"Writers have lamented our reliance on what consumers say rather than what they do.  We collect data through obtrusive surveys rather than unobtrusive observation or secondary data.  Mark Pigott described what's wrong with this practice:  'So often we're obtaining data and information in environments that taint the value because they're being acquired in artificial circumstances.'

What consumer researchers need to do is observe consumers in their native environments -- making choices, purchasing and consuming.  It's time we started observing actual consumer behavior instead of playing obtrusive word games with respondents about what they say they perceive, believe, think or intend through unnatural laboratory experiments and surveys.

I am skeptical of research that asks people to respond by checking off points on scales in non-purchase or non-choice situations.  I also question results garnered in response to linguistic descriptions or to hypothetical products, situations or advertisements.  How can these responses possibly reflect reality?

In consumer research, we too often short-circuit these necessary conditions for external validity.  We have an all-too-easy escape hatch to the tedious business of observation:  It's called language.  We simply ask subjects about their beliefs, opinions, perceptions, attitudes and intentions and then make gargantuan assumptions that these obtrusive measures of intentions will predict behavior."