Innovations as Catalysts for Organizational Change: Shifts in Organizational Cognition and Search

Innovations as Catalysts for Organizational Change: Shifts in Organizational Cognition and Search – Statistical Data Included

Henrich R. Greve

This paper uses data on radio format changes to test hypotheses on innovations as catalysts for nonmimetic change in organizations. Innovations are difficult to interpret using existing schemata, causing organizations to search for information on the opportunities and threats implicit in observed innovations. Such search may lead to mimetic adoption of the innovation or, more likely, to more varied nonmimetic change. Results show an effect of innovations on the rate of nonmimetic change in radio markets, with innovations in large or nearby markets having greater effect and innovations by large organizations having less effect. The social and competitive relations of the innovator to a given organization are thus modifiers of the catalytic effect. These findings have implications for theories of innovation, competition, and organizational isomorphism. [*]

A central theme in the study of social systems is the perturbations caused by “newness.” In organizational inquiry, the newness is often product and process innovations (e.g., Quinn, 1986; Van de Ven, 1986), and innovative behavior is a strategic activity by which organizations gain and lose competitive advantage (von Hippel, 1988; Jelinek and Schoonhoven, 1990). This perspective has resulted in a stream of research that describes the benefits, obstacles, and patterns of implementation of the new activities. What has garnered less attention is the more general role of innovations as catalysts for shifts in interpretation, decision making, and action in the organizational fields in which the innovation occurs. Managers observe the behavior of other organizations in their organizational field, and their observation of innovative competitive activity can have two effects on their organization. The first, well-studied effect is imitation of the new behavior with the intent to gain its benefits. The second, less-s tudied effect is non-imitative actions as the organization embarks on a wider search because the innovation changes cognitions about the appropriateness or exigency of taking action. In this paper we explore the extent to which innovations generate nonmimetic responses, an effect that is much more important than the scant attention given to it suggests. Research on such catalytic change provides a path toward a theory of organizations interacting with their environment, not just reflecting it through mimetic behavior, and is an effective approach to studying managerial cognition and organizational search.

Organizations face information environments with accuracy ranging from precise information to speculation (Barnard, 1938). [1] Precise information allows organizations to act according to the standard norms of rational decision making (March, 1994). Their managers investigate, optimize, and plan their actions to maximize the expected rewards. As the information decays from precise to uncertain, managerial attention, interpretation, and judgment become increasingly important in the decision process. Although managers may wish to avoid uncertainty (Cyert and March, 1963), they frequently face new stimuli that inject uncertainty into the future consequences of alternative actions, including the continuation of their current actions.

The process of decision making in uncertain environments revolves around a cycle of environmental scanning, interpretation, and learning (Daft and Weick, 1984). Managers scan the environment by receiving information about the actions of relevant others, such as customers, suppliers, competitors, and regulators. They interpret it using their cognitive schemata, structures that encode past experience and guide future actions (Axelrod, 1972; Fiske and Taylor, 1991). They learn by determining their future actions, either by continued exploitation of their existing activities or by changing their activities to search for better rewards (March, 1991). Managers in this cycle react to the innovations of competitors because the meaning of such innovations is uncertain, casting doubt on traditional interpretations of the environment and suggesting opportunities for organizational search.

Search as a tool of organizational learning takes different forms depending on the information environment. When managers are confident of their interpretive schemata, search consists of passive scanning, often with a bias toward confirming existing schema (Higgins and Bargh, 1987; Fiske and Taylor, 1991:149-152). When schemata are held with less confidence, however, the behavior changes from passive scanning to active probing of the environment, as individuals face nonroutine problem solving (Larkin, 1983; VanLehn, 1993). This process of active search (Argyris, 1996) can generate new strategies and discredit old ones. It is a precursor to organizational change that involves organizational resources and enacts new organizational roles in the environment, so the heuristic rules for initiating search are a very important part of organizational learning. As March (1988: 3) stated, “Since only a few alternatives, consequences, and goals can be considered simultaneously, actions are determined less by choices amon g alternatives than by decisions with respect to search.” Search is triggered by increased perceived environmental uncertainty (Stinchcombe, 1990: 3-5; Weick, 1995: 95-96), leading to a simple heuristic of increasing search after observing other organizations making innovations.

EFFECTS OF INNOVATION ON SEARCH AND CHANGE

Activities are deemed innovative if they differ significantly from current or recent activities. In organizations, innovations may change the incumbent skills, standard practices, technology, services, and products of the firm. Innovations are likely to present significant cognitive problems when they involve new core concepts or new relations among core concepts (Henderson and Clark, 1990), suggesting that newness relative to the knowledge of the focal organization may determine the difficulty of adoption. Organizations with established routines and practices face a difficult task of incorporating change amidst continuity (Child and Smith, 1987; Walsh and Ungson, 1991; Bartunek, 1993). An established practice has little ambiguity in execution and a known history of returns, while an innovation has highly uncertain future returns. Organizations also tend to commit to activities that they have done well in the past and are structured to exploit the practices they consider to be competencies (March, 1991). Henc e, the bias of most organizations is to continue their current activities.

Innovation as Weapon

Innovations are rare, but when they occur, the consequences for the innovator and the competitors are often critical. Organizations make product innovations to move into new industries, try out new technologies, or market entirely new products and make process innovations to extract greater rents or gain advantages over competitors (Burgelman and Sayles, 1986). Organizations use innovations as weapons to gain resources, as seen most clearly in the introduction of radical new technologies that undermine the current technological regime. These technologies are often not introduced by the dominant firms in an industry but by new entrants (Tushman and Anderson, 1986). Such innovations increase the search behavior of existing firms and new entrants, as they seek to understand and master the technology and its market implications. Only when firms have settled on one way of using the technology, a dominant design, does the intensive search end (Utterback and Abernathy, 1975). This argument of punctuated equilibrium in technological evolution has been generalized to include other kinds of changes in competencies, such as changes in market strategy and organizational structure (Tushman and Romanelli, 1985; Romanelli and Tushman, 1994).

The ability to use innovations as competitive weapons gives organizations incentives to develop competencies to make their own innovations or understand those of their competitors (Cohen and Levinthal, 1990). Decisions to obtain the competencies to develop or absorb innovations are especially important to organizations, as competencies are deeply ingrained, difficult to replicate, and can be exploited over long periods (Prahalad and Hamel, 1990). Since innovative behavors have both short-term and long-term consequences, it is difficult to evaluate their consequences so as to identify the optimal solution. Instead, decision makers may rely on a simplified rule of selecting the first alternative that satisfies their criteria (March and Simon, 1958: 140-141). Such a satisficing rule is sensitive to the timing of decisions and the order of considering alternatives, making the heuristic rules for initiating search for alternatives consequential for the decision (Cyert and March, 1963: 120-121). Innovations made by others draw decision makers’ attention, increasing the chances that they will choose an alternative in the neighborhood of recent innovations by others.

Innovation as a Target of Imitation

The processes of mimetic adoption that frequently follow innovations show very clearly that innovations are noticed and acted on (Rogers, 1995; Strang and Soule, 1998). An innovative strategy is mimetically adopted because the fact that others have adopted it suggests that they, based on the information available to them, believe it will be profitable. This increases its estimated value for managers who have little information available on its consequences (Bikchandani, Hirshleifer, and Welch, 1992). Seeing others adopt an innovation also focuses decision makers’ attention on the strategy, so the managers are likely to enter discussions on alternatives to the current strategy (Cyert and March, 1963; Cohen, March, and Olsen, 1972; March, 1981). By setting the agenda and influencing judgments, innovations become targets of imitation. Thus, innovations are spread by mimetic adoption, and the likelihood of each actor adopting is determined by an organization’s social proximity to prior adopters (Rogers, 1995; Gre ve, 1996; Strang and Soule, 1998).

Mimetic adoption is a rather limited form of organizational learning from the experience of others. Diffusion studies assume that organizations learn by adopting the same behaviors that other organizations exhibit, overiooking the role of experimentation and search, but three mechanisms can cause search to go beyond the simple imitation envisioned in diffusion studies. First, even intended imitation may result in innovative behaviors, as organizations may inaccurately imitate the innovation they observed (March, 1981). Second, managers avoid competing with other organizations using the same strategy, so market strategies are usually not adopted when they would be in direct competition with those of other organizations in the market (White, 1992; Greve, 1996; Miller and Chen, 1996; Baum and Haveman, 1997). Third, managers can draw inferences about the structure of the market from the innovations of others and may adopt alternative strategies to exploit the same structure.

Innovation as Springboard

The key to understanding nonmimetic learning from others is to view managers as boundedly rational actors (March and Simon, 1958; Cyert and March, 1963) who will make deliberate strategies but whose analytical and knowledge boundaries lead to the use of decision-making heuristics. An oftenoverlooked aspect of this core idea is the connection between cognition and action. The actions taken as a result of decisions at one point in time affect the future decisions of the focal organization, as well as organizations with which they are independent, in ways that are difficult to predict. Interpretations, meanings, valuations, and the relevance of future options shift when organizations act. Thus, organizations are boundedly rational not only because of their managers’ cognitive limits but also because their actions alter the feasibility of future options.

Decision-making heuristics are a form of organizational cognition that often reflects the shared perspectives of the organizational members (Lyles and Schwenk, 1992). They include rule-like responses to stimulus but also high-level schemas for interpreting information, evaluating outcomes, and judging the likelihood of outcomes (Daft and Weick, 1984). While such interpretation is predicated on actions, as Weick (1990:293) argued, “The emphasis is on actions that provide a pretext for thinking, and not the reverse.” Cognitions do not change spontaneously. They change because of information received through the actions of others and through actions taken by the focal organization, but they rarely undergo complete transformations (Gersick and Hackman, 1990; Starbuck, 1993). What is changed most readily is the evaluation of information (Weick, 1995), the appropriateness of actions (March, 1994), and the foci of attention (Ocasio, 1997).

