those w h o study the replication process—explaining whether and h o w the replication of some exemplary program "model" occurs in other sites. Craig H. Blakely and his col- leagues (Blakely and others, 1987) examined seven innovative programs in education and criminal justice adopted in at least twenty sites nationally. They followed a procedure developed at the University of Texas Research and Development Center for Teacher Education in 1981. This procedure "conceptualized social programs as consisting of a finite number of components or parts. Program fidelity could then be defined as the num- ber or proportion of finite program components that were implemented." Components were observable activities, materials, or facilities and had to be verifiable through inter- views with staff members and clients of the implementing organization. Each component was also to be "logically discrete from other components" and, if not discrete, then it "should not depend on the implementation of another component" (p. 260).
The degree of presence of each component w a s rated as ideal, acceptable, or unaccept- able (p. 258). The authors found a .38 correlation between fidelity of replication and effec- tiveness. This finding does not necessarily show that fidelity increases effectiveness, how- ever. It is possible that components were sacrificed not because implementers were in any sense unfaithful to the original model but simply because they were constrained by cir- cumstances to implement what they realized w a s a regrettably imperfect, but it w a s hoped acceptable, version of the model. The .38 correlation shows that more of some- thing increases effectiveness, but it is not clear that the something is the model. The essence of the model is the clever trick for taking advantage of some latent potential, or opportunity, for extracting value on the cheap. Counting the number of implemented components is no way to characterize how well the model was implemented.
Skills and Abilities
My focus is less on the personal skills of craftsmen than it is on the nature of craft materials and on the nature of craft practice: I focus on lumber and the ways to choose it rather than on architects and builders, to return to the home-building analogy. But I need to discuss craft skills and abilities to some extent, because they are obviously integral to the whole craft process.
G R O P I N G A L O N G . The best craftsmen are skilled at perceiving
opportunities in the materials at hand and surveying the array of pos- sible smart practices that might be used to exploit them. Some of this is done through pure rational and self-conscious planning, but more is done experimentally, through trial and error, in a process Robert D. Behn has called "groping along."37
Behn's principal specimen of groping along is Ira Jackson, a young and inexperienced appointee as Commissioner of Revenue in Massa- chusetts in 1983. Behn accepted Jackson's self-deprecating description of himself as a person without ideas, without a strategy—just someone who experimented and retained what worked as he transformed the Department of Revenue from a laughingstock into an apparently suc- cessful tax collection agency. In the end Jackson was able to fashion a crisp slogan and a smart strategic idea for the agency: "honest, firm, and fair revenue collection."38 But this was not an idea that Jackson grasped at the beginning of his tenure. He groped his way into it.
Jackson was groping not just at the level of ideas but at the level of action as well, a point Behn was less definite about, preferring to call it "trying . . . numerous things. Some work. Some do not. Some are par- tially productive and are modified to see if they can be improved. Finally, what works best takes hold."39 It was at this level that Jackson was most successful: Once he had accumulated enough successes it was not too difficult to examine them with a view toward conceptualizing the common thread.
Valuable as the groping along metaphor is, its use by Behn and oth- ers in the subsequent literature tends to underestimate the role of cer-
37. Behn (1988).
38. Behn (1988, p. 644). The full, and very interesting, details of Jackson's successful turnaround efforts are described in Behn (1987).
tain kinds of craft materials in affording a kind of tactile guidance to the actors doing the groping—the latent potential, one might say, of certain materials to signal "hotter" and "colder." In the revenue collection case Jackson was groping within a well-structured domain of potential strategies and tactics. He instituted a series of tough, well-publicized enforcement actions against tax evaders. This is a strategy that must have occurred to every tax collector since the pharaohs. At the same time he added a twist that m a d e this strategy suitable for a liberal democracy: He represented his increased pressure on tax evaders as a form of service to "the hundreds of thousands of honest taxpayers in Massachusetts."40 At relatively low levels of tax compliance, people who pay through continual collections at the source are at risk of feeling like suckers relative to those who can avoid the tax collector's grasp by means of once-a-year self-assessment. It was part of Jackson's shrewd- ness to appreciate that the majority could be induced to support tougher action against a seemingly unworthy minority even if this tougher action might spill over to eveiyone else as well. As it happened,
there were m a n y d o m a i n s a n d m a n y classes of tax evaders—from
restaurants that pocketed customers' meal taxes to yacht owners who dishonestly declared out-of-state addresses for their crafts—that could be targeted in this way.
In the ICC case, groping along occurs first in a typically lengthy, com- plex, multiparty discussion process in which participants signal one another about their degree of interest in collaborating on some range of projects whose blurry definition is clarified as the discussion proceeds.41 The groping is facilitated to the extent that the participants are mutually intelligible to one another and willing to be reasonably honest. These two conditions are often hard to meet simultaneously. At the individual level much of the groping is done by actors trying to conceal their true interests or to hang on to their professional and bureaucratic parapher- nalia even while testing to see how far they might go in disclosure with- out incurring too much risk and cost for their agencies and themselves. This is true even of actors motivated by value-creating purposes. A
40. Quoted in Behn (1987, p. 17). Jackson said he was trying to prevent "a new social disease. It has become the 'in' thing to cheat on taxes rather than the w o n g thing." One of his favorite slogans w a s that tax cheating w a s not a victimless crime (p. 17).
41. One might say that in the ICC arena planning occurs through groping along,