In graduate school I was taught to distrust surveys of consumer behavior. “Look at what they do, not what they say,” was the lesson. The Wall Street Journal reinforces that training with an eye-opening article about people lying to poll-takers. One striking example is the discrepancy between answers to a live person on the telephone compared to entering answers on a computer form:
Went to church last week: 56% on the phone, 25% online
Gambled last month: 21% on the phone, 28% online
Donate to charity regularly: 67% on the phone, 57% online
Brush teeth twice daily: 78% on the phone, 64% online
Exercise regularly: 58% on the phone, 35% online
Drank alcohol last week: 39% on the phone, 53% online
Keep this in mind whenever you see a consumer survey. Even more importantly, when a newspaper reporter draws a conclusion from interviewing a handful of people, be very, very skeptical.