Innovations by others are learning opportunities because they are imperfectly understood stimuli that can change the appropriateness of the focal organization’s current activities. Innovations are novel, vivid, and unexpected events. Organizations have well-established routines to handle typical environmental stimuli (Nelson and Winter, 1982) but lack patterned responses to unexpected events (Barr, 1998). Innovations create an opportunity for choice and widen the normal sets of choices. As March and Simon (1958: 35) stated, “A stimulus may have unanticipated consequences because it evokes a larger choice set than was expected, or because the evoked set is different from that expected.” Because of this widening of choice sets, “The rate of innovation is likely to increase when changes in the environment make the existing organization procedures unsatisfactory” (March and Simon, 1958: 185).

Search and Change

Innovations suggest to decision makers that existing organizational procedures may be inadequate, which creates a problem for the organization. Routine responses are inadequate because innovations are novel, causing organizations to search for solutions through nonroutine probing for information on the innovation and its implications for the focal organization (Cyert and March, 1963: 120-121). An innovation also creates decision-making opportunities by causing discussions about alternative strategies, increasing the probability that a decision to change the organization will be made (Cohen, March, and Olsen, 1972). Such active search is a temporary organizational change that will have enduring consequences if it leads to the adoption of change as a solution to the problem. Because active search can uncover solutions and because decision-making opportunities can result in decisions, innovations make organizational change more likely.

The relationship between an innovation and organizational change is not deterministic, since organizations may fail to uncover promising solutions and may fail to act on solutions they do uncover. Managerial risk aversion (Cyert and March, 1963) and organizational inertia (Hannan and Freeman, 1977) may prevent change and are especially likely to do so in high-performing, old, or large organizations. High performance causes managers to be risk-averse, making the organization less likely to change (Greve, 1998b). Old or large organizations tend to be inert and unlikely to change (Hannan and Freeman, 1984; Kelly and Amburgey, 1991). Managers’ interpretations may also interfere in the change process, making organizations whose managers interpret the innovation as a threat more likely to reinforce the current strategy (Staw, Sandelands, and Dutton, 1981: Dutton and Jackson, 1987; Ocasio, 1995).

The effect of innovations on organization-level change can be studied by taking into account these organizational modifiers, but an alternative approach is to model the population-level effect of innovations on change (Miner and Haunschild, 1995). A population of organizations will have members with varying performance, inertial forces, and interpretations, so the aggregation to the population level weakens the effect of organization-level moderating variables. A population of organizations may be delineated such that its member organizations observe the same innovations, making the effect of an innovation on the members homogenous. Since the members of a population observe each other and develop shared conceptions of appropriate actions, change itself may even be legitimized at the population level. Thus, the role of innovations in activating the search processes that cause change can be studied by examining how innovations influence the rate of change in an organizational population:

Hypothesis 1 (H1): Innovations will increase the rate of change in an organizational population.

Information Diffusion and the Location of Change

As in ordinary diffusion processes, the social structure of the organizational field determines the availability and salience of information on innovations and thus influences where it is likely to be acted on. Proximity and size are important aspects of the social structure. First, location near each other makes observation easier, since interpersonal networks and media reports give more information on nearby events, so organizations that are geographically proximate to innovation events are likely to make changes. Second, the relative social and economic dominance of large places such as metropolitan areas leads to more efficient information transmission from large places to small, rather than the other way around, so innovations in large places will have greater effects on the rate of change. These geographical structures result in information transmission and influence from large cities to small cities and among nearby cities (Pred, 1977; Rogers, 1995). Localized diffusion has been shown in the spread of business practices (Jaffe, Trajtenberg, and Henderson, 1993; Pouder and St. John, 1996; Suchman, 1995; Greve, 1998a), but there is little evidence available on the diffusion from large cities to smaller ones (Pred, 1977).

Hypothesis 2 (H2): Innovations in the same geographical regions will increase the rate of change in an organizational population more than innovations farther away.

Hypothesis 3 (H3): Innovations in large cities will increase the rate of change in an organizational population more than innovations in small cities.

Competitive Structure and Strategic Change

Many industries are separated into multiple geographical markets because regulations or transportation costs separate customers. In analyzing the effect of innovations on strategic behavior in such multimarket industries, it is important to distinguish between effects on organizations in the same market as the innovating organization and effects on organizations in other markets. This distinction is important because the competitive impact of an innovative strategy can be great on organizations in the same market, but absent or negligible in other markets. Thus, innovations outside a focal market are just opportunities to learn and initiate search behavior, but innovations inside the focal market are threats to the organization’s current position. The local effect cannot be treated as a case of pure learning but must be considered a mix of learning and competition. To take this effect into account, we formulate separate hypotheses on the effect of innovations in local markets.

Innovations in the local market are both cognitive and competitive challenges, suggesting new ways of competing and threatening the established order of competition. All organizations in the market may become more likely to search, through the same cognitive processes that affect their response to innovations in general. Additionally, there is a competitive effect on organizations whose market shares are hurt by the innovation. Reduced market shares cause them to fall below their status quo (Kahneman and Tversky, 1979), increasing their risk-taking behavior and the likelihood of making changes (Greve, 1998b). Organizations generally respond to challenges in their market by making competitive counterattacks (Chen and Miller, 1994), so local innovations are likely to lead to change:

Hypothesis 4 (H4): Local innovations will increase the rate of change in the local market.

There is also heterogeneity in the response to local innovations. Normal diffusion arguments would imply that large and successful organizations are imitated (e.g., Haveman, 1993), but this argument has to be modified in local markets, where adopters of the same innovation will compete with each other. Local innovations are simultaneously cognitive opportunities and competitive threats, and when the innovator is large or successful, the threat aspect will become salient. Search caused by threatening events will cause responses that are restricted to well-established existing options, such as increased commitment to the current strategy (Staw, 1981; Dutton and Jackson, 1987; Ocasio, 1995). Innovations by large organizations are likely to be viewed as threats and to elicit a commitment response (Staw, Sandelands, and Dutton, 1981). Consequently, other organizations may be less likely to change in response to innovations if the innovator is a large firm. Similarly, if the innovator obtains a large market share, the innovation is more threatening and thus more likely to lead to a commitment response. Thus, innovations from large organizations or organizations with large market shares are less likely to cause changes:

Hypothesis 5 (H5): The larger the local innovating organization, the lower the impact of its innovation on the rate of change in the market.

Hypothesis 6 (H6): The higher the market share of the local innovating organization, the lower the impact of its innovation on the rate of change in the market.

STUDY

Radio Broadcasting

We tested these propositions with data on format changes in radio broadcasting. This is a good setting for testing the hypotheses, as radio markets are geographically separated and demographically differentiated. The separation allows comparison of innovations within and outside the focal market. The differentiation makes radio markets difficult to compete in, creating a great need for innovation and search behavior by the stations. Radio markets are differentiated because audiences differ in tastes for music and other programming material (Steiner, 1952). Taste differences correlate with demographic characteristics but are so diverse and difficult to observe that radio programming is “extraordinarily complex and dynamic” (Keith, 1987: xvi). Radio stations usually select a format, which is a package of music or non-music material, scheduling, and announcer style, and broadcast that format all day and night (Keith, 1987; Leblebici, 1995). The audience of a station depends on the station’s choice of format as w ell as the formats of competing stations. A good choice of format can yield a large audience and high advertising revenue, but it is difficult to find a format that appeals to a large audience and is not already being used.

This paper uses data on format adoptions from 1984 through 1992 to trace the impact of four innovative music formats: Soft Adult Contemporary (Soft AC), Soft Urban Contemporary, Urban Contemporary, and New Age. Soft AC is described as “almost wholly non-current, soft rock originals; can also be mixed with adult standards” (M Street Corp., 1992: 11). It is designed to appeal largely to female members of the baby-boom generation. Soft Urban is non-contemporary, with many oldies and a selection of Black music and sometimes jazz. Urban Contemporary is also oriented toward Black music but has a contemporary selection and younger listeners. New Age is mostly instrumental originals by contemporary artists, sometimes mixed with soft jazz.

Each of these formats may have caused search behaviors by radio stations, and they may even have reinforced each other through a process of cognitive cross-fertilization. The earliest and most successful of the formats, Soft Adult Contemporary, demonstrated that the female members of the baby-boom generation responded to quiet music, short and unobtrusive announcements, and long sweeps of music without the frequent interruptions by commercial and informational material (news, weather) seen in many other formats. A radio manager seeking to draw inferences from this innovation could obtain different ideas that might spur search behavior. First, bright-sounding personality-type announcers are not popular with all listeners, so one might try more subdued announcements in other formats (some Country and Western stations do this). Second, long sweeps of music are popular with the audience, so this form of programming could be tried in other formats as well (many stations have experimented with the duration of music sweeps). Third, there is a market for background music, and this market can register in the audience measurement. Station managers often suspect that background listening is underreported because the dominant audience measurement agency, Arbitron, uses a diary survey method and some respondents forget what stations they have listened to. New Age was also a background listening format, and its adoption would have been supported by this reasoning. Fourth, the baby-boom generation is so large that even narrowly focused formats can be profitable. New Age and Soft Urban Contemporary appeal to this age group, and their adoption would have been justified by such reasoning. As in these examples, some ideas from seeing innovations by others result from awareness of new dimensions of the market, allowing organizations to differentiate themselves in ways that would not be possible in a market with fewer known preference dimensions (Peli and Nooteboom, 1999).

