Abstract
This article focuses on normatively uncertain innovations and asks when and how organizational status affects the adoption and implementation of these innovations. I argue that middle-status organizations perceive normative uncertainty more as an opportunity for gain, whereas high- and low-status organizations perceive it more as a threat of loss and competition. Consequently, middle-status organizations are likely to be the first to adopt normatively uncertain innovations and to implement them in a way that emphasizes normative uncertainty. Event history and negative binomial regression analyses of the US newspaper industry from 1993 to 2007, the period after the emergence of digital media that was normatively uncertain, support my theoretical arguments. Middle-status newspapers launched their websites faster than other newspapers did. In addition, they emphasized interactivity on the web, which could potentially change the identity of newspapers from sole news producers to facilitators.
Keywords
The diffusion of an innovation or a new practice depends, among many other factors (see Rogers, 2003 for review), on the social status of the incumbent firms potentially adopting the innovation and the normative compatibility of the innovation itself (e.g. Rogers, 2003; Strang and Meyer, 1993). The status of a firm refers to its position in a (hierarchical) market (Jensen et al., 2011; Podolny, 2005), whereas the normative compatibility of an innovation is defined as “the degree to which [it] is perceived as being consistent with the existing values and norms” (Rogers, 2003: 15). When innovations are normatively compatible or normative, such as accounting firms’ offering of consulting services, high-status firms are known to adopt these innovations first (e.g. Davis and Greve, 1997; Greenwood and Suddaby, 2006). Low-status firms are, in contrast, likely to be the first to adopt normatively incompatible or counter-normative innovations, such as hostile corporate takeovers (e.g. Hirsch, 1986; Leblebici et al., 1991). Regardless of normative compatibility, middle-status firms are often viewed as those least likely to adopt innovations first (e.g. Menzel, 1960; Phillips and Zuckerman, 2001). A widespread assumption in this line of research is that potential adopting firms clearly understand the normative compatibility of an innovation and respond differently to it depending on their status. For some innovations, however, a shared understanding of whether the innovations are normatively compatible has not been fully established, raising the important strategy question of when and how firms respond to normatively uncertain innovations.
Normative uncertainty means that the normative compatibility of an innovation is open to more than one interpretation, so consensus regarding its normative compatibility has not yet been reached among audiences (i.e. those who are interested in the innovation), including potential adopting firms and customers (cf. Benders and Van Veen, 2001). A multifaceted innovation that contains aspects of diverse normative compatibility is an example of such a normatively uncertain innovation. Potential adopting firms thus have some flexibility in how they interpret the normative compatibility and how they strategically respond to the innovations (Ansari et al., 2010; Westphal et al., 1997). I examine how the status of incumbent firms constrains or facilitates their responses to these innovations. Organizational status often affects how firms perceive and interpret innovations, because different status positions encompass different sets of market identities that control and coordinate how organizations should behave (Jensen et al., 2011; Podolny, 2005). The status of adopting firms, therefore, is critical for understanding the two-stage process of when firms adopt normatively uncertain innovations and how they differently implement the innovations (Chandler, 2014; Okhmatovskiy and David, 2012; Park et al., 2011).
In this article, I focus on middle-status firms and examine how their responses to normatively uncertain innovations may differ from those of high- and low-status firms. Middle-status firms tend not to be the first to adopt innovations, because they often have fewer resources than high-status firms and face stronger conformity pressures than low-status firms (e.g. Menzel, 1960; Phillips and Zuckerman, 2001). The strong constraints on innovation faced by middle-status companies, however, conflict with their economic and/or social desire to improve their performance by being the first to adopt innovations (cf. Kennedy and Fiss, 2009). I argue that normative uncertainty relaxes the often-observed innovation constraints on middle-status firms, and therefore, middle-status firms may attempt to change their current disadvantageous position by being the first to adopt this type of innovation. Moreover, middle-status firms emphasize the normatively uncertain aspect of the innovations in implementation to better free themselves from existing innovation constraints. High- and low-status firms, in contrast, are likely to highlight the normatively compatible and incompatible aspects, respectively, to exploit the innovations in ways that are most viable and beneficial to them.
The empirical context of this study is US daily newspapers’ reactions to digital media, the electronic version of news publishing on the web (Gilbert, 2005, 2006). The normative compatibility of digital media was not easy to determine when digital media first emerged (Black, 1994; Singer, 2003). Some experts believed that by enabling newspapers to provide multisensory news to readers, digital media could enhance professional journalism, the core of existing newspaper norms that focuses on the process of reporting accurate, objective, relevant, and fair news (Pavlik, 2008; Schudson, 1978; South, 1999). Others were excited by an opportunity for civil journalism created by digital media’s interactive features (Jenkins, 2006). Still, others questioned the future of professional journalism because they believed that digital media blurred the wall between editorial and advertisement, which represents a violation of another important foundation of professional journalism: maintaining editorial independence from commercial pressures (France, 1999; Glaberson, 1994). The normative uncertainty of digital media caused variations with respect to when and how newspapers of different status levels responded to digital media. Thus, this setting is appropriate for testing my hypotheses about how status affects the adoption and implementation of normatively uncertain innovations.
