Opinion Surveys

Soliciting customers for opinion surveys can be laborious, particularly if you're conducting research over the phone. That's why many polling agencies farm out their work to independent contractors or freelance agencies. The trick to making sure that your opinion survey data remain pure is to train your call center operators in the vernacular and technicalities of your poll.

This doesn't mean introducing your operators to every nuance of your opinion experiment. Indeed, giving them too much information may actually make them less objective moderators. Simply teach your operators the basics, and train them to record customer feedback accurately. Since opinion surveys can be swayed by operator sentiment, you may wish to conduct separate bias tests to weed out any operator bias effects.

Taking Survey Bias into Account

The phraseology of your opinion questions can also impact your results. Leading questions, for instance, can radically alter your survey ratings. At the same time, even minor biases can yield catastrophic results for opinion surveys. In Malcolm Gladwell's book Blink, the author describes a situation in which a number of African-Americans were asked to identify their race before taking a GRE practice test.

Simply as a result of admitting that they were black before they took these tests, the students scored significantly lower on their GREs. In other words, such an overwhelming societal bias against African-Americans exists, that these students scored lower just by identifying their race. As frustrating as these results may be, there are ways to fight opinion biases. When subsequent surveys showed students positive African-American roll models, like Martin Luther King and Colin Powell, their racial identification bias made them score higher than the control group scored.