Steve Mitchell on the accuracy of Online and Interactive Voice Response (IVR) polling:
Mitchell Interactive provides corporate, association, governmental, and media clients with scientific random sampling polling using standard operator assisted surveys, IVR (Interactive Voice Response) surveys and Online qualitative and quantitative research.
Unlike most of the polls you see on the Internet, the Mitchell Interactive Online surveys are conducted using a broad array of different checks and balances to assure that they are scientifically accurate.
IVR and Online surveys are conducted for our media clients because they provide an accurate - but affordable --- method of survey research. IVR polling is being successfully conducted by Rasmussen and Survey USA. It is still more accepted by the media than Online polling because it is considered to be a more scientific methodology. In order to assure that Mitchell Interactive polling is used by as many media as possible, we use IVR for our media polling.
However, we believe that our Online polling is just as accurate. Currently, the two largest online polling companies are Harris Interactive and Zogby International. We are proud to follow the lead of these four companies in using Online and IVR technology.
There are still a number of people who are skeptical of Online polling for political purposes. They just don't believe Online polling can be accurate. We do not share that skepticism. For years, we have been encouraged to use Online polling in political campaigns. For years, we said "No!" The primary reason we did not want to do it is because we felt Online polling simply could not be as accurate as telephone polling.
As we move forward with our Online polling, we feel confident that in a few years it will be a more acceptable method to obtain accurate political polling data. Until it is, we will use it in combination with IVR (Interactive Voice Response) polling to make the media more comfortable in using the results. Actual results of the IVR polling will be used in determining the survey results that are released.
Working for the news media for more than 20 years, we are extremely aware of the importance of accuracy. Our company was the most accurate media pollster in every state in which we polled between 1986 and 2006. Our national poll in 2004 was more accurate than those conducted by other known polling companies. Our telephone survey conducted the Saturday and Sunday before the 2004 November election was more accurate in projecting the final results in the Bush v. Kerry race nationally than Harris, Reuters/Zogby, ABC/Washington Post, Newsweek Magazine, American Research Group, Marist College, and Fox News.
With the type of accuracy we have always shown, you may ask why risk a new methodology? Hundreds of millions of Americans have Online access and we are able to reach the same percentage of Americans Online as we do over the phone. We have worked very hard to assure that the Mitchell Interactive polling will be every bit as accurate as the phone polling we have always conducted. The methodology for IVR/Online polling is different, but you still end up with a very accurate sample and therefore an accurate survey. We constantly compare demographics to make certain the Online respondents mirror the respondents of telephone surveys. In order to track the demographics, we compare up to 20 different demographic characteristics such as age, race, political party affiliation, etc.
There are a number of important assets to Online/IRV
polling:
- You will generally have a larger sample, providing greater accuracy and larger sub-samples (women, men, etc.)
- With the larger number of respondents from IVR and Online polling you lower the Margin of Error.
- The challenge of dealing with voters who only have a cell phone is no longer an issue, because they are more likely to be on-line.
- As the Internet matures, the tendency for people to rapidly change email addresses has diminished. Today, we are seeing much greater stability in email addresses.
- Mitchell Interactive does not pay political respondents. Although we have a rewards system for our consumer panelists, we do not provide any financial compensation for Online political respondents.
- We do not send out mass emails to the same people over and over. Instead, we pick samples from our huge database and send them the survey.
- All of our respondents have opted in to do the survey. We have all the demographics on them. We keep track of how often they have participated and do not allow them to participate too often.
There are a number of proprietary steps we take to ensure our IVR/Online polling is as accurate as our telephone survey.
Most importantly, when my name is on a poll, I will take
every possible step to assure its accuracy. Therefore,
when we do an Online poll, we will make sure it is every
bit as accurate as our telephone polling. |