Status and normatively uncertain innovations
The status of a firm not only functions as a signal of product quality (Kim, 2012; Podolny, 1994) but also provides a social or market identity that defines a broader set of normative expectations for the firm (Jensen et al., 2011; Phillips et al., 2013). 1 Market norms refer to “what the members [of a given market] … should do, ought to do, are expected to do, under given circumstances” (Homans, 1950: 123). Importantly, market norms tend to be established around the behaviors of high-status firms, given the visibility and representativeness of their position for the market (Merton, 1968; Murphy, 2002). 2 High-status law firms, for example, are expected to provide corporate clients with legal advice in corporate, taxation, and securities law and to avoid the domains of low-status law firms, such as family and personal injury law, which are consistent with a normative understanding of law firms (Phillips and Zuckerman, 2001). Since organizational status defines normative expectations, it is not surprising that status affects the diffusion of innovations. Most research show that depending on the normative compatibility of innovations, high- or low-status firms tend to be the first to adopt them (Menzel, 1960; Phillips and Zuckerman, 2001; Rogers, 2003). High-status firms are more likely to adopt normatively compatible innovations that maintain and enhance the current norms, whereas low-status firms tend to be the first to adopt economically attractive, but normatively incompatible, innovations that often directly challenge and undermine the current market norms.
High-status firms tend to be the first to embrace normatively compatible innovations, as these innovations help them maintain their central roles in the industry (Davis and Greve, 1997; Greenwood and Suddaby, 2006; see also Kim, 2012). High-status firms possess positional advantages—such as the best access to information—to most quickly and effectively adopt innovations if necessary (Jensen et al., 2011; Podolny, 2005; Shipilov, 2005). Elite accounting firms, for example, were among the first to offer management advisory services to their auditing clients, which “was justified in normative terms as being consistent with providing a high level of service to customers” (Dacin et al., 2002: 48). For high-status firms, in other words, it is important that innovations be compatible with the existing values and norms that define the core identity of their high-status position. New technologies in the newspaper industry, such as telegraphy and photography, for example, did not change the norms of professional journalism, although they revolutionized the content and appearance of newspapers (Fenton, 2010; Pavlik, 2008). 3 The introduction of these technologies reinforced the status hierarchy by allowing early adopting high-status newspapers to increase their quality (the core of professional journalism) relative to lower-status newspapers, which had relatively fewer resources and incentives that were needed to invest in these new norm-enhancing technologies.
High-status firms are less likely to adopt normatively incompatible innovations because these innovations can potentially undermine their advantageous position (Greenwood and Suddaby, 2006; Phillips et al., 2013; Podolny, 2005). 4 Low-status firms, however, are generally least committed to the existing market norms, and they have the least to lose from adopting even normatively incompatible innovations. To the extent that they perceive economic benefits, low-status firms often actively adopt normatively incompatible innovations (Hirsch, 1986; Kraatz and Zajac, 1996; Leblebici et al., 1991). Importantly, normative incompatibility also provides assurance to low-status organizations that members of other status groups will not actively respond to the innovations because higher-status organizations tend to care more about normative or social pressures (Phillips et al., 2013; Podolny, 2005). This expectation of little or no competition further increases the interest of low-status organizations in adopting normatively incompatible innovations first. In the US radio industry, in the 1920s, for example, small low-status local radio stations first initiated the practice of directly advertising products and services on the air waves, which was not fully accepted at the time (Leblebici et al., 1991). Within the UK National Health Service, low-status primary care providers and non-doctors also initiated projects that diverged from the norms of the field, such as primary-care-provider-based stroke rehabilitation services (Battilana, 2011).
While high-status firms are less likely than low-status firms to adopt normatively incompatible innovations, middle-status organizations are even less likely than high-status firms to adopt such innovations; this phenomenon is known as middle-status conformity (Menzel, 1960; Phillips and Zuckerman, 2001). In other words, middle-status firms experience the strongest conformity pressures to adhere to “common standards and shared understandings” of how a player should behave in a given market, that is, the current market norms (Phillips and Zuckerman, 2001: 384). High-status firms generally enjoy secure membership in the market, as they represent their market categories based on the visibility of their structural position (Merton, 1968; Murphy, 2002; Phillips et al., 2013). As mentioned, low-status firms tend not to worry about maintaining the status quo and have the least to lose from not following existing conformity pressures. High- and low-status firms, thus, are relatively free from these conformity pressures, whereas middle-status firms often face the strongest pressures to claim and maintain their membership in the market. Indeed, middle-status conformity has been observed in numerous settings, including the markets for investment advice and legal services (Phillips and Zuckerman, 2001) and the Hollywood film industry (Perretti and Negro, 2006).
