8.3.5 Model Constituency



Because of all the diverse interests that can surface in an intervention program, it is necessary that any such program have a broad support base or constituency. In a program that is strongly model-oriented, one must build a constituency for the model itself. Such a constituency comprises the model users and other individuals and groups that support the model implementation effort.

Building a constituency for a model is often difficult. The urban analyst is likely to feel most comfortable interacting with a "bilingual person" in the agency, one who speaks "model language" as well as "agency language." Such people are rare and highly valued. They are professionally mobile-apt to be promoted or to relocate voluntarily to another agency. An urban analyst who builds his or her implementation process solely through this one person surely has manufactured a weak link in the implementation process. Any program that requires more than, say, one year to complete, runs a high risk of losing such a person, resulting in a "vanishing advocate" [CHAI 75]. (Recall the Boston police simulation mini case study.)

Developing a broader constituency by holding workshops and training sessions and soliciting feedback has other important benefits. Agency personnel, through feedback, often provide inputs that can turn an otherwise naive modeling study into a realistic and sound analysis. Such feedback can relate to safety factors (e.g., permissible U-turns for school buses8), concerns over supervision (e.g., as in the swapping of police patrol officers in southern Manhattan), legal issues (e.g., in the selection of a jury pool), correct interpretation of recorded data (e.g., as in the accuracy of recorded travel times), and probable response of rank-and-file workers to a proposed innovation (e.g., to a redeployment of personnel). The process of eliciting feedback can transform otherwise skeptical, perhaps even hostile, personnel into fully participating supportive allies of the study. Supervisory personnel, as well as in-the-field workers, can become advocates. A broad mutual investment in a program is not likely to lead to subversion during implementation. And a model either created or at least modified in the presence of feedback from personnel is more likely to deserve broad-based support.

The needs and priorities of the user and other managers should also be considered when building a constituency. Key decision makers, including the user, have short- and long-range goals that can have a beneficial or detrimental effect on an implementation effort. Remember the commissioner of education whose short-range goal was demonstrating frugality, leading to a long-range goal (passing a multi-million-dollar bond issue)? Or a long-range goal of an urban official may be reappointment or reelection, and an urban modeling analysis may be viewed positively or negatively in this light. Some of these goals may be apparent to the urban analyst; many come to light only after the fact. A common mistake here is for the analyst to focus on long-range projects while ignoring short-term problems. Public decision makers are often under pressure to "put out fires," and analytic assistance is usually very much appreciated. Short-term assistance, even if it appears to delay progress in long-range projects, has two benefits: (1) it helps the decision maker address the current problem, and (2) perhaps more fundamentally, it establishes credibility for the analyst in the eyes of the decision maker, thereby enh ancing his interest in the long-range projects.

A strong model constituency is no guarantee against the vanishing advocate syndrome. Sometimes a modeling effort is associated with a particular administrative regime; viewed as an instrument of reform, it develops political attributes. Should the current regime-the one supporting the modeling effort-be replaced, the effort could be discarded because of its links with that regime. However, any innovation viewed as a reform runs similar risks. The development of a wide base of support minimizes this risk. 8 The students performing the school busing study interviewed bus drivers to acquire this information.