Political season drives home the point of how many decisions are made based on polling. It's a great time to reflect on what separates good polling from bad information.
The keys to getting accurate and usable research data are:
• Using the right survey or combination of survey techniques;
• Interviewing a representative sample of your target audience; and
• Asking direct, fair and single-topic questions.
Research breaks down into quantitative and qualitative techniques. Quantitative research involves enough interviews to produce reliable statistical results. Qualitative research engages smaller groups of people who provide insight and context on advertising, new product features and pubic issues.
A lot goes into selecting the right technique or combination of techniques for a particular need. One of the most elemental aspects of the choice is getting a representative sample to interview.
Several years ago, Portland homebuilders wanted to embark on a marketing PR campaign to reach young families about buying their first home. The poll they commissioned wound up interviewing a large number of older respondents and relatively few young adults with new families. The "findings" were worthless to the task at hand.
Selecting the right kind of research technique also can be undermined by asking loaded questions. The whole idea of research is to find out what people think, not trying to conspire to get the "right answer." Straightforward, unbiased questions can elicit accurate responses that give you what you need to know, not necessarily what you want to hear.
Questions also can be overloaded with too many variables, so you really aren't sure which variable tipped the respondent's answer in one direction or the other. Evaluating single variables in a single question offers a higher level of confidence you will understand what makes your target audience's mind click.
The choice of research techniques is influenced by, among other factors, technological shifts. Fewer people have landline telephones in their homes. Those that do are shielded from unwanted calls by features such as Caller ID and voice mail.
The pendulum has clearly swung to use of smartphones, which, in addition to all their other sweet features, cannot be dialed randomly, erasing those annoying dinner-time interruptions to answer a poll. So research professionals are switching to Web-based and other techniques, each with their own merits and demerits.
To get effective and actionable results from research, you need to be a smart consumer. The pollster you need is the pollster who can explain clearly why he or she has selected a particular research type, how a representative sample will be achieved and the kinds of questions that will be asked.