Normative uncertainty and the effect of status on adoption
The adoption of normatively uncertain innovations, however, cannot be fully explained by the existing literature based on the normative compatibility criterion. I argue that normatively uncertain innovations create unique opportunities, especially for middle-status firms. As mentioned, normative uncertainty indicates a lack of consensus among audiences regarding the normative compatibility of the innovations. This uncertainty stems either from the coexistence of the normatively various elements of an innovation or from uncertainty about whether a certain element of the innovation is normatively compatible (see also Benders and Van Veen, 2001).
When digital media were first introduced, for example, disagreement existed regarding whether it would bolster or challenge the prevailing norms of professional journalism (e.g. Black, 1994; Lynch, 1998; Singer, 2003). Some argued that digital media were normatively compatible because it would help newspapers to deepen readers’ understanding of traditional stories by providing more accurate and unbiased information (Pavlik, 2008; South, 1999). Others thought that digital media were normatively incompatible because excessive information might dilute the overall quality of news on the web (Jones, 2009) or it would become more difficult to maintain editorial independence on the web (France, 1999). It also had normatively uncertain interactivity, which could help newspapers “link with the community”; at the same time, however, interactivity could distract readers from “the main thread of the article” (South, 1999). Similarly, in the early 20th century, American opera companies believed that American operas would enhance the identity of the United States (Saunders, 1932), which was not directly related to opera but had a positive connotation at the society level (normative uncertainty). They were also concerned about presenting American operas because the boundary between American operas and musicals was not always clear, and presenting musicals (a lowbrow art form) was regarded as normatively incompatible for opera companies (highbrow art organizations) (Levine, 1988).
Importantly, normative uncertain innovations imply that conformity pressures with respect to the innovations have not yet been fully established. The basic assumption of middle-status conformity, which mainly prevents middle-status firms from adopting even innovations that are not the primary interest of high-status firms, is that audiences have a clear normative understanding of their identities and activities (see Phillips and Zuckerman, 2001). The typical understanding of opera in the United States, for example, stems from Italian opera, which is the classical music equivalent of a Billboard hit in popular music: “[opera] should be tuneful, it should have recognizable trios and quartets … and it should be in a 2,500 seat house—all of the standards set up by Puccini” (Kim and Jensen, 2011; Rourke and Hossack, 2000: 15). Normative uncertainty, however, means that audiences do not agree on whether the current set of norms can be applied to the innovations in question or whether they must either modify the current norm system or create a new one. In other words, it is unclear what to expect from these innovations and from the adopting firms (cf. Ansari et al., 2010; Benders and Van Veen, 2001). Conformity pressures related to normatively uncertain innovations, therefore, have yet to be effectuated, creating an interesting theoretical situation in which the middle-status conformity hypotheses cannot fully explain when and how different-status firms differently respond to normative uncertainty.
Given this normative uncertainty, high- and low-status firms often perceive these innovations as threats and are therefore reluctant to be the first adopters, albeit for different reasons. First, high-status firms may perceive normative uncertainty as a potential threat to their advantageous social position. Since high-status firms often face stronger penalties if their main audiences believe that they are not committed to them (Phillips et al., 2013), they are reluctant to be the first to adopt innovations that could contain normatively incompatible elements. The uncertain nature also means that these innovations could potentially displace the current system that provides high-status firms with their positional advantages (Jensen et al., 2011; Podolny, 2005; Rogers, 2003). Importantly, high-status firms typically have the best access to necessary resources (Jensen et al., 2011; Podolny, 2005). These advantages allow them to wait and observe the behaviors of first movers before responding to normative uncertainty (Conner, 1988; Shipilov, 2005). If the responses by other first movers appear promising, high-status firms can (or believe they can) invest their resources in these innovations to quickly catch up with the first movers. Second, for low-status firms, normative uncertainty indicates that they are not guaranteed to be the only players that have considered adopting these innovations. When low-status firms adopt normatively incompatible innovations, they are the only firms interested in responding to them (Podolny, 2005); thus, their positional disadvantages are relatively inconsequential. The threat of competition with better-positioned organizations, however, is significant for low-status firms when adopting normatively uncertain innovations.
In contrast, I argue that normative uncertainty increases the possibility that middle-status firms actively respond to these innovations. For other firms, middle-status firms may strategically seek innovation opportunities to improve their economic and/or social performance (Kennedy and Fiss, 2009; Rogers, 2003). Current competitive conditions are often not advantageous to middle-status firms, given the nature of their mid-level position within the stratified market. They tend to face competition not only from other middle-status firms but also, to a lesser extent, from high- and low-status firms (e.g. Haveman, 1993). Middle-status firms are in a situation in which they can potentially benefit most from an innovation to the extent that it provides them with an avenue for moving up the status hierarchy or helps them create a new market in which they are the leaders; however, their existing innovation constraints mainly prevent them from actively adopting innovations first. I argue that normative uncertainty provides one of the few unique opportunities for these middle-status firms to strategically change their disadvantageous position. Because of normative uncertainty, high-status firms—which tend to have more resources than middle-status firms to more successfully adopt any innovation—are unlikely to be the first to adopt these innovations. In addition, the conformity pressures related to these innovations—which often prevent them from actively adopting innovations that are not the main interest of high-status firms—are not fully established.
