Hum, most consumers are smarter than I give them credit for.
Update (8/27/2014): The article does a marginal job describing how the survey was conducted (it was on-line), noting that it was a weighted survey, and describing the error factors (+/- or plus/minus factor). But it says nothing about what questions were on the survey or how they were phrased. It is easy to influence how respondents answer a survey by how the questions are worded. Real scientific surveys (and I suspect this wasn’t – because a “voluntary on-line survey” isn’t random) are vetted for non-discriminatory, non-influencing, neutral language.