Using a number of different sources, we were able to estimate the births of these formats. Interviews in stations programming Soft Adult Contemporary and New Age set the origin of those formats to about 1983 and 1987, respectively. Our source for the format adoption data, the M Street Journal, noted Soft AC adoptions beginning in 1984 and New Age adoptions since 1986 (only two were adopted that year, which explains why the broadcasters interviewed were unaware of them). Soft Urban has appeared in the M Street Journal since 1984, though with only two adoptions in 1984 and three in 1985. This slow trickle of adoptions suggests that the format was new at the time. The start of Urban Contemporary is more difficult to place because it has evolved by updating the music selection of older Black-oriented formats. The distinctive feature of current Urban Contemporary is that it includes rap music, which one source estimates did not have significant play in radio until the late 1980s (Lopes, 1992). Some stations starte d using Urban Contemporary earlier, however, and the format has been seen in the M Street Journal since 1984. These starting dates do not exclude the possibility that some innovators were using these formats earlier, but they do mark the approximate time that these formats were becoming noticed in the industry, suggesting that they were viewed as innovative during the study period.

Sample and Observation Period

We analyzed the search behavior of radio stations in U.S. radio markets where the audience shares were measured by the Arbitron audience measurement service. We used audience shares from a single source, Duncan’s American Radio, and the data reflect Duncan’s selection of which markets to follow. Duncan omits nearly 100 of the 261 markets measured by Arbitron in 1991, mostly small markets (below 200th rank), resulting in a sample size of 162 markets. Missing data on the advertising revenue reduced this to 157 markets and 4,970 market-quarter observations, 214 less than the 5,184 (eight years, four quarters, 162 markets) than would have been possible with no missing data. Data on format changes were obtained from the M Street Journal, a weekly newsletter reporting format changes and other events from the radio industry. This newsletter has national coverage, which assures that we would miss no innovations when measuring the nonlocal independent variables.

The dependent variable is the quarterly count of format changes of each market. The observation window for the independent variables is one year, so we counted the number of format changes in the year prior to the focal quarter. This means that innovative format adoptions of all of 1984 are used to predict change in the first quarter of 1985, the innovative adoptions of the three last quarters of 1984 and the first quarter of 1985 are used to predict change in the second quarter of 1984, and so on. Considerations of the reaction time and attention of radio stations guided the timing choices. Radio stations need some time, estimated as at least three months, to plan and implement format changes, so a three-month observation window for the dependent variable should ensure that contagion within the observation window does not occur. The response to recent events is greater than the response to earlier events, and a one-year observation window of the independent variables should reflect the attention span of deci sion makers. This is not meant to suggest that radio managers are forgetful, only that events more than a year old have low strategic importance.

Model and Variables

The theory and hypotheses predict that the likelihood of a given organization making a format change is affected by the number of local and nonlocal innovations and the characteristics of these innovations. This theory can be used to build a model of the count of changes of each market in a given time period. First, each station’s probability of change is affected by the number of local market innovations [I.sub.m] and nonlocal innovations [I.sub.i] (i means industry), which can be specified as:

P(C) = [[alpha].sub.0] + [sum][[alpha].sub.m][I.sub.m] + [sum][[alpha].sub.i][I.sub.i].

Second, each innovation’s influence depends on its characteristics, such as the size of the corporation that makes it. Modeling the influence of a given innovation as a sum of an intercept and a vector of variables leads to [[alpha].sub.z][I.sub.z] = [[alpha].sub.z0][I.sub.z] + [[alpha].sub.z1] [I.sub.z][x.sub.z1] + [[alpha].sub.z2][I.sub.z][x.sub.z2], where z is either m (market) or i (industry) and [x.sub.1] and [x.sub.2] are covariates describing innovation characteristics. This leads to:

P(C) = [[alpha].sub.0] + [sum][[alpha].sub.m0][I.sub.m] + [sum][[alpha].sub.m1][I.sub.m][x.sub.m1] + [sum][[alpha].sub.m2][I.sub.m][x.sub.m2] + [sum][[alpha].sub.i0][I.sub.1] +

[sum][[alpha].sub.i1][I.sub.i][x.sub.i1] + [sum][[alpha].sub.i2][I.sub.i][x.sub.i2].

The coefficients ([alpha]) can be taken outside the sums, so the covariates that affect each station are the number of innovations locally and non-locally and the sum of the covariates that influence how influential these innovations are. Thus, the variables in the analysis will be counts of innovations locally and non-locally and sums of the innovation characteristics. Similar methods of aggregating influences that differ by measurable covariates have been used to model the diffusion of innovations (Strang and Tuma, 1993) and organizational competition (Barnett, 1997).

The number of changes in a market can be modeled as a Poisson process in which the expected number of events is the product of the number of stations in the market and the influence on each station (exponentiated to avoid predicting negative numbers of events). Since markets may have unobservable differences that affect the number of changes in the market, it is useful to include a random effect that differs for each market but stays the same over time within each market. This can be done with Guo’s (1996) negative multinomial regression model for clustered random effects:

[[mu].sub.mt] = [N.sub.mt] exp([[alpha].sub.0] + [[alpha].sub.m0][sum][I.sub.mt] + [[alpha].sub.m1][sum][I.sub.nt][X.sub.m1t] + [[alpha].sub.m2][sum][I.sub.mt][X.sub.m2t] +

[[alpha].sub.i0][sum][I.sub.it] + [[alpha].sub.i1][sum][I.sub.it][X.sub.i1t] + [[alpha].sub.i2][sum][I.sub.it][X.sub.i2t]) [[theta].sub.m].

Here, [N.sub.m] is the number of stations in the market, and [[theta].sub.m] is the market random effect. The model was estimated by the estimation program supplied by Guo. Including influence from prior events and unmeasured heterogeneity in a Poisson-class model is a well-known technique from research on organizational foundings (Ranger-Moore, Banaszak-Holl, and Hannan, 1991; Land, Davis, and Blau, 1994) and political conflict (Rasler, 1996). Our method is distinctive in analyzing multiple markets over time with disturbances that differ by market (Guo, 1996) and in measuring how the influence of each event differs depending on event characteristics.

Dependent variables. The analyses have five different events as dependent variables. The most inclusive event consists of all changes in format or production, effectively counting all entries in the M Street Journal. This is called all changes. To investigate whether different types of change produced different reactions to innovations, we also analyzed four subevents separately. The first subevent, innovative change, consists of entry into one of the innovative formats, Soft Adult Contemporary, New Age, Urban Contemporary, and Soft Urban Contemporary. This event represents mimetic behavior, and since it omits the most common formats (Adult Contemporary and Country and Western) and many less common but older formats (such as Easy Listening and Album-oriented Rock), it has substantially fewer events. The second subevent, format change, consists of all format changes except entry into innovative or satellite formats, and is thus a nonmimetic change of format. The third, satellite entry, counts all format change s in which the destination format is a satellite format. These are provided by programming services that sell, for money or for a portion of the advertising time, ready-made programming in a number of different formats. Satellite programming was a novel way to reduce operating costs by eliminating announcers and programming staff, so satellite entry is both an operational and a format change. The fourth, production change, is an operational change that consists of all changes among the production modes of live, simulcast, or satellite that do not also change the format. These categories are mutually exclusive, so each event occurs only in one subanalysis.

Independent variables. The following independent variables were entered, all measured as sums of the relevant covariates over the year before the current period. Market-level innovations previous year is the number of adoptions of innovative formats in the local market. innovations weighted by market share is the sum of market shares (after the change) for these innovations. Innovations weighted by size of corporation is the sum of the corporate sizes, measured as number of stations, of the corporations owning the stations making the changes. Nonlocal innovations previous year nationally is the number of adoptions in other markets. Innovations previous year, regionally is the number of nonlocal innovations in the same region of the country (using similar regions as Walker, 1969). Innovations in top-lO markets is the number of innovations in the 10 largest (by population) markets. [2] The nonlocal variables do not include events in the local market, so each innovation is only counted once.

Control variables. We added variables capturing organizational and market characteristics. First, branches of multimarket organizations are likely to be influenced by their corporate centers and other branches. Such organizations are likely to seek to transfer what they learn about competition in one market to their other markets, which leads them to copy innovations and smaller changes across branches (Darn Argote, and Epple, 1995; Greve, 1995, 1996; Ingram and Baum, 1997). Thus, the number of branches in the market will positively affect changes in the market and needs to be controlled for. Second, because radio formats usually target specific demographic groups, we also controlled for the demographic composition of the market. From the 1990 census data (metropolitan or urbanized area level, depending on the market) we computed the percentage of persons who were male, White, ages 20-34, ages 35-49, and ages 50 and over. Two of these age groups are of special interest. The cohort of persons then 35-49 was vi ewed as lucrative and was the target of much format innovation, and the cohort of persons 50 and over was viewed as a declining market with few opportunities. Thus, one might expect a high percentage of persons between 35 and 49 to increase the rate of changes and a high percentage over 50 to decrease it.

Innovations accumulated is the number of innovations nationwide from 1984 to the quarter preceding the focal quarter. This variable controls for traditional diffusion arguments in which the total number of innovations adopted legitimizes the innovation (and perhaps also other changes). Non-innovative changes are the number of changes in the local market during the previous year that are not adoptions of innovations. This variable was included because format changes may occur as responses to competitive moves by other organizations, even if these are not innovative (Chen and Miller, 1994). Advertising revenue growth was coded from Duncan’s American Radio and controls for the economic conditions of each radio market. This variable is an estimate made by Duncan, as radio stations rarely release advertising revenue figures, but since it is an aggregate for the market, it is more precise than an estimate of individual station earnings would be.

Organizations in general are more likely to change when their performance is below their aspiration level (Lant and Montgomery, 1987; Lant, Milliken, and Batra, 1992), and radio stations change more often when their market share is lower than the mean share in the market (Greve, 1998b). Individual stations’ market shares cannot be included in the model, since each observation is a market, so we included the semi-variance of market shares, which is the sum of squared negative deviations from the mean market share. Since a high semi-variance means that many stations perform below the market mean, we expected a positive coefficient estimate. [3]

Data on radio format adoptions and audience shares have previously been used to analyze the diffusion of innovative formats (Greve, 1995, 1996, 1998a) and risk taking in organizational change (Greve, 1998b). The studies on diffusion showed clear evidence of mimetic processes, including mimetic adoption of innovations made in other markets, but the study of risk taking revealed an interesting anomaly. While adoptions of innovative formats were important in the sense that they spread new market positions across markets and increased consumer choice, they constituted a small minority of the total change behavior. Nine out of ten changes were not to innovative formats, but it was clear that they were attempts to improve the competitive positions of the focal stations, since they were nearly always made when the stations had low performance relative to their aspiration level (Greve, 1998b). Thus, most of the change behaviors were not imitations of innovations. This observation was important to recognizing the need for a theory in which innovations cause general search behavior and change in addition to imitation.

