A Tool to Explain Affirmative Action

The lessons of the grocery shelf also have something to say about affirmative action.

The New York Times, "Economic Scene" , January 30, 2003

Deciding which shampoo or toothpaste to buy seems a long way from the emotionally charged debate over affirmative action. But an analytical tool developed by marketing scholars to analyze how consumers make brand choices can in fact illuminate that debate.

People have limited time, memory and attention. So when they make buying decisions, they simplify their choices.

"On the shelf you may have 30, 40 brands of shampoo, or 20, 30 brands of toothpaste," explained Jagdish N. Sheth, a marketing professor at the Goizueta Business School of Emory University. But consumers don't take the time to examine every possible choice.

Rather, they reduce their selection to a smaller set of options, based on experience and exposure. "Through learning over time, consumers are really efficient in terms of reducing their transaction costs," Professor Sheth said.

In the 1960's, he and John A. Howard, the Columbia University marketing scholar who died in 1999, developed the idea of the "evoked set" to describe this process of selection.

Shoppers start not with every single brand they are dimly aware of but with a group of options -- the evoked set -- uppermost in their minds.

"An evoked set consists of the brands in a product category that the consumer remembers at the time of decision making," according to "Marketing: Best Practices," a textbook edited by K. Douglas Hoffman.

(An alternative term, "consideration set," is sometimes used for the same concept and sometimes for the smaller set of choices that remain after consumers eliminate unacceptable options from the evoked set.)

Ask a grocery shopper to name toothpaste brands, for instance, and you'll probably hear "Crest and Colgate." Only when pressed to name others will the shopper come up with, say, Rembrandt and Mentadent. Crest and Colgate are the evoked set, the one from which most shoppers will choose to buy -- especially if they aren't looking at snappy product displays for other brands.

The downside of this process is that the results depend on exactly how we sort the possibilities into categories.

"The way this information is recorded in memory can influence consumers' preference for brands, and whether the brand will be considered for purchase," Barbara E. Kahn and Leigh McAlister, two marketing professors, wrote in "Grocery Revolution" (Addison Wesley, 1997).

If, for instance, a store arranges yogurt first by brand (like Dannon and Yoplait) and then by flavor within each brand, consumers will tend to select their flavors from the same brand.

On the other hand, the authors write, "If the products had been displayed with all the strawberry yogurts together, then all the lemon-lime yogurts, and so forth, consumers would most likely choose which flavors they wanted first, and then choose which brand name they would most like for that particular flavor."

Similarly, American supermarkets display meats by animal type -- beef, chicken, pork, etc. -- and then by cut. In Australia, by contrast, grocers arrange meats by the way they might be cooked, and stores use more descriptive labels, like "a 10-minute herbed beef roast." The result is that Australians buy a greater variety of meats.

How we classify goods changes how we make consumer choices. "The composition of the set of final possibilities can have subtle effects on choice," write Professor Kahn of the Wharton School at the University of Pennsylvania, and Professor McAlister of the University of Texas at Austin. As a result, "brand choices can be influenced without changing the actual preference for a brand per se, but merely by changing the content of the consideration set."

What is true for yogurt and meat is true for Supreme Court appointments, award nominees, TV talking heads, corporate board members, conference speaker selections and many mundane hiring decisions.

Decision makers start with an evoked set of possibilities -- the people who immediately spring to mind. Who makes it into that evoked set depends in part on how people are categorized on the mind's "grocery shelf."

Last summer, for instance, The New York Times ran an article on Hollywood's search for young action heroes. Old standbys like Arnold Schwarzenegger and Harrison Ford were getting a bit long in the tooth, leading studios to turn to newcomers like Matt Damon and Vin Diesel. The piece left the impression of a vast generation gap, with no heroes from the latter half of the baby boom.

But one huge action star was inconspicuously absent: Wesley Snipes, born in 1962. Another, Will Smith, born in 1968, was mentioned only in passing.

The evoked set of "action stars" didn't overlap with the evoked set of "black movie stars." There was no racial hostility at work, just the limits of human minds and the categories they create.

Overcoming those limits is the argument for a certain type of affirmative action -- not quotas or preferences, but an active effort to select from the full range of possible candidates, not merely the first evoked set. (This analysis does not apply easily to cases like college admissions, where the selection is made from a large pool of people who actively present themselves for consideration.)

If you are looking for the best possible conference lineup, just listing the speakers who immediately come to mind may inadvertently exclude good candidates. You should also search through the other categories your mind uses to classify people.

Some of those categories may be the politically fraught ones of race and sex. But, depending on the context, this "affirmative action" might include others, like geography, political persuasion, age, educational background or professional discipline.

The goal is not to meet numerical targets but to make the final selection from a broad enough sample to ensure not only fairness but quality. What's efficient for picking toothpaste isn't good enough for people.