In summary, middle-status firms may perceive normatively uncertain innovations more as an opportunity for gain, whereas high- and low-status firms perceive them more as a threat of loss and competition. Opportunity perceptions often encourage innovation adoption, while threat perceptions promote commitment to the status quo (Gilbert, 2005; Kennedy and Fiss, 2009; Thomas et al., 1993). I thus hypothesize below that middle-status firms with opportunity perceptions adopt normatively uncertain innovations more rapidly than others with threat perceptions, which is different from the traditional middle-status conformity hypothesis that predicts that middle-status firms are least likely to innovate first (cf. Menzel, 1960; Phillips and Zuckerman, 2001):
H1. Middle-status organizations are more likely than high- or low-status organizations to be the first to adopt a normatively uncertain innovation.
Status and implementation of normatively uncertain innovations
The discussion so far has centered on when different-status firms adopt normatively uncertain innovations. It is also important to examine variance in how firms implement the innovations once they have adopted them (Chandler, 2014; Okhmatovskiy and David, 2012; Park et al., 2011). Specifically, I examine how organizational status affects not only the adoption decisions but also the implementation decisions of normatively uncertain innovations. The adaptation literature has emphasized that many innovations invite substantial modifications and, in turn, vary in implementation (Ansari et al., 2010; Bijker et al., 1987; Westphal et al., 1997). When implementing total quality management (TQM) practices, for example, US hospitals could select one of the standard approaches focused on different aspects of TQM programs (Westphal et al., 1997). They could develop their own if none of the standardized approaches suited their needs. Normatively uncertain innovations are no exception in inviting organizational discretion, as normative uncertainty is likely to create a situation “where expectations are vague and ill-defined” (Chandler, 2014: 1726). Depending on their own interpretations and needs, firms can eclectically emphasize different elements of the innovations in implementation (Ansari et al., 2010; Benders and Van Veen, 2001), potentially shaping the shared understanding of how the innovations and adopting firms will be defined (cf. Bijker et al., 1987).
As previously mentioned, middle-status firms tend to perceive normatively uncertain innovations as an opportunity to differentiate themselves without confronting their typical innovation constraints. It is therefore beneficial for middle-status firms to further emphasize the normatively uncertain aspect of the innovations. This “innovative” attempt can potentially help them to establish a new set of normative expectations related to the innovations and to the adopting firms. By focusing on the normatively uncertain aspect, in other words, middle-status firms attempt to create a situation in which they can better free themselves from their current innovation constraints. Of course, they may not be able to completely change the current normative expectations; by implementing them in a normatively uncertain way, however, middle-status firms attempt to, at least partially, alter or add a new dimension to the current set of norms and expectations. They can also be the first to exploit this uncertain and potentially distinctive aspect if such advantages actually exist. In the French gastronomy industry, for example, young but not peripheral chefs “who had to climb up the ladder of stardom” (Rao et al., 2005: 974) were the first to initiate the nouvelle cuisine movement. They created an identity of nouvelle cuisine with normatively uncertain dimensions that often contrasted with those of classical cuisine, which was the dominant cuisine at the time. The creation of the new identity made it possible for these chefs to be relatively free from conformity pressures imposed by the then-existing norms of classical cuisine and helped them climb the status ladder in a distinctive manner.
It is important to distinguish a normatively uncertain implementation from a normatively incompatible one. Market norms are defined “under given circumstances” (Homans, 1950: 123), and normatively incompatible approaches mean that the attempts directly challenge the existing market norms under these given circumstances and are followed by social sanctions (Homans, 1950). Normative uncertainty, however, may change these given circumstances. Normatively uncertain implementations, therefore, are often not clearly defined by the current market norms and are unlikely to be followed by social punishments. Emphasizing positive connotations at the broader society level, such as democracy or efficiency, is an example of implementing an innovation in a normatively uncertain manner, especially if the current market norms do not contain these positive connotations (see also David et al., 2013; Haveman and Rao, 1997). In the context of the newspaper industry, the interactive aspect of digital media does not align well with the traditional notion of newspapers as news creators: “anybody can post anything on the Net” (Fancher, 1995: A2). However, it does not directly detract from newspapers’ pursuit of professional journalism (Schultz, 1999; Zeng and Li, 2006), and it supports the democratic distribution of the news creation process (Neuman et al., 1994). In other words, an interactivity-enriched web page is a normatively uncertain implementation of digital media. My next hypothesis is therefore about the inverted U-shaped relationship between status and normatively uncertain implementation:
H2. Middle-status organizations are likely to implement the normatively uncertain innovation in a normatively uncertain way.
