The Role of Context Parameters in The Validity of Data

The way to see ecological validity and its potential effects on consumer judgment has direct methodological implications. In the field of sensory and consumer science, studies looking at the validity of contextualized experiments fall into two categories: those that approach the issue of ecological validity as a whole (the experimental context consist of a combination of the environment and the task performed and, attempts to keep most of them as close to natural as possible) and those that focus on specific factors that are found to have an impact on the measures and, try to make these more ecologically valid.

The studies following a global approach compare scores on food liking and choices in different natural environments (restaurants, canteens, prisons) with those obtained on laboratory or central location settings showing differences on hedonic scores. Those differences are usually related to the degree of discrimination among products – consumers being more discriminant in natural settings than in laboratory settings – or to the higher scores on natural settings versus laboratory settings.

On the other hand, the studies focusing on context variables compare how the addition of contextual variables in controlled experiments can affect food liking and choice. We may first notice that several classifications of contextual variables have been proposed: Rozin & Tuorila (1993) divide contextual variables into either product and non-product variables and subdivide them in simultaneous and temporal contextual factors; Meiselman (1996), proposes to distinguish between three categories of variables (the situation, the individual and the product); whereas Stroebele & De Castro (2004), divide the contextual variables into social context variables, physical surroundings, time related characteristics and distraction and/or television viewing. From these studies, it is difficult to fully disentangle the various factors and isolate a specific context effect. The relevance of those contextual variables thus remains unclear. To date, the lack of knowledge of the combined effects of these contextual variables on consumers’ responses compromises the ability to identify causal relationships through experimental approaches.

In practice, a consequence of this is that participants to a test may not perceive the study context the way the researcher assumes they would. This questions the ecological validity as defined by Brofenbrenner.

The issue seen as a whole would naturally lead to global changes in the test design, while dividing context into the separate variables would bring targeted improvements of the experimental setup, keeping the rest of the task and environment potentially non ecological.