograph. As we do more and more photographs it becomes a moving picture, and we might suggest where the trend is leading. But with the rise of so many independent voters, we have many more last-minute deciders, many of whom make their final decision on Election Day. My hunch is that this election will stay close until the very end, but then it will break big one way or the other. I think the burden is on Sen. Obama. If he can persuade the electorate that he represents leadership for the future, he has the potential to win in a landside. So many states are close and can go either way, including some that have previously not been competitive. On the other hand, if Obama cannot make the case, Americans will probably rely on the man that they know better, and in that sense McCain can win in a landslide. This is the scenario that occurred in 1980 when Ronald Reagan faced Jimmy Carter. Once Americans were comfortable with Ronald Reagan, he won in a landslide."

RETURN TO PREVIOUS PAGE