Although my first two hypotheses focus on when and how middle-status firms differently respond to normatively uncertain innovations compared to other-status firms, I further argue that high- and low-status firms may emphasize different aspects of normatively uncertain innovations when implementing them. Specifically, the well-established normative compatibility criterion can be applied to explain how high- and low-status firms implement these innovations differently. First, it is strategically advantageous for high-status firms to emphasize the normatively compatible aspects of the innovations. Their positional advantages help them become the first to utilize or monetize normatively compatible innovations, indicating that they can also enjoy these advantages if the innovations are normatively shaped (Phillips and Owens, 2004). Positional advantages and the persistent nature of status hierarchies also enable high-status firms to wait and watch the behaviors of middle-status firms before actively embracing normatively uncertain aspects (Conner, 1988; Shipilov, 2005). Once they adopt a normatively uncertain innovation, low-status firms should, and will, focus on the innovation’s normatively incompatible aspects, to which other status groups generally (are forced to) devote less attention (Leblebici et al., 1991; Podolny, 1994). With respect to other implementation approaches, low-status firms face greater direct competition pressures, as these attempts receive active reactions from other-status firms, lowering their relative interest in them. High- and low-status firms’ implementation of the innovations in normatively compatible and incompatible ways, respectively, can be further amplified by their threat perceptions, which tend to lead them to stick to the same routines even under new competitive conditions (Gilbert, 2005). High- and low-status firms, therefore, may emphasize different aspects of normatively uncertain innovations in implementation, as proposed below:
H3a(b). High- (low-) status firms are likely to implement the normatively uncertain innovation in a normatively compatible (incompatible) way.
Methods
Daily newspapers in the United States and sample
The empirical setting of this study is the US daily newspaper industry from 1993 to 2007. Newspapers containing news and information began to be published regularly in Europe in the early 17th century and were introduced to the United States soon thereafter (Picard and Brody, 1997). With the introduction of new technologies, such as improvements in presses, newspapers began to reach broad groups of readers, and two types of journalism—“journalism as information” and “journalism as entertainment”—have become prevalent since the late 19th century (Schudson, 1978: 88). Journalism-as-information focuses on reporting fair and objective information and has helped establish the objectivity standard of professional journalism. Journalism-as-entertainment, in contrast, focuses on telling entertaining stories to reach broader readers and advertisers, often based on sensationalism. The difference in the two models’ quality expectations has been pervasive and has created “a set of tiered market segments … ordered by status” (Phillips and Zuckerman, 2001: 389). The New York Times, a representative of journalism-as-information, has been “a badge of respectability,” whereas journalism-as-entertainment has been associated with low-status newspapers or tabloid journalism in extreme cases (Jones, 2009; Schudson, 1978: 117). The persistent differences between the two types of journalism have thus created a stable status hierarchy in the newspaper industry.
The market stratification and the status hierarchy have not changed substantially since the 19th century, although there have been several technological innovations in the newspaper industry, such as the advent of telegraphy and photography (Fenton, 2010; Pavlik, 2008). The introduction of digital media, however, was believed to have had profound effects on the newspaper industry (Boczkowski, 2005; Gilbert, 2005, 2006). Digital media were introduced with the World Wide Web in 1993 and rapidly diffused to most US newspapers (Boczkowski, 2005). The News and Observer in North Carolina and The San Jose Mercury News in California were among the first major daily newspapers to launch their own websites, and other elite newspapers, such as The New York Times and The Wall Street Journal, soon followed. In relation to this study, digital media were normatively uncertain when it first emerged. On one hand, digital media were praised as the future of journalism because they could enhance the quality of journalism by providing readers with more objective information (Pavlik, 2008; Peng et al., 1999; South, 1999). They also offered an unprecedented opportunity for civil journalism by encouraging readers to become active participants in news creation (Jenkins, 2006; Neuman et al., 1994). Some experts, however, were concerned that editorial independence could be more easily challenged, as the clear wall between content and commerce was more easily blurred on the web (France, 1999; Gaw, 1997; Schudson, 1978). When commenting on The San Jose Mercury News’ electronic experimentations, for example, The New York Times worried that “the transition to electronic journalism is bringing with it some challenges to old rules [of journalism]. [The Mercury News’ electronic experiment], for instance, includes news and business employees, breaching the traditional wall that separates the two at most newspapers” (Glaberson, 1994: D6; see also Appendix 1).
With this normative uncertainty of digital media, US newspapers developed their web pages in various ways, once they had adopted digital media (Boczkowski, 2004, 2005). As with print media, digital media are an important platform for newspapers to publish their news articles and create advertisement space; they encompass pre-existing elements of print newspapers, such as the scope of news content and an advertisement-focused business model (Picard and Brody, 1997). Digital media also contain several unique elements, including interactivity and multimedia capability (Boczkowski, 2005; Peng et al., 1999). By focusing on either new or pre-existing elements of the digital media, print newspapers have developed their own websites, which has led to variance in implementation.