RESULTS

Table 1 provides descriptive statistics and correlation coefficients for the independent and dependent variables of the models. Approximately 49 total innovative format adoptions were made per year in the overall industry, which is much less than one per market per quarter. Innovations are rare occurrences in radio markets, as is expected if they represent significant changes in a station’s strategy and operations. The correlation matrix shows modest to high correlations among the national-level innovation variables, but no other high correlations. The demographic variables are time-constant and market-varying and thus have zero correlation with the time-varying and market-constant national innovation variables.

Aggregate Model of Change

Table 2 reports the estimates of the models of aggregate changes spurred by innovations. Model 1 only has the market-level control variables and provides a base model of search driven only by local market opportunities and station characteristics. The only significant demographic variable is percent aged 50 years and over, which is negative as expected. Advertising revenue growth, branch stations, and semivariance of market share are all significant, with the expected signs. Favorable economic conditions buffer stations from changing, while branch status and low performance make them more likely to change. The number of new stations per year and non-innovative changes per year are also significant and positive, suggesting that radio stations respond to threats in their local market by making format changes.

Model 2 examines whether local market innovations affect the rate of change. We predicted that innovations the previous year would promote change (H4), but their effect on the rate of change would be less if the innovation was adopted by large corporations (H5) or successful stations (H6). Hypotheses H4 and H5 were supported at the .01 and .05 significance levels, respectively, while H6 was not significant but had the expected negative coefficient. It appears that previous innovations do spark changes, and the characteristics of who has implemented the change do matter. The impact of size but not market share is similar to Freeman and Lomi’s (1994) findings on bank foundings. They found that capacity, as measured by the size variable, was the significant influence, rather than demand, as measured by market share. Here, we also find that innovations by larger corporations are seen as threatening, but it seems that stations with large market shares can still be considered vulnerable to competitive action. The s ignificance levels and relative magni-tudes of the control variables stayed the same.

Model 3 examines whether nonlocal innovations affect the rate of change. Our first three hypotheses predicted that previous innovative activity would spark industrywide (H1) and regional change (H2), which would be especially strong if innovation occurred in a large market (H3). Model 3 shows that all three hypotheses are supported at the .05 significance level or better. Thus, the most straightforward tests of our arguments were substantiated. Innovations at the national level, which have no direct impact on the competitive situation in the local market, cause an increased rate of change. The effect is strengthened by the geographic proximity and size of the market in which the innovation originated. The visibility of the innovation influences the attention and impact of innovations made outside the focal market. These findings are robust, as model 4 entered all variables at once, yielding reductions in the level of significance for the local variables, while the nonlocal variables maintained their significa nce. The coefficients of several market-level control variables lost their significance when the nonlocal innovation variables were included.

The local-level results show that innovative activity has a mixed effect on change in the local market. Innovations cause organizations to broaden their search activities, but their effect on search decreases when the innovator is in a dominant position, as measured by corporate size. At the industry level, innovations broaden search activity, and this effect is heightened when it occurs within the same region or in a highly visible market. The results thus support the general model of innovation as a catalyst for changes due to active search. Because a key part of our argument is that the changes observed represent more than imitation of the innovation, to complete our inquiry, we need to examine whether non-imitative changes were affected by innovations in the same way as the aggregate number of changes.

Disaggregated Models of Change

Table 3 examines the disaggregated analysis of change in response to innovations. Model 1 is the same as model 4 from table 2 and is included for ease of comparison. The next four columns show the estimates of models with the same independent variables and different dependent variables. Innovative change represents imitation and the other changes (format, satellite entry, and production changes) represent non-imitative change. For our overall argument to hold, innovations must influence the non-imitative changes modeled in columns three, four, and five.

Model 2 shows innovative changes as a result of past innovations, or simple imitation. No local market level variables are significant, so it appears that innovative activity does not spur innovative change in the same market. This is consistent with our earlier discussion of the competing influences of perceived opportunity and exhaustion of capacity, but it could also result from the low number of events in this model, since innovative changes are infrequent. Innovations outside the focal market but in the same region or in large markets are significant, supporting hypotheses H2 and H3. The main effect is not significant, however, suggesting that only the most easily observed innovations are imitated. Such selective imitation of the most salient adopters has some precedents in recent diffusion studies (Burns and Wholey, 1993; Haveman, 1993; Davis and Greve, 1997). Unlike the aggregate model, model 2 shows a significant effect of the branch-station control variable, so innovations are transferred among organ izations in branch systems.

The models of format changes, satellite entry, and production changes examine non-imitative change, in which the dependent variables are changes other than imitating the innovation. The nonlocal variables show positive and significant effects of national (H1), regional (H2), and large-market innovations (H3) for satellite and production changes. Change by satellite entry or production change seems to be spurred by innovations observed in other markets, and nearby markets are more influential than markets farther away, as one would expect from observability criteria. The local variables show that innovations spur format changes (H4) but do so less as the size of the innovating branch system increases (H5), showing that format change occurs in response to innovations, but to a lesser extent if the innovator is not a branch station.

The findings show an interesting difference between results for format changes, in which only local innovations are significant, and those for satellite and production changes, in which only nonlocal innovations are significant. It seems that format change is the regular response to locally threatening innovations, while satellite and production entry is spurred by nonlocal opportunities. This finding is interesting because some have theorized that threatening events lead to low-risk responses (Dutton and Jackson, 1987), while others have found that threatening events are vigorously responded to (Chen and Miller, 1994). Since format changes are generally more risky than satellite and production changes, our finding supports the latter argument. An explanation specific to radio broadcasting is also possible, however. Satellite programming and one form of production change (local market agreements, in which two stations with different owners share programming) were innovative ways of programming radio stations during this period and may have been legitimated by the observation of innovative formats.

The control variables produced few results, which is not surprising, since the model has a high number of variables relative to the number of events. The demography variable percent aged 50 and over is always negative and sometimes significant. Percent male and percent White are significant for one and two outcomes, respectively. The semi-variance of market shares is positive and significant for format changes only, suggesting that they occur in response to low market share. Branch stations appear to be prone to make innovative changes and are unlikely to enter satellite formats. Both findings were expected, since branch systems transfer innovative formats among their members and have the resources to produce live programming rather than purchase satellite feeds. New stations appear to spur format and satellite changes, which could be a result of the new stations themselves entering satellite formats in response to poor performance of live formatting. Non-innovative changes are positively related to satellite entry and production changes, so very dynamic markets are also markets in which stations try new forms of production.

The results strongly support our conceptualization of innovations as change catalysts in the industry, as innovations clearly increase the search intensity and rate of change. Unexpectedly, the type of change differed by the location of the innovation, as new forms of production occurred as a response to nonlocal innovations, while new formats were tried in response to local innovations, suggesting that more threatening innovations resulted in stronger, or at least more risky responses. Innovations by noncompeting organizations were especially influential when they occurred in markets highly visible to managers, such as nearby or large markets. Innovations were less influential when they were done by large corporations in the market, suggesting that a rigid response occurred when an innovation was threatening. These findings were robust across outcomes and specifications. [4]

DISCUSSION AND CONCLUSION

Do innovations by others just cause mimicry or do they have a wider influence? We have argued that innovations by others encourage change by suggesting new opportunities and making search more acceptable to organizations. When faced with uncertainty, organizations interweave search and action as they conduct active search to gain feedback and information on the shifting external environment. We found that prior innovations at both the nonlocal and local market levels result in changes by industry participants. Interestingly, the weakest effects were in the model of innovations. It seems that not only are organizations doing more than mimicking innovation but that most of the activity occurring is not innovation but more traditional strategic and operating changes. The extent to which a specific innovation spurs change varied with the associated corporation and market characteristics. These moderating effects were as we predicted from theory of the salience of nonlocal innovations and the preemption of opportu nities for local market innovations. Innovations in large or nearby markets were especially influential because they were salient; local innovations by large corporations were less influential because they were threatening and caused commitment responses.

Non-imitative Change

These findings suggest a new research direction that differs from the current emphasis on convergence among organizations. Institutional theory predicts the convergence of forms (Scott, 1995), and macroculture arguments describe convergence of beliefs based on value networks among communities of organizations (Abrahamson and Fombrun, 1994). These are different theories of how managers come to think and act the same way. Institutional theory describes mimetic isomorphism as the process of an organization modeling itself after organizations perceived to be legitimate or successful as a response to uncertainty (DiMaggio and Powell, 1983). Macroculture focuses on the introduction of uncertainty and predicts that cultural homogeneity will lead to greater diffusion of innovations and a lower rate of adoption of innovations outside of the community (Abrahamson and Fombrun, 1994). Both arguments predict adoptions of behaviors similar to the innovation introduced into the system. By contrast, we have argued that innov ations cause many non-imitative changes.

Our organization-level argument mirrors micro-level research on minority influence in group situations. It is well documented that exposure to the dissenting views of a minority fosters broader thought around an issue and stimulates divergent rather than convergent search (Nemeth, 1986). Nemeth and Rogers (1996) found that minority dissent stimulated search for more and broader information. Gruenfeld and Kim (1998) found that minority opposition in Supreme Court rulings caused the majority to display greater cognitive flexibility and consider issues from multiple perspectives. Innovation can also be conceptualized as a minority view because it is outside of the dominant behaviors of the industry and can be seen as unfamiliar or even inappropriate by other actors. Taylor (1998) found that autonomous innovation attempts outside of a firm’s current strategic direction foster far-ranging top management discussion and debate about organizational actions focused not only on the support or rejection of the innovativ e attempt. The introduction of an innovative minority provokes cognitive activity across all dimensions, including but not limited to the focus of the innovation.

