The Edge: Competitive Edge Blog

The Rising Cost of Public Opinion Polling

Print Friendly, PDF & Email

The cost of conducting quality public opinion research has risen over the past decade. Note that lousy polling has remained fairly cheap. Two primary trends account for it: the American public’s growing tendency to no longer answer unsolicited phone calls, and the switch from landlines to cell phones. In the face of these trends, Competitive Edge is working to control costs while continuing to innovate for quality.

Response Rates

Americans don’t like to answer calls from an unknown number. It wasn’t always this way.  Many of us are old enough to remember the – dare I call it – excitement of getting a phone call, not to mention a long-distance call, from an unrecognized number.  Regardless of the amount of time on the phone we’re talking about, people are now less willing to give their time to a stranger. The ubiquity of call-blocking and caller-ID systems has made avoiding calls a lot easier.

Phone polling response rates – which just ten years ago were at 30-40% – have dropped to roughly 5%. This translates into dramatically higher polling costs, because it takes so many attempts to get someone on the line. Somewhat countering this phenomenon is that when a potential respondent is reached, they are just as willing to share their opinions as they were a decade ago. Paradoxically, because it’s more difficult to get someone on the line, people who are contacted now participate at a reasonably high rate.  But that does not mitigate the fact that it now takes 30 or 40 phone numbers to conduct an interview with just one individual, up from roughly from a ratio of 15:1 in the 2000’s and 5:1 when CERC was founded in 1987.  A pollster now needs 10,000 numbers or more to complete a 300-person survey. Even with auto-dialers, that’s an awful lot of work hours to wade through those phone numbers.

Cell Phones

And auto-dialers come in less handy these days.  Just a short ten years ago, 10% to 12% American households relied only on a cell phone. Now, more than one-half of American households have shifted entirely to cell phones. This would be OK for the public opinion polling industry, were it not for the Telephone Consumer Protection Act of 1991’s prohibition on using autodial equipment to dial cell phone numbers. When calling cell phone numbers, call centers cannot rely on a piece of equipment that autodials randomly-selected telephone numbers before immediately routing answered calls to live interviewers. Instead, interviewers must dial cell phone numbers manually. When those calls go unanswered, as they more frequently do, multiple hand-dial callbacks to those numbers are required.

Here’s the thing though:  good polling requires a dogged pursuit of that randomly-selected respondent. The quick and sloppy way to do polling – whether it’s online or over the phone – is to contact as many people as possible. With that approach, only the “cream” gets skimmed off the top; that is, the people who are easily reachable and eager to do a survey get into the sample, while those who are more difficult to reach are left out.  Good data collection avoids the easy way out.

How Polling Firms Have Responded

Some polling firms cap the total number of cell phone calls in a survey (say to 30%) and then make statistical weighting adjustments to meet the overall distribution. Others have shifted to online polling methods, with the corresponding loss of true scientific probability sampling that this entails. Still others have jumped on Interactive Voice Response (IVR) polling, which is just a fancy way to say automated polling where participants interact with a recorded voice, not a live human being. Problems inherent with IVR polling include the fact that it can’t call cell phones and encourages false responses from participants. Not surprisingly, IVR suffers from significantly lower than average response rates. If you think the response rates for live calling are bad, just think about the sad response rates that come from IVR polling.

How Competitive Edge Helps You Win

Competitive Edge is committed to helping clients win. We refuse to compromise research quality by adopting methodologies that simply don’t yield accurate or valuable data. The cost of bad research is incalculably higher than the cost of good research.  Why? Because bad research will lead clients to make bad decisions – and make those bad decisions confidently. Survey research is like a map. Maps give you confidence.  A good map will get you where you want to go, but with a bad map, like bad research, you’ll confidently march off in the wrong direction and end up with dire consequences. CERC will suggest measures to help clients keep research costs within budget, but not if the research approach compromises the integrity of the data and results.

We maintain our own call center with a staff of 100+ fully-trained interviewers. This affords us complete oversight and control of the public opinion research process, without the need for outside vendors. With more than three decades of experience, we’ve developed operational efficiencies that empower us to perform without compromise, and we’re quick to field a project because we can devote more resources directly to the project itself, not its coordination. Whether a polling project calls for landlines, cell phones, on-line or mixed-mode, we deliver the method and format that yields sound data married to reasonable cost.

Take online polling, for instance. We do employ online research – and it does lower the cost of research – but we do it in ways that maintain the ultra-critical randomness of probability samples. Beware of firms who offer sweetheart deals with purely web-based survey techniques. They deal in non-probability samples that carry a bevy of assumptions, and make calculating the margin of sampling error for such polls impossible. That’s right, you can’t have a margin of sampling error without a probability sample and probability samples must be methodically random. CERC instead has tightly integrated our computer-assisted telephone interviewing (CATI) and computer-assisted web interviewing (CAWI) systems so that cell phone, landline and online data is seamless woven together in random samples.

Another area in which we offer definite competitive advantage is bilingual surveys. More than one-quarter of our interviewers are fluently English-Spanish bilingual. With locations in San Diego, CA and El Paso, TX we are plugged into the bilingual community like few other research firms. This greatly helps us to keep bilingual survey costs low, because we do not contract out bilingual interviewers, questionnaire designers or translators.

In short, Competitive Edge invests in great people, great processes and great equipment. This may cost a little more than the other firm that is rounding off his edges to stay in the game, but the result is research you can trust. Great decisions come from good data, and good data is going to cost more.


This entry was posted in Campaigns, Polls. Bookmark the permalink.

Related Articles

One Response to The Rising Cost of Public Opinion Polling

  1. Michael Lee says:

    Before I read this article, I did not know that it was illegal to use an autodialer machine to call cell phones. I agree that it would be much more difficult to use opinion polls. I have problems getting people to answer calls at all, let alone to strangers.

Leave a Reply

Your email address will not be published. Required fields are marked *


Featured Research Download

Are you targeting the "persuadables" of your campaign? Download our seven step process to winning an election.