Textbook Publishing Case
Putting a Lid on the Garbage Can

In an interesting study of decision making by editors in the college textbook publishing industry, Levitt and Nass (1989) concluded that the garbage can model fit well many aspects of the choice behavior.

Interviews with editors of the ten best-selling introductory textbooks in physics and sociology found the editors consistently describing their work in gambling terms ("a lottery with bad odds," "a crapshoot") and that decision making was best described as "guesswork, intuition, and opinion."

A sociology editor expresses this feeling of ambiguity and confusion:

"Editors can become schizophrenic. You think a manuscript is good and it doesn't make money. Then you get a manuscript that you think is bad, and it makes money -- but not always"
(quoted in Levitt and Nass 1989, p. 192)

In textbook publishing, decision situations are characterized by ill-defined preferences, unclear technology, and fluid participation. Preferences are ill-defined because interpreting and differentiating between success and failure is highly equivocal and malleable.

For example, one editor stated that a physics book which sold poorly was still considered a success because "it was important for the company to have an entry, any entry, in the physics market." (p. 193) The organizational technology is unclear because the connections between means and ends are unclear.

There are no specifiable procedures or formula for producing a successful textbook, and editors often work in disciplines in which they were not trained. Participation is fluid because editors change departments and publishing houses relatively frequently. Levitt and Nass observed that

"it is part of the occupational culture of editors that being fired (even more than once) is no indication of incompetence." (p. 195)

Besides, the gestation period of textbooks is three to five years, with projects often being handed down to new successors. As predicted by the garbage can model, timing is an important element in deciding about projects. For example, introductory texts sell better in their second or later editions, but the decision to do this depends on whether the publisher is planning a new entry into that market segment.

Serendipitous or random events often play a significant part in the acquisition of manuscripts, so that text-book editors recognize the importance of maintaining strong links with academics. Problems, solutions and participants also track each other through time, as for example, when academics in the artificial intelligence field claim that they would teach introductory AI courses if there were a suitable text, while editors counter that such texts would be produced if there were courses (Levitt and Nass 1989).

Despite the disorderly decision process, the textbooks produced as outcomes show significant levels of homogeniety in terms of the ordering of contents and topics. It was as though a lid was being placed on the garbage can. Levitt and Nass (1989) showed that textbook homogenization was the result of the forces of coercive, mimetic, and normative isomorphism (DiMaggio and Powell 1983). Thus, established paradigms define the essential contents and topics that should be included (normative isomorphism); editors imitate successful textbooks produced by others (mimetic isomorphism); and the organizational structure of publishing houses often mirror that of the higher education institutes who are their customers (coercive isomorphism).