Innovation and Decision-making Uncertainty

One of our fundamental assumptions, and the emphasis of this study, is that innovations are sources of uncertainty in a market environment. Uncertainty in decision-making situations prevents organizations from applying fully rational decision-making processes. Issues of uncertainty in innovation have been examined at the technological level with an emphasis on connection with past knowledge (Henderson and Clark, 1990) and on the uncertainty of acquiring the appropriate skills (Lippman and Rumelt, 1982). Here, we link the literatures of innovation with those of decision making, as innovation is an endogenous shock generated by the activities of others in the industry. The shock results in industry participants undertaking activities despite limited information. The results are particularly striking for nonlocal innovation, which, with no direct competitive impact, sparked changes in the market.

Importantly for this study, we found that nonmimetic changes ranging from strategic (format change) to operational (satellite and production changes) occur as a result of innovation. This shows that innovations can cause organizations to look for other ways to compete, and not only in the direction of the innovation. This is an interesting middle ground between theory suggesting that organizational change occurs as organizations develop their own established competencies (Reed and DeFilippi, 1990; Barney, 1991) and theory suggesting that organizational change occurs as a result of pressures to imitate others (DiMaggio and Powell, 1983). Innovations have an uncertain competitive impact, represent unfamiliar values and meanings, and have unclear appropriateness (March, 1994). The resulting ambiguity gives rise to other innovations and new practices that shape the context of future behaviors.

Since most changes were not imitations of the innovations that triggered the search, these findings clearly differ from those explained by mimetic isomorphism (DiMaggio and Powell, 1983). It is possible, however, that our findings reflect that innovations legitimize the process of format change in addition to the particular format changes that are introduced. Investigation of whether change can be institutionalized seems like a promising research topic for institutional theory but would require a reorientation from the present focus on the outcomes of change to the process of change as the dependent variable. It also requires shifting the emphasis from isomorphism to variation-generating institutional processes (e.g., Suchman, 1995; Rura-Polley, 1999). While the theory of this paper concerns innovations in a competitive market, its foundation in the bounded rationality of managers facing an uncertain environment is the same as the cognitive branch of institutional theory (Scott, 1995). Thus, applying these ar guments to predict change in institutional environments seems like a promising path to broaden the scope of institutional theory.

Another difference from studies of mimetic isomorphism is our finding that the search effect was dampened by competition. Threat rigidity responses caused innovations by large firms in the same market to be less likely to lead to search, so the process of catalytic change slowed down when large organizations adopted innovations. Interestingly, large organizations were particularly likely to transfer innovations among markets since their presence in multiple markets enables their managers to discover innovations early (Greve, 1996), so this dampening effect occurred quickly. Our findings thus provide additional evidence for the often-observed pattern of a period of technological ferment followed by stability (Utter-back and Abernathy, 1975) and suggest that the mechanism driving this pattern is positive feedback through catalytic change slowed by the dampening effect of threat rigidity responses to competition.

Our study also suggests that those conducting diffusion studies must be careful in identifying the adoption of innovative practices, as organizations may cloak more traditional changes in the rhetoric of the innovation (Westphal and Zajac, 1994). Most of the observed changes were nonmimetic, but if we had stopped at coarser measures, we might have coded the behavior as imitation, or at least diffusion of innovative behavior. The finer-grained models showed that the innovative behavior was imitated by some, but nonmimetic behavior was also caused by the cognitive challenge of innovations.

Industry Cognition, Innovation, and Strategic Action

These findings suggest that innovations influence cognitions according to the main theses of our theory. First, innovations change managerial cognitions on what actions are possible and beneficial and thus allow organizational change to take place. For cognition to be affected, managers must be linked as a community of self-defined participants. Our finding that managers respond to actions of organizations outside of their local market show that such a community exists and influences their strategic action. The effect of innovations on search was cognitively moderated, as it was affected by both salience and by strategic considerations of whether there were open opportunities or not. Such effects depend on how managers structure their attention and how they think about the community of competitors in their industry. Thus, the importance of innovations varies depending on how managers process them cognitively, and their effect is due to their influence on managerial cognitions about organizational opportunitie s. The new stimulus can change managerial schemas directly or cause search that stimulates updating of schemas.

Scope of Theory and Research Opportunities

The role of innovations as catalysts of change may help explain the many sudden changes in rates of innovative and non-innovative change seen in various industries. Sometimes it is even possible to pinpoint one innovation that starts an industry down the path of change. A case in point is the “beer war” in Japan, which started in 1986 with the reformulated Asahi Draft (a precursor of the better-known Asahi Super Dry), a beer with a sharper taste than the then-dominant lager beers and whose success among young people revealed cohort differences in beer preferences. This innovation was followed by several waves of innovation and search, such as packaging innovations, seasonal beers, and non-beer malt beverages, along with investments in organizational capabilities to make product change a greater part of the organizational activities of the producers (Craig, 1996). A similar industrywide search process may follow the redefinition of healthcare implicit in such pharmaceutical innovations as hair regrowth compoun ds and male potency pills.

The theory of catalyst effects has the potential to explain firms’ responses to innovations in uncertain environments and seems particularly relevant to industries with high uncertainty in customer preferences (such as many consumer industries) or technology (such as industries with recent or rapidly changing technology). The theory may be less appropriate for industries in which uncertainty is low or in which linkages between suppliers and customers make it difficult for one party to initiate change unilaterally. Some subassembly industries may be of this form, since the main task for such firms is to be responsive to the demands of the dominant organizational form. In such cases, it may be more appropriate to view innovations in the dominant industry as drivers of search behavior in the dependent industries. The role of automobile manufacturers in driving search in related industries, such as ceramics (for engine parts) or batteries (for electric automobiles), is an example of search driven by interindustry connections rather than by the intraindustry process described here. Though the underlying cognitive process should be similar in such cases, the specific predictions would differ.

In this paper, we sought to support the basic hypothesis that innovations lead to active search and to go further by showing how this effect is mediated by the social structure of the industry. While the evidence is promising on both counts, the study has some limitations and oversights that future research may correct. First, it is a common observation that all innovations are not equally influential, and the innovation literature has explained this by the theory of competence-destruction (Tushman and Anderson, 1986) and architectural change (Henderson and Clark, 1990). Because only four innovations occurred in these data, we were unable to test whether differences in innovation characteristics modified their effect on search intensity, but it follows from the theory that such differences should be found. We hope that this limitation can be addressed by future research that draws on a larger pool of innovations. Second, we were not able to show whether the direction of active search was toward markets or act ivities related to the observed innovation. This is also a prediction of the model and should be testable with more fine-grained data on market structure or firms’ resource allocations. The strategic groups literature contains work measuring the direction of strategic change (e.g., Fiegenbaum and Thomas, 1995), and such techniques can be adapted for further study of active search. Third, we did not explore organizational differences in the initiation of actionable search but believe that the usual inertia and social embeddedness effects should hold also for this outcome. We have only provided evidence on a small proportion of the hypotheses that can be derived from the theory, leaving much work for future research.

The importance of innovations is clearly understated if only imitation is considered, since the newness introduced into the system has so many other effects. Despite the wealth of research on innovative activity, researchers have only started to appreciate the full range of impact that innovations represent in competitive markets. A cognitive view provides a promising perspective for studying the effect of innovations because it explicitly models management’s limited attention and inability to act based on analysis and evaluation. Managerial actions are based on decision-making heuristics triggered and altered by stimuli such as innovations. Distinct from economic arguments, the cognitive perspective posits behavior driven not by proactive analysis but by reaction to scarce and ambiguous information. Viewing innovative activity through this lens shifts the investigation from how to exploit the opportunities represented by an innovation to the ways innovations shift managerial perceptions, understandings, appr opriate behavior, and risk taking. This line of inquiry is especially appropriate for understanding organizational behavior in the fast-changing uncertain environments faced by many organizations.

Henrich R. Greve [“Innovations as Catalysts for Organizational Change: Shifts in Organizational Cognition and Search”] is an associate professor in the Institute of Policy and Planning Science at the University of Tsukuba, Tsukuba-shi, Ibaraki 305-8573 Japan (e-mail: greve@shako.sk.tsukuba.ac.jp). His current research includes work on the effect of aspiration-level risk on investment decisions and on the estimation of heterogeneous social influence in hazard rate models (with N. B. Tuma). Recent publications include “Market Niche Entry Decisions: Competition, Learning, and Strategy in

Tokyo Banking, 1894-1936″ (Academy of Management Journal, forthcoming) and “Branch Systems and Nonlocal Learning in Populations” (Advances in Strategic Management, 16:57-80). He received his Ph.D. in organizational behavior from the Graduate School of Business, Stanford University.

Alva H. Taylor [coauthor, “Innovations as Catalysts for Organizational Change: Shifts in Organizational Cognition and Search”] is a visiting assistant professor at the Kellogg Graduate School of Management, Northwestern University, 2001 Sheridan Road, Evanston, IL 60208 (e-mail: a-taylor3@nwu.edu). His current research interests are the social and cognitive influences of innovation attempts and adoption, with particular interest in the domain of entrepreneurial activity and e-commerce. He is presently involved in projects examining the role of cognition in the formation of entrepreneurial communities, and responses to large-scale radical change in the financial and entertainment industries. He is completing his Ph.D. in organizational behavior at the Graduate School of Business at Stanford University.

(*.) We are grateful for comments from Richard Harrison, Mark Mizruchi, Keith Murnighan, James G. March, Anne S. Miner, Marc Ventresca, Ed Zajac, participants in the Kellogg Graduate School of Business OB Strategy Seminar, Christine Oliver, Linda Johnson, and three anonymous reviewers.