Specifically, three elements of digital media—interactivity, multimedia, and advertising practices—are important for this study. Newspapers were known to emphasize these elements to different extents on their websites, leading to substantial empirical variations among newspapers (Boczkowski, 2004). More importantly, these three elements capture different normative aspects of digital media (see also Appendix 1, which contains the contemporaneous quotes for the three elements of digital media). First, (interpersonal) interactivity represents the most normatively uncertain aspect of digital media among the three elements. Interactivity refers to bi- or multi-directional communication among sources and recipients, and newspapers and newspaper readers have communicated with one another much more easily since the emergence of digital media (Schultz, 1999; Zeng and Li, 2006). 5 Since their inception, newspapers were producers and readers were consumers; indeed, a newspaper “institutes a structured break between the production of symbolic forms and their reception. The capacity of recipients to intervene in or contribute to the process of production is strictly circumscribed” (Thompson, 1995: 29). Interactivity, however, allows previously passive recipients to become active news creators, and it thus potentially changes the identity of newspapers from sole news producers to facilitators. It strongly supports the democratic distribution and dissemination of news by anyone who is interested in news (Jenkins, 2006; Neuman et al., 1994). From the normative perspective, however, interactivity is not well defined by professional journalism, which focuses on the objectivity and independence of the newspapers’ news creation process (Jones, 2009).
Second, multimedia capability represents the most normatively compatible element of digital media among the three. Developing multimedia-enriched websites, such as publishing news articles with video and/or audio “as an adjunct to a story” (Layton, 2007: 26), was praised for enhancing stories with more descriptive scenes and providing more accurate and unbiased information, thereby supporting the existing norms of professional journalism (Pavlik, 2008; South, 1999). Of course, reporters needed to retrain themselves to publish multimedia-enriched articles, which could change how newspapers delivered news. In contrast to interactivity, multimedia capability does not change what newspapers do: newspapers are still sole news producers, and readers are passive recipients. The multimedia capability of digital media, in other words, can be applied by newspapers in normative terms to support what newspapers should do: professional journalism.
Finally, placing excessive emphasis on advertising has been regarded as financially tempting (or necessary) but normatively incompatible with professional journalism, which emphasizes the editorial independence of newspapers as important watchdogs of governments and/or corporations (Schudson, 1978). Corporations and governments are the main financial resource providers of newspapers, which can potentially hinder editorial independence. To maintain editorial independence, print newspapers have structurally separated the newsroom and advertisement departments, which is often called “the Great Wall between content and commerce” (France, 1999: 122). News reporters and sales employees, for example, are not supposed to work together (Glaberson, 1994). Print newspapers also “do not run display advertising on front pages, citing the need to mark the independence of articles from the newspapers’ advertisers” (Gaw, 1997: 1D). Advertising is regarded as a necessary evil, at least for higher-status newspapers, and these normative expectations have not changed on the web (Arant and Anderson, 2001). The technical ease of blurring the wall between content and commerce, such as mingling articles with commercial links, and a lack of committed resources on the web, however, have forced newspapers to directly face an ethical dilemma, creating substantial variance in how newspapers handle advertisements on the web (Lynch, 1998).
The US newspaper industry is therefore an appropriate empirical setting in which to study the hypotheses. I consider a newspaper as having adopted digital media, a normatively uncertain innovation, once it launched a website. The first hypothesis, which states that middle-status firms are the first to adopt a normatively uncertain innovation means that there is a U-shaped relationship between newspaper status and their website launch (i.e. adoption time). Based on the normative uncertainty of interactivity, having an interactive website means emphasizing the normatively uncertain aspect of digital media. Having a multimedia-enhanced website indicates highlighting the normatively compatible aspect, while including more advertisements on the web’s front pages refers to implementing it counter-normatively. I, therefore, hypothesize that there is an inverted U-shaped relationship between status and interactivity. In addition, there is a positive (or negative) relationship between status and the development of a multimedia- (or ad-) enriched website.
The sample includes 207 large daily print newspapers in the United States. I exclude online-only newspapers because I am interested in the behaviors of incumbents. I also focus on large daily newspapers because small newspapers with circulations below 25,000 often have different journalistic objectives, business models, newsroom practices, and customers than large newspapers (Picard and Brody, 1997). To obtain a sample of large daily US newspapers that have similar newsroom practices and have been similarly affected by the emergence of digital media (Meyer, 2009), I sampled the largest 100 newspapers by circulation (including the top 100 by daily circulation, the top 100 by Sunday circulation, and the largest newspapers in each state) from 1993 to 2006, which resulted in a group of 131 large newspapers (the top-100 sample). Studies also show that newspapers that have received or have been nominated for the Pulitzer Prize, the most prestigious award in journalism, often employ similar reporting practices and have similar journalistic ambitions, and they differ from non-awardees in these aspects (Bogart, 2004; Hansen, 1990; Harris, 2007). To capture medium-sized, but significant, newspapers, I added newspapers that had ever been awarded or nominated for the Pulitzer Prize to the top-100 sample, which resulted in a basic sample of 207 newspapers.
