Political polling and targeting of voters is no longer a one-size-fits-all business. There are a variety of options available based on your goals and your budget. I have explored all of these options, and our company, IBoPE Zogby International, finds that the greatest demand is for interactive polling.
Yes, operator-generated phone calls to voters are still the gold standard, but like gold they are costly. (They also fail to reach many typically young cell phone–only users.) Automated calling is cheaper, but turns off many respondents, and the medium tends to limit the number of questions and demographic categories you can ask about.
Interactive polling is much less costly, produces results more quickly, and is capable of generating much larger samples and yielding more demographic categories than the other available methods. One downside is that it can be challenging to create a sufficient panel size in some states and in races below the state level. Also, interactive poll samples do not have universal coverage because not everyone has Internet access.
At the same time, calling landlines alone offers far from universal coverage. Today, landline penetration is approximately 63 percent, which is close to where it was in the early 1960s, before serious phone polling began. Even if you merge in cell phone users as 25 percent of the sample, you still fall short of universal coverage. The fact that both landline and cell phone use involve self-selection makes pure probability samples out of reach. In addition, you have to take into account the legal and practical challenges to reaching cell phone users. Online access penetration is much greater, with 91 percent of likely voters and 82 percent of all adults currently on the Internet.
As noted above, our increased use of interactive polling is driven by the preferences of our clients, and has expanded ahead of schedule from what we envisioned when we first went to online in the 1990s. But the fact is that most companies and NGOs don’t have six-figure budgets for polling and find less expensive interactive polls reliable. The need for rapid turnaround is also a factor; our clients often want results within days, and an interactive poll that goes in the field on Tuesday can easily bring several thousand responses by Thursday. Readers of this publication care about measuring voter opinion reliably. We have stuck our necks out by using interactive polls on elections, and have had success with U.S. Senate races and national results. To show how far interactive polling has come, consider the fact that in the 2010 midterm elections, two interactive pollsters, IBoPE Zogby and YouGov/Polimetrix, came closest to predicting the actual generic ballot vote total for congressional races.
One explanation for this accuracy is that interactive polling also produces very reliable likely voter models. People who make the effort to give their opinions online also tend to be more motivated to vote. Again, the caveat is that not everyone has online access, but this need not be an insurmountable impediment to accuracy.
An overlooked value of interactive polling is targeting and especially microtargeting of voters. Over three nights, an interactive poll can garner responses from 6,000 likely voters to as many as forty demographic and twenty behavioral questions. This offers vast possibilities for targeting. Such a poll can yield samples in the hundreds of groups such as women with young children, small business owners, investors, gun owners, vegetarians, college students, and on and on. It can also produce smaller, but still meaningful, samples for subsets such as physicians, gin drinkers, skiers, etc.
We can all say with certainty that the days of using landlines to conduct polls are numbered. Until some other technology proves itself, interactive polling is the only viable successor. It may not be perfect, but interactive polling has already demonstrated that it is reliable and it will only become more so over time.
John Zogby is chairman of the board and chief insights officer of IBOPE Zogby International.