(1.) See Barnard’s (1938) Appendix, section III, “The Nature of the Material to Which the Mind is Applied.”

(2.) Initially, we used a continuous measure of market size that added the ranks of markets. That measure gave the same results as those we present here, but we decided not to use it because it was highly correlated with two other measures in the analysis.

(3.) The regular variance could also be used but is less appropriate, since high positive deviations from the market mean should not lead to changes. We tried it and found that it did give lower estimates than the semi-variance. We also tried the square root of the semi-variance and got similar but weaker results.

(4.) Because we were concerned that the positive correlations of the national and regional innovation variables could cause artifactual results, for all events we estimated the saturated model in table 3 and models dropping one of each of the variables–regional innovations, innovations weighted by corporate size, and innovations in top-10 markets. All results were maintained in these specifications. We did not drop annual number of innovations, as the coefficients of the weighted variables do not make sense without this main effect.

REFERENCES

Abrahamson, E., and C. J. Fombrun

1994 “Macrocultures: Determinants and consequences.” Academy of Management Review, 19: 728-755.

Argyris, C.

1996 “Actionable knowledge: Design causality in the service of consequential theory.” Journal of Applied Behavioral Science, 32: 390-406.

Axelrod, R. M.

1972 Framework for a General Theory of Cognition and Choice. Berkeley, CA: Berkeley Institute of International Studies, University of California.

Barnard, C. I.

1938 The Functions of the Executive. Cambridge, MA: Harvard University Press.

Barnett, W. P.

1997 “The dynamics of competitive intensity.” Administrative Science Quarterly, 42:128-160.

Barney, J.

1991 “Firm resources and sustained competitive advantage.” Journal of Management, 17: 99-120.

Barr, P. S.

1998 “Adapting to unfamiliar environmental events: A look at the evolution of interpretation and its role in strategic change.” Organization Science, 9: 644-669.

Bartunek, J. M.

1993 “The multiple cognitions and conflicts associated with second order organizational change.” In J. K. Murnighan (ed.), Social Psychology in Organizations: Advances in Theory and Research: 322-349. Englewood Cliffs, NJ: Prentice Hall.

Baum, J. A. C., and H. A. Haveman

1997 “Love thy neighbor? Differentiation and agglomeration in the Manhattan hotel industry, 1898-1990.” Administrative Science Quarterly, 42: 304-338.

Bikhchandani, S., D. Hirshleifer, and I. Welch

1992 “A theory of fads, fashion, custom, and cultural change as informational cascades.” Journal of Political Economy, 100: 992-1026.

Burgelman, R. A., and L. R. Sayles

1986 Inside Corporate Innovation: Strategy, Structure, and Managerial Skills. New York: Free Press.

Burns, L. R., and D. R. Wholey

1993 “Adoption and abandonment of matrix management programs: Effects of organizational characteristics and interorganizational networks.” Academy of Management Journal, 36: 106-138.

Chen, M.-J., and D. Miller

1994 “Competitive attack, retaliation and performance: An expectancy-valence framework.” Strategic Management Journal, 15: 86-102.

Child, J., and C. Smith

1987 “The context and process of organizational transformation: Cadbury Limited in its sector. Journal of Management Studies, 24: 565-593.

Cohen, M. D., J. G. March, and J. P. Olsen

1972 “A garbage can model of organizational choice.” Administrative Science Quarterly, 17: 1-25.

Cohen, W. M., and D. A. Levinthal

1990 “Absorptive capacity: A new perspective on learning and innovation.” Administrative Science Quarterly, 35:128-152.

Craig, T.

1996 “The Japanese beer wars: Initiating and responding to hypercompetition in new product development.” Organization Science, 7: 302-321.

Cyert, R. M., and J. G. March

1963 A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-Hall.

Daft, R. L., and K. E. Weick

1984 “Toward a model of organizations as interpretation systems.” Academy of Management Review, 9: 284-295.

Darr, E. D., L. Argote, and D. Epple

1995 “The acquisition, transfer, and depreciation of knowledge in service organizations: Productivity in franchises.” Management Science, 41: 1750-1762.

Davis, G. F., and H. R. Greve

1997 “Corporate elite networks and governance changes in the 1980s.” American Journal of Sociology, 103: 1-37.

DiMaggio, P. J., and W. W. Powell

1983 “The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields.” American Sociological Review, 48: 147-160.

Dutton, J. E., and S. E. Jackson

1987 “Categorizing strategic issues: Links to organizational action.” Academy of Management Journal, 12: 76-90.

Fiegenbaum, A., and H. Thomas

1995 “Strategic groups as reference groups: Theory, modeling and empirical examination of industry and competitive strategy.” Strategic Management Journal, 16: 461-476.

Fiske, S. T., and S. E. Taylor

1991 Social Cognition. New York: McGraw-Hill.

Freeman, J., and A. Lomi

1994 “Resource partitioning and foundings of banking cooperatives in Italy.” In J. A. C. Baum and J. V. Singh (eds.), Evolutionary Dynamics of Organizations: 269-293. New York: Oxford University Press.

Gersick, C. J. G., and J. R. Hackman

1990 “Habitual routines in task-performing groups.” Organizational Behavior and Human Decision Processes, 47: 65-97.

Greve, H. R.

1995 “Jumping ship: The diffusion of strategy abandonment.” Administrative Science Quarterly, 40: 444-473.

1996 “Patterns of competition: The diffusion of a market position in radio broadcasting.” Administrative Science Quarterly, 41: 29-60.

1998a “Managerial cognition and the mimetic adoption of market positions: What you see is what you do.” Strategic Management Journal, 19: 967-988.

1998b “Performance, aspirations, and risky organizational change.” Administrative Science Quarterly, 44: 58-86.

Gruenfeld, D. H., and R H. Kim

1998 “Dissent and decision making on the U.S. Supreme Court: Conflict with a minority predicts cognitive complexity in the majority.” Working paper, Kellogg Teams and Groups Center, Northwestern University.

Guo, G.

1996 “Negative multinomial regression models for clustered event counts.” In A. E. Raftery (ed.), Sociological Methodology, 26: 113-132. Cambridge, MA: Blackwell.

Hannan, M. T., and J. Freeman

1977 “The population ecology of organizations.” American Journal of Sociology, 82: 929-964.

1984 “Structural inertia and organizational change.” American Sociological Review, 49: 149-164.

Haveman, H. A.

1993 “Follow the leader: Mimetic isomorphism and entry into new markets.” Administrative Science Quarterly, 38: 593-627.

Henderson, R. M., and K. B. Clark

1990 “Architectural innovation: The reconfiguration of existing product technologies and the failure of established firms.” Administrative Science Quarterly, 35: 9-30.

Higgins, E. T., and J. A. Bargh

1987 “Social cognition and social perception.” Annual Review of Psychology, 38: 369-425. Palo Alto, CA: Annual Reviews.

Ingram, P., and J. A. C. Baum

1997 “Chain affiliation and the failure of Manhattan hotels, 1898-1980.” Administrative Science Quarterly, 42: 69-102.

Jaffe, A. B., M. Trajtenberg, and R. Henderson

1993 “Geographic localization of knowledge spillovers as evidenced by patent citations.” Quarterly Journal of Economics, 108: 577-598.

Jelinek, M., and C. B. Schoonhoven

1990 The Innovation Marathon. San Francisco: Jossey-Bass.

Kahneman, D., and A. Tversky

1979 “Prospect theory: An analysis of decision under risk.” Econometrica, 47: 263-291.

Keith, M. C.

1987 Radio Programming: Consultancy and Formatics. Boston: Focal Press.

Kelly, D., and T. L. Amburgey

1991 “Organizational inertia and momentum: A dynamic model of strategic change.” Academy of Management Journal, 34: 591-612.

Land, K. C., W. R. Davis, and J. R. Blau

1994 “Organizing the boys of summer: The evolution of U.S. minor-league baseball, 1883-1990.” American Journal of Sociology, 100: 781-813.

Lant, T. K., F. J. Milliken, and B. Batra

1992 “The role of managerial learning and interpretation in strategic persistence and reorientation: An empirical exploration.” Strategic Management Journal, 13: 585-608.

Lant, T. K., and D. B. Montgomery

1987 “Learning from strategic success and failure.” Journal of Business Research, 15: 503-518.

Larkin, J. H.

1983 “The role of problem representation in physics.” In D. Gentner and A. Collins (eds.), Mental Models: 75-98. Hillsdale, NJ: Erlbaum.

Leblebici, H.

1995 “Radio broadcasters.” In G. R. Carroll and M. T. Hannan (eds.), Organizations in Industry: Strategy, Structure and Selection: 308-331. New York: Oxford University Press.

Lippmanm, S. A. and R. P. Rumelt

1982 “Uncertain imitability: An analysis of interfirm differences in efficiency under competition.” Bell Journal of Economics, 13: 418-438.

Lopes, R D.

1992 “Innovation and diversity in the popular music industry, 1969 to 1990.” American Sociological Review, 57: 56-91.

Lyles, M. A., and C. R. Schwenk

1992 “Top management, strategy and organizational knowledge structures.” Journal of Management Studies, 29: 155-174.

M Street Corp.

1992 M Street Radio Directory. New York: M Street Corp.

March, J. G.

1981 “Footnotes to organizational change.” Administrative Science Quarterly, 26: 563-577.

1988 Decisions and Organizations. Oxford: Basil Blackwell.

1991 “Exploration and exploitation in organizational learning.” Organization Science, 2: 71-87.

1994 A Primer on Decision Making: How Decisions Happen. New York: Free Press.

March, J. G., and H. Simon

1958 Organizations. New York: Wiley.

Miller, D., and M.J. Chen

1996 “Nonconformity in competitive repertoires: A sociological view of markets.” Social Forces, 74: 1209-1234.