The first adoption hypothesis focused on the period between 1994, the year when the first newspapers in the sample launched their websites, and 2003, the year when the last newspapers in the sample launched their websites. However, it still required several years for digital media to become a significant player in the market. In the very early stages, newspapers often used websites to provide simple contact information (Li, 2006). They began to actively use their websites around the year 2000, and the forms of online news publishing significantly varied (Boczkowski, 2004). As of 2007, the newspaper industry had not yet fully formed an expectation of how newspapers on the web should look (Li, 2006; Pavlik, 2008), which indicates that conformity pressures in the new market had not yet been put into effect. I thus focused on the period between 2001 and 2007 to examine the next implementation hypotheses. 6 The Editor and Publisher International Yearbook and the Internet Archive were the primary data sources for this study. The Yearbook, an annual US daily newspaper directory, offers print circulation data and organizational information. It also provides the website address of a newspaper if it has one, enabling access to the website launch date on an annual basis. I obtained the website launch year for 201 newspapers. I also examined the websites of the sampled newspapers from 2001 to 2007 using the Internet Archive, a website containing archived US daily newspaper web pages (http://www.archive.org), and I collected the front pages of these newspapers per quarter (four front pages per year, with some missing years). The final sample for the second hypotheses included 1195 (firm-year) observations for 177 newspapers.
Dependent and independent variables
My dependent variables were the year in which a newspaper launched its website and Interactivity, Multimedia Capability, and Ad Space. First, the year in which a newspaper company launched its website was coded as “1,” and the variable was coded as “0” for each year until such an adoption had been made. Based on the event history analysis model using a single event, newspapers that launched their websites in a given year were dropped from the analysis in subsequent years. Second, Interactivity and Multimedia Capability were measured by applying a coding schema to each newspaper’s front web page (see Appendix 2). The coding schema was created based on several journalism studies (Lowrey, 2003; McMillan, 1998; Zeng and Li, 2006). I also emphasized the visibility of items on the front page because the front page is an appropriate unit for measuring a firm’s strategic emphasis on different elements of digital media (Foust, 2009). I awarded one point if a web page contained a component in the coding schema that contributed to interactivity (or multimedia capability). Then, I summed these points to obtain the score of interactivity (or multimedia capability), which had a maximum of nine (or six) points per page. The annual Interactivity (or Multimedia Capability), that is, the sum of four pages, ranged from 0 to 36 (24) points. Finally, I counted the number of separate ad spaces on each front page and totaled the scores of four pages to obtain annual Ad Space. Newspapers in print tended to have very limited advertisement space on their front pages, but advertisements on the web could be more easily mingled with news articles, creating significant variance in the number of separate advertisement spaces on the web (Gaw, 1997). A higher index indicated a greater degree of interactivity (multimedia capability or ad space).
The main independent variable was the cumulative number of the Pulitzer Prizes a newspaper had ever been awarded (Pulitzer [ln]). Since 1917, the Pulitzer Prize in journalism has been awarded annually for achievement in US text-based journalism (Harris, 2007). Winning a Pulitzer Prize is regarded as the highest honor a newspaper can receive; being named a finalist is also highly respected (Bogart, 2004). As of 2007, The New York Times had been awarded the Pulitzer Prize 93 times, followed by The Washington Post and The Los Angeles Times. In contrast, 52 newspapers in the top-100 sample (of 131 newspapers) had never received a Pulitzer Prize and 33 of these newspapers had never been nominated as finalists. 7 Considering the skewed distribution and the difference between awards and nominations, I took the logarithm of the sum of the cumulative number of Pulitzer awards and the half-weighted number of nominations, as in the equation below
I focused on the cumulative number of awards as a status indicator for several reasons. First, winning the Pulitzer Prize is rare for most newspapers (Alexander, 2010). Whether a newspaper has ever been awarded the Pulitzer and how many times it has received one in its history are commonly used by audiences—including newspapers, readers, and advertisers—to establish their quality expectations regarding newspapers. Newspapers indeed emphasize their total number of awards, regardless of the achievement date (Harris, 2007). Second, audiences tend to consider award-winning newspapers differently from non-awardees, creating a subgroup of their own (Hansen, 1990). For example, readers’ confidence in a newspaper often significantly increases if the newspaper receives the Pulitzer (Alexander, 2010). Third, award-winning newspapers are not affected by the defections of reporters from winning articles because entries and awardees are newspapers; winning the Pulitzer is regarded more as a collective achievement by the newspaper than as an individual honor awarded to the journalist (Bogart, 2004; Harris, 2007). As an alternative status measure, I calculated the percentile measure of a newspaper based on Pulitzer [ln]. Pulitzer Rank ranges from zero to one, and a higher number indicates a higher-status newspaper. The New York Times had a value of one for the entire period of study, indicating that it always occupied the highest position. The high correlation between the two independent variables demonstrates the persistence of the status hierarchy (Table 1). It also justifies why people tend to focus on a simpler measure of the cumulative number instead of the rank measure (cf. Jensen et al., 2011).
