The Brands We Know

Why do you drink Coca-Cola? Why do you wash your clothes in Tide? Why do you drink an expensive imported beer instead of a no-name brand?

These questions have interested me for a while, but they have remained dormant. As my academic training has provided more tools to understand emotion and memory, the notion of brands has remained on mind.

Recently -- with the help of a group of talented undergraduate students -- I began revisiting the topic. Eventually, we will do some scientific research on the topic. For now, I am collecting informal data to report here.

I am trying to determine what brands mean to people. How are they semantically represented in the mind? To do that, however, one has to commit to a definition of "meaning," and this borders on philosophy. In my lab, we are looking back toward history. In 1957, Osgood, Suci, and Tannenbaum published The Measurement of Meaning. This seminal work outlines the semantic differential, which is a tool wherein two adjectives are separated by some number of equal intervals, usually 7. So for example, a semantic differential might look like:

good __:__:__:__:__:__:__ bad

For any and all things, I could ask you this question. So, if you thought "X" were particularly good, you could check the space right next to "good" and so on. Osgood et al. used 50 such pairs of adjectives across a variety of stimuli. This produces an enormous data set. They were able to reduce the data to manageable size with a mathematical technique called Factor Analysis, which (to oversimplify) allows us to reduce the number of dimensions driving a data set. Thus, for example, if I ask you 50 semantic differentials (e.g., good-bad, pleasant-unpleasant, exciting-calming), do you respond based upon 50 unique "factors" in your mind? Or are there fewer factors driving your responses? Factor analysis provides one tool to try to reduce these dimensions.

Interestingly, Osgood et al. repeatedly found that most of the pattern of responding (i.e., variance) could be described by three dimensions. They found these same three dimensions with a wide variety of stimuli. After examining the pairs of adjectives most associated with each factor, Osgood et al. labeled the dimensions "evaluation," "activity," and "potencty." This 1957 book is one of the most cited works by emotion theorists who subscribe to a dimensional theory of emotion. Thus, we tend to call these dimensions, "hedonic valence," "arousal," and "dominance."

My simple question for now is whether we can describe mental representations of brands in the same three dimensions. As I said, in the coming year we will conduct systematic research in this area. For now, I am collecting some data to write about here, a journalistic exercise. If you have any interest in how we come to know brands, stay tuned.