Miner, A. S., and P. R. Haunschild

1995 “Population level learning.” In L. L. Cummings and B. M. Staw (eds.), Research in Organizational Behavior, 17: 115-166. Greenwich, CT: JAI Press.

Nelson, R. R., and S. G. Winter

1982 An Evolutionary Theory of Economic Change. Boston: Belknap Press of Harvard University.

Nemeth, C. J.

1986 “Differential contributions of majority and minority influence.” Psychological Review, 93: 23-32.

Nemeth, C. J., and J. Rogers

1996 “Dissent and the search for information.” British Journal of Social Psychology, 35: 67-76.

Ocasio, W. C.

1995 “The enactment of economic adversity: A reconciliation of theories of failure-induced change and threat-rigidity.” In L. L. Cummings and B. M. Staw (eds.), Research in Organizational Behavior, 17: 287-331. Greenwich, CT: JAI Press.

1997 “Towards an attention-based theory of the firm.” Strategic Management Journal, 18: 187-206.

Peli, G., and B. Nooteboom

1999 “Market partitioning and the geometry of the resource space.” American Journal of Sociology, 104: 1132-1153.

Pouder, R., and C. H. St. John

1996 “Hot spots and blind spots: Geographical clusters of firms and innovation.” Academy of Management Review, 21: 1192-1225.

Prahalad, C. K., and G. Hamel

1990 “The core competence of the corporation.” Harvard Business Review, May-June: 79-91.

Pred, A.

1977 City-Systems in Advanced Economies. London: Hutchinson.

Quinn, J. B.

1986 “Innovation and corporate strategy: Managed chaos.” In M. Horwitch (ed.), Technology in the Modern Corporation: A Strategic Perspective: 167-183. Elmsford, NY: Pergamon.

Ranger-Moore, J., J. Banaszak-Holl, and M. T. Hannan

1991 “Density-dependent evolution in regulated industries: Founding rates of banks and insurance companies.” Administrative Science Quarterly. 36: 36-65.

Rasler, K.

1996 “Concessions, repression, and political protest in the Iranian revolution.” American Sociological Review, 61: 132-152.

Reed, R., and R. J. DeFillippi

1990 “Causal ambiguity, barriers to imitation, and sustainable competitive advantage.” Academy of Management Review, 15: 88-102.

Rogers, E. M.

1995 Diffusion of Innovations, 4th ed. New York: Free Press.

Romanelli, E., and M. L. Tushman

1994 “Organizational transformations as punctuated equilibrium: An empirical test.” Academy of Management Journal, 37:1141-1166.

Rura-Polley, T.

1999 “Constructing variation: Insights from an emerging organizational field.” In A. S. Miner and P. Anderson (eds.), Advances in Strategic Management, 16: 249-275. Greenwich, CT: JAI Press.

Scott, W. Richard

1995 Institutions and Organizations. Thousand Oaks, CA: Sage.

Starbuck, W. H.

1992 “Learning by knowledge-intensive firms.” Journal of Management Studies, 29: 713-740.

Staw, B. M.

1981 “The escalation of commitment to a course of action.” Academy of Management Review, 6: 577-587.

Staw, B. M., L. E. Sandelands, and J. E. Dutton

1981 “Threat-rigidity effects in organizational behavior: A multi-level analysis.” Administrative Science Quarterly, 26: 501-524.

Steiner, P. O.

1952 “Program patterns and preferences, and the workability of competition in radio broadcasting.” Quarterly Journal of Economics, 66: 194-223.

Stinchcombe, A. L.

1990 Information and Organizations. Berkeley, CA: University of California Press.

Strang, D., and S. A. Soule

1998 “Diffusion in organizations and social movements: From hybrid corn to poison pills.” In Annual Review of Sociology, 24: 265-290. Palo Alto, CA: Annual Reviews.

Strang, D., and N. B. Tuma

1993 “Spatial and temporal heterogeneity in diffusion.” American Journal of Sociology, 99: 614-639.

Suchman, M.

1995 “Localism and globalism in institutional analysis: The emergence of contractual norms in venture finance.” In W. R. Scott and S. Christensen (eds.), The Institutional Construction of Organizations: International and Longitudinal Studies: 39-63. Thousand Oaks, CA: Sage.

Taylor, A.

1998 “Layered selection environments: Drivers for internal strategic change in high technology firms.” Working paper, Kellogg Graduate School of Management, Northwestern University.

Tushman, M. L., and P. Anderson

1986 “Technological discontinuities and organizational environments.” Administrative Science Quarterly, 31: 439-465.

Tushman, M. L., and E. Romanelli

1985 “Organizational evolution: A metamorphosis model of convergence and reorientation.” In B. M. Staw and L. L. Cummings (eds.), Research in Organizational Behavior, 7: 171-222. Greenwich, CT: JAI Press.

Utterback, J., and W. Abernathy

1975 “A dynamic model of product and process innovation.” Omega, 33: 639-656.

Van de Ven, A. H.

1986 “Central problems in the management of innovation.” Management Science, 32: 639-656.

VanLehn, K.

1993 “Problem solving and cognitive skill acquisition.” In M. E. Posner (ed.), Foundations of Cognitive Science: 527-579. Cambridge, MA: MIT Press.

von Hippel, E.

1988 The Sources of Innovation. New York: Oxford University Press.

Walker, J. L.

1969 “The diffusion of innovation among the American states.” American Political Science Review, 63: 880-899.

Walsh, J. P., and G. R. Ungson

1991 “Organizational memory.” Academy of Management Review, 16: 57-91.

Weick, K. E.

1990 “Cognitive processes in organizations.” In L. L. Cummings and B. M. Staw (eds.), Information and Cognition in Organizations: 287-320. Greenwich, CT: JAI Press.

1995 Sensemaking in Organizations. Thousand Oaks, CA: Sage.

Westphal, J. D., and E. J. Zajac

1994 “Substance and symbolism in CEOs’ long-term incentive plans.” Administrative Science Quarterly, 39: 367-390.

White, H. C.

