|
Home Page Zogby’s interactive online polling model is shining as an example of reliability, accuracy, and innovation in a field that is facing significant change. Pollster John Zogby: Zogby International’s groundbreaking new Zogby Interactive poll is the wave of the future. The strength of the results of the 2006 Midterm elections, and in the 2004 Presidential election has validated the method. Zogby Interactive accurately predicted the winner in 17 of 18 U.S. Senate races in 2006, a tumultuous year to say the least. In that one race that Zogby missed – the very tight contest in Missouri between incumbent Republican Jim Talent and Democratic challenger Claire McCaskill, Zogby Interactive was still within the margin of error for the poll. In 2004, Zogby’s online polling was on the money in 85% of the states that it polled, while, state by state, the poll was within 4 points on average. Zogby International began development of the Zogby Interactive poll 1998, and is now at the top of the industry, along with Harris Interactive. Zogby is alone in its pioneering of political polling at the statewide level. As a pioneer in the field, Zogby expects a certain degree of skepticism of this groundbreaking method. While caller ID, cell phones and rapidly falling response rates foreshadow a day where telephone polling is less preferable, Interactive polling is well on its way to becoming as accurate as telephone polling. In 2006, an internal Zogby analysis found that its interactive surveys were actually more accurate than its telephone polling. Zogby is continuing to undertake rigorous research and development of its Interactive polling, as explained in detail below, to be ready to take the polling industry into the next generation. And though Zogby’s Interactive polling is not yet perfect, its results , and Zogby anticipates even greater accuracy in the 2008 elections, and beyond. How does interactive polling work? Zogby starts with the premise that the vast majority of Americans now visit the Internet on a regular basis. Zogby does a lot of online polling work for corporate, government, non-profit, media, and other clients, as well as for the world of politics. In terms of politics, industry studies show that about 74% of all American adults and 91% of likely voters visit the Internet regularly, which makes this polling methodology particularly effective for political use. Zogby has amassed a database of respondents that numbers in the hundreds of thousands and is growing by the day. These people have agreed to take online surveys from time to time, for no compensation whatsoever. They never know when they will be invited to take a survey, nor do they know the subject. This database is constantly expanding, and Zogby technicians are constantly cleaning the database of obsolete entries and updating data points on respondents. When a survey is initiated, a random sample is drawn from this pool of potential respondents, and Zogby sends them an invitation to participate in a survey via email, which includes a link that will take them to the survey on Zogby’s own secure servers. The link expires after one use, which is just one of many security measures Zogby has in place to guarantee the veracity of the polling methodology. When needed, Zogby’s interactive polling is supplemented by telephone polling to guarantee a full and proper representation of all demographic groups. Once the desired number of respondents has completed a survey, Zogby technicians process the results, weighting each poll according to its standard practices to reflect the age, gender, geography, race, religion, political affiliation of the population a particular survey is designed to emulate. Other weights may also be applied, depending on the survey. Zogby keeps an on-going record of data points on each respondent, so that, if any glaring inconsistencies develop over a period of multiple polls, that respondent can be eliminated from the pool. But that is not the end of it. In a step that is unique in the industry, Zogby then extracts a random sample of completed interactive survey results and sends them to its own on-site call center in Upstate New York for verification. These 2% of interactive respondents are then verified by Zogby’s live telephone operators to guarantee the interactive surveys are reliably and accurately recording responses online. We believe this step is a key to the accuracy of the Zogby polling. Why use interactive polling? The way people around the world communicate is changing significantly and rather quickly, as landline telephones give way to cell phone use and Internet communications makes the postal service almost obsolete. The ability to move text, pictures, audio, and video across the Internet at lightning speed has made it the preferred mode of communications – at least for now. This much is obvious. The point of developing a reliable and accurate interactive polling methodology is to take advantage of the new way in which people are communicating, both to be more efficient and effective. The truth about polling via landline telephones is that it is getting increasingly difficult to find people at home and ready to talk – what with the now-commonplace voicemail and call-screening devices and programs in place. We find it still possible to get a good sample of respondents via landline telephones, but it is taking more time and many more calls to do so. Zogby has ruled out the conducting of polls over cell phones, for a myriad of reasons too numerous to outline here. Simply put, we do not believe it makes sense to do so. As cell phones have encroached on the use of landlines, they have not had an adverse impact on Internet usage, which makes us believe that Internet polling will have a long career in the field of public opinion research. Interactive polling has several inherent advantages as well. It is fast – in that thousands of respondents can be invited to take a survey at the same time, making it possible to conduct very fast surveys and still get large sample sizes. By comparison, a telephone survey must be done with one telephone survey interviewer talking to one respondent at a time, so there are limits on how quickly a survey can be completed. Along this same line, online surveys can also generate much larger sample sizes, which allows analysts to break out sub-groups and still have large enough samples in those sub-groups to draw meaningful conclusions. Interactive surveys can include longer, more detailed questions, because the respondent can re-read sections that they may not have understood without the embarrassment of having to ask a telephone interviewer to re-read the question to them. In this same way, online surveys can include more nuanced questions. Interactive surveys can also include audio and video for the respondents to review and react to. This is a unique benefit that no other polling methodology can match. 2008 National and State by State Results 2006 Survey Results for Gov. Sen. and House Races 2006 Results for Gov Primary in OH |