1992 Identity and Control: A Structural Theory of Social Action. Princeton, NJ: Princeton University Press.

Descriptive Statistics and

Correlation Coefficients

Variable Mean S.D. 1 2 3

1. All changes 1.08 1.37

2. Innovative changes 0.07 0.29 .364

3. Format changes 0.70 1.03 .861 .157

4. Satellite entry 0.20 0.48 .475 .051 .112

5. Production changes 0.10 0.35 .426 .067 .147

6. Percent male 48.51 0.95 .120 .065 .109

7. Percent white 82.59 11.34 -.159 -.098 -.145

8. Percent 20-34 years 25.31 2.57 .127 .085 .107

9. Percent 35-49 years 20.79 1.32 .123 .067 .114

10. Percent 50 and over 25.07 4.34 -.148 -.081 -.129

11. Advertising revenue growth 1.06 0.30 -.020 -.015 -.000

12. Semi-variance of market shares 0.84 0.43 -.072 -.039 -.067

13. Branch stations 5.90 5.40 .321 .187 .258

14. New stations previous year 0.37 0.72 .154 .046 .107

15. Non-innovative changes previous year 3.77 3.26 .389 .163 .296

16. Innovations previous year 0.26 0.60 .222 .123 .168

17. Innovations weighted by market share 0.60 2.06 .080 .057 .049

18. Innovations weighted by corporate size 0.90 3.57 .148 .088 .113

19. Innovations, accumulated 134.7 117.3 .197 .079 .075

20. Innovations previous year nationally 48.54 24.62 .179 .082 .063

21. Innovations previous year regionally 8.92 8.12 .162 .103 .064

22. Innovations in top-10 markets 6.93 3.02 .113 .074 .037

23. Number of stations 32.37 13.73 .342 .165 .308

Variable 4

1. All changes

2. Innovative changes

3. Format changes

4. Satellite entry

5. Production changes .120

6. Percent male .047

7. Percent white -.064

8. Percent 20-34 years .062

9. Percent 35-49 years .026

10. Percent 50 and over -.073

11. Advertising revenue growth -.020

12. Semi-variance of market shares .000

13. Branch stations .087

14. New stations previous year .123

15. Non-innovative changes previous year .202

16. Innovations previous year .096

17. Innovations weighted by market share .043

18. Innovations weighted by corporate size .050

19. Innovations, accumulated .185

20. Innovations previous year nationally .178

21. Innovations previous year regionally .158

22. Innovations in top-10 markets .122

23. Number of stations .100

Variable 5 6 7 8 9

5. Production changes

6. Percent male .029

7. Percent white -.026 -.284

8. Percent 20-34 years .027 .586 -.377

9. Percent 35-49 years .053 .145 -.014 .258

10. Percent 50 and over -.032 -.600 .422 -.793 -.336

11. Advertising revenue growth -.036 .015 .041 .052 .075

12. Semi-variance of market shares -.050 -.150 .078 -.053 -.022

13. Branch stations .189 .094 -.196 .156 .187

14. New stations previous year .082 .141 -.109 .080 .061

15. Non-innovative changes previous year .238 .172 -.218 .172 .200

16. Innovations previous year .140 .120 -.189 .149 .113

17. Innovations weighted by market share .060 .050 -.087 .091 .001

18. Innovations weighted by corporate size .104 .098 -.113 .104 .138

19. Innovations, accumulated .232 .001 -.002 .003 -.001

20. Innovations previous year nationally .200 -.001 .000 -.003 -.011

21. Innovations previous year regionally .145 -.131 -.156 .007 -.062

22. Innovations in top-10 markets .107 .003 -.003 .001 -.006

23. Number of stations .158 .065 -.262 .124 .209

Variable 10

5. Production changes

6. Percent male

7. Percent white

8. Percent 20-34 years

9. Percent 35-49 years

10. Percent 50 and over

11. Advertising revenue growth -.044

12. Semi-variance of market shares -.048

13. Branch stations -.071

14. New stations previous year -.125

15. Non-innovative changes previous year -.212

16. Innovations previous year -.148

17. Innovations weighted by market share -.076

18. Innovations weighted by corporate size -.084

19. Innovations, accumulated .004

20. Innovations previous year nationally .001

21. Innovations previous year regionally .059

22. Innovations in top-10 markets .003

23. Number of stations -.066

Variable 11 12 13 14 15 16

10. Percent 50 and over

11. Advertising revenue growth

12. Semi-variance of market shares -.018

13. Branch stations .034 -.322

14. New stations previous year -.029 .047 .060

15. Non-innovative changes previous year -.019 -.107 .449 .383

16. Innovations previous year -.026 -.087 .340 .154 .347

17. Innovations weighted by market share -.012 -.016 .167 .057 .138 .607

18. Innovations weighted by corporate size -.015 -.136 .365 .044 .247 .611

19. Innovations, accumulated -.151 -.018 .200 .129 .278 .164

20. Innovations previous year nationally -.114 .016 .190 .127 .247 .189

21. Innovations previous year regionally -.069 .143 .101 .086 .222 .198

22. Innovations in top-10 markets -.093 .030 .132 .096 .135 .124

23. Number of stations .060 -.311 .700 .074 .497 .281

Variable 17 18 19 20 21 22

17. Innovations weighted by market share

18. Innovations weighted by corporate size .445

19. Innovations, accumulated .063 .086

20. Innovations previous year nationally .153 .128 .776

21. Innovations previous year regionally .144 .076 .525 .607

22. Innovations in top-10 markets .086 .087 .416 .578 .328

23. Number of stations .087 .262 .018 .008 .043 .056

Negative Multinomial Models of All

Changes per Market [*]

Model

Independent variable 1 2 3

Intercept -4.122 -4.197 -5.938

(1.496) (1.730) (2.351)

Local market controls

Percent male 0.002 0.002 -0.002

(0.460) (0.484) (0.472)

Percent white 0.050 0.051 0.075

(1.001) (1.009) (1.468)

Percent 20-34 years -0.028 -0.028 -0.023

(1.089) (1.108) (0.946)

Percent 35-49 years -0.021 -0.020 0.022

(0.672) (1.041) (0.723)

Percent 50 and over -0.041 [oo] -0.041 [ooo] -0.035 [oo]

(2.521) (2.940) (2.258)

Advertising revenue growth -0.195 [oo] -0.189 [oo] -0.052

(2.310) (2.296) (1.043)

Semi-variance market shares 0.203 [ooo] 0.203 [ooo] 0.103

(3.051) (3.109) (1.564)

Branch stations 0.042 [ooo] 0.043 [ooo] 0.002

(6.405) (5.564) (0.313)

New stations previous year 0.036 [o] 0.033 [o] 0.034 [o]

(1.828) (1.682) (1.692)

Non-innovative changes 0.015 [ooo] 0.012 [oo] -0.007

previous year (2.865) (2.097) (1.205)

Local market innovations

Innovations previous year 0.095 [ooo]

(3.320)

Innovations weighted by -0.0097 [oo]

size of corporation (2.398)

Innovations weighted by -0.0099

market share (1.166)

Nonlocal innovations

Innovations, accumulated 0.0012 [ooo]

(6.964)

Innovations previous year 0.0025 [oo]

nationally (2.383)

Innovations previous year 0.0080 [ooo]

regionally (2.888)

Innovations in top-10 0.0125 [oo]

markets (2.273)

[varphi] 5.702 [ooo] 5.686 [ooo] 5.939 [ooo]

(5.896) (5.983) (6.413)

Log likelihood -4085.168 -4080.229 -3974.089

[[chi].sup.2] test against 1 20.97 [ooo] 201.58 [ooo]

Independent variable 4 Hypothesis

Intercept -6.433

(2.290)

Local market controls

Percent male -0.002

(0.441)

Percent white 0.074

(1.448)

Percent 20-34 years -0.023

(0.923)

Percent 35-49 years 0.023

(0.746)

Percent 50 and over -0.035 [oo]

(2.179)

Advertising revenue growth -0.051

(1.010)

Semi-variance market shares 0.105

(1.569)

Branch stations 0.002

(0.405)

New stations previous year 0.031

(1.552)

Non-innovative changes -0.010

previous year (1.626)

Local market innovations

Innovations previous year 0.068 [oo] H4: +

(2.363)

Innovations weighted by -0.0078 [o] H5: –

size of corporation (1.916)

Innovations weighted by -0.0108 H6: –

market share (1.251)

Nonlocal innovations

Innovations, accumulated 0.0012 [ooo]

(6.589)

Innovations previous year 0.0029 [ooo] H1: +

nationally (2.712)

Innovations previous year 0.0075 [ooo] H2: +

regionally (2.663)

Innovations in top-10 0.0117 [oo] H3: +

markets (2.109)

[varphi] 5.865 [ooo]

(6.308)

Log likelihood -3970.168

[[chi].sup.2] test against 1 212.14 [ooo]

(o.)p [less than] .10

(oo.)p [less than] .05

(ooo.)p [less than] .01

(*.)Dependent variables are measured at 3-month intervals; independent variables are measured during the 1-year interval preceding the observation period. The data have 157 markets and quarterly observations for 8 years for a total of 4,970 observations. The number of stations per market is not entered as a covariate but is entered as the exposure factor for each observation (Guo 1996).

t-ratios are in parentheses below coefficient estimates.

Negative Multinomial Models of Format

Changes per Market [*]

Dependent variable

All Innovative Format

Independent variable changes changes changes

Intercept -6.433 -10.230 -7.198

(2.290) (3.276) (3.111)

Local market controls

Percent male -0.002 -0.004 -0.001

(0.441) (0.661) (0.376)

Percent White 0.074 0.099 [o] 0.080 [o]

(1.448) (1.718) (1.847)

Percent 20-34 years -0.023 0.004 -0.013

(0.923) (0.106) (0.576)

Percent 35-49 years 0.023 0.038 0.001

(0.746) (0.812) (0.027)

Percent 50 and over -0.035 [oo] -0.053 [oo] -0.016

(2.179) (2.042) (1.075)

Advertising revenue growth -0.051 -1.346 -0.130

(1.010) (1.441) (1.245)

Semi-variance market shares 0.105 0.113 0.176 [oo]

(1.569) (0.779) (2.446)

Branch stations 0.002 0.038 [ooo] -0.001

(0.405) (3.394) (0.126)

New stations previous year 0.031 0.022 0.055 [o]

(1.552) (0.300) (1.954)

Non-innovative changes -0.010 0.022 -0.059

previous year (1.626) (1.218) (0.675)

Local market innovations

Innovations previous year 0.068 [o] 0.060 0.111 [ooo]

(2.363) (0.617) (2.623)

Innovations weighted by -0.0078 [o] -0.0137 -0.0149 [oo]

size of corporation (1.916) (1.034) (2.241)

Innovations weighted by -0.0108 0.0066 -0.0111

market share (1.251) (0.229) (0.920)

Nonlocal innovations

Innovations, accumulated 0.0012 [ooo] -0.0003 0.0003

(6.589) (0.372) (1.177)

Innovations previous year, 0.0029 [ooo] 0.0003 -0.0014

nationally (2.712) (0.086) (0.862)

Innovations previous year, 0.0075 [ooo] 0.0352 [ooo] 0.0052

regionally (2.663) (4.148) (1.362)

Innovations in top-10 0.0117 [oo] 0.0652 [ooo] -0.0016

markets (2.109) (3.043) (0.202)

[varphi] 5.865 [ooo] 20.626 10.265 [ooo]

(6.308) (0.698) (4.331)

Satellite Production

Independent variable entry changes

Intercept -6.790 -9.077

(1.530) (2.018)

Local market controls

Percent male 0.002 0.018 [ooo]

(0.576) (2.653)

Percent White 0.029 0.058

(0.323) (0.707)

Percent 20-34 years 0.001 -0.055

(0.280) (1.381)

Percent 35-49 years -0.030 0.016

(0.884) (0.336)

Percent 50 and over -0.031 [o] -0.039

(1.779) (1.589)

Advertising revenue growth 0.030 -1.320

(0.254) (1.519)

Semi-variance market shares 0.137 -0.106

(1.266) (0.685)

Branch stations -0.029 [ooo] 0.017

(3.327) (1.572)

New stations previous year 0.104 [ooo] 0.037

(2.631) (0.637)

Non-innovative changes 0.028 [oo] 0.049 [ooo]

previous year (2.290) (3.109)

Local market innovations

Innovations previous year 0.008 0.125

(0.117) (1.549)

Innovations weighted by -0.0016 -0.0096

size of corporation (0.172) (0.903)

Innovations weighted by 0.0018 -0.0010

market share (0.099) (0.407)

Nonlocal innovations

Innovations, accumulated 0.0019 [ooo] 0.0033 [ooo]

(4.933) (5.512)

Innovations previous year, 0.0073 [ooo] 0.0152 [ooo]

nationally (3.209) (4.901)

Innovations previous year, 0.0105 [o] 0.0136 [o]

regionally (1.883) (1.884)

Innovations in top-10 0.0047 [ooo] 0.0532 [ooo]

markets (3.853) (3.003)

[varphi] 10.222 [ooo] 7.597 [oo]

(2.841) (2.343)

[varphi] is the inverse of the variance of the gamma-distributed market error term.

(o.)p [less than]. 10

(oo.)p [less than]. 05

(ooo.)p [less than]. 01; t-ratios are in parentheses below coefficient estimates.

(*.)Dependent variables are measured at 3-month intervals; independent variables are measured during the 1-year interval preceding the observation period. The data have 157 markets and quarterly observations for 8 years for a total of 4,970 observations. The number of stations per market is not entered as a covariate, but is entered as the exposure factor for each observation (Guo 1996).

[varphi] is the inverse of the variance of the gamma-distributed market error term.

COPYRIGHT 2000 Cornell University, Johnson Graduate School

COPYRIGHT 2000 Gale Group