Summary statistics and bivariate correlations.
STD: standard deviation; HS: high-status.
Control variables
Adoption hypothesis
I used a number of control variables for the first hypothesis. First, at the online news production level, I controlled for whether a newspaper used delivery methods other than print or online. Certain newspapers chose to have a preliminary text-based electronic presence with a major national online service, such as America Online. I coded Online Service as “1” if the company provided an electronic service using an online service provider. I also coded Audiotex to indicate whether a newspaper used a telephone voice-information system. At the newspaper level, newspaper size or resource availability affects the newspaper’s operations, including technology adoption and implementation (Schultz, 1999; Zeng and Li, 2006). Although digital media are characterized by low setup and maintenance costs (Peng et al., 1999), smaller newspapers could have insufficient resources to launch their own websites. I measured newspaper size using the natural logarithm of each newspaper’s average circulation. Since newspapers can be published on weekdays, Saturday and/or Sunday, or some combination of these times, I calculated daily Average Circulation [ln] to measure its size (Meyer, 2009)
I also controlled for firm performance, Average Circulation Change, as financially constrained newspapers may be either more reluctant or more willing than other newspapers to adopt a new technology. The variable was defined by the change in average circulation divided by the previous-year average circulation. 8 I also included the Age [ln] of a newspaper to control for any age effect on adoption. Subscribing to news services, especially supplemental ones offered by The New York Times or The Los Angeles Times/Washington Post, can also provide a newspaper with information and/or affiliation benefits (Kingsley, 1992), which can affect its decision regarding when and how to adopt digital media. Therefore, I controlled for two news service–related variables (High-Status News Service Subscription and Number Change in News Services). The regional scope of coverage is another factor that can influence adoption behavior (Zeng and Li, 2006); thus, I created National Newspaper for national distribution. According to the Yearbook’s categorization, there are four national newspapers in the sample: Christian Science Monitor, Investor’s Business Daily, USA Today, and The Wall Street Journal. I also controlled for two Spanish-language newspapers: La Opinión and El Nuevo Herald (Spanish Newspaper).
Finally, the broader environment in which a newspaper is located can affect its adoption decisions. First, at the parent-company level, media ownership can affect its activities, including website launches and development policies (Beam, 1993; Smolkin, 2007). I thus controlled for the effects of media ownership by coding Family Ownership, Media Conglomerate, and Dual-Stock Structure for different types of newspapers and parent companies: whether a newspaper (or a parent company) was family-owned, whether it was a member of a larger media chain, and whether its parent company had a dual-class stock structure, often used to protect non-market-based objectives, such as journalistic ambition (Smolkin, 2007). I also included two parent-company dummy variables for firms renowned for actively adopting new technologies (Knight–Ridder and Hearst Corporation). In the late 1980s, for example, Knight–Ridder Newspapers, the parent company of The San Jose Mercury News in the 1990s, and The Hearst Corporation, the parent company of Houston Chronicle, actively participated in developing electronic publishing methods (Boczkowski, 2005). Second, at the county level, I first controlled for two effects of the county in which a newspaper was located. As the primary markets for most of the newspapers were their county markets and market size often affects performance or activities (Meyer, 2009; Zeng and Li, 2006), I included County Population [ln] as a control variable. The rate of higher education denotes the percentage of the population holding a bachelor’s degree or higher, which is claimed to approximately capture the Internet penetration rate (Yi, 2008). I thus controlled for the proportion of highly educated people in a given county (County High Education Rate). This high education rate could also be a proxy for the tech hubs. For example, the rate of Santa Clara County in California, where Silicon Valley is located, was two standard deviations greater than the mean. The county-level data from the US Census Bureau from 1990 and 2000 were matched to the newspaper that primarily represented the county. Finally, to control for the effect of Prior Adopters on a newspaper’s adoption, I added the total number of newspapers in the sample that had launched websites in the year prior to the observation year.
Implementation hypotheses
I also used a number of different variables to control for alternative explanations for the implementation hypotheses. At the online news production level, newspapers require a new set of skills to implement digital media. Their experience on the web and the strength of their technical staff can affect their tendency to emphasize different aspects of digital media (Boczkowski, 2004; Zeng and Li, 2006). Based on website launch dates, I calculated Web Presence Length in years. To capture the strength of the online staff, I examined whether a newspaper had an Online Newsroom division and the number of managerial roles in that division (Online Newsroom Size). A separate online division is identified as a viable means of balancing potential conflicting interests between print and digital media (Gilbert, 2005, 2006). Investment in the newsroom, measured by the number of roles in a certain division, is also generally used as an indicator of journalistic quality (Rosenstiel and Mitchell, 2004). At the newspaper and parent company levels, I included the same 11 control variables used previously (Average Circulation [ln], Average Circulation Change, Age [ln], High-Status News Service Subscription, Number Change in News Services, National Newspaper, Family Ownership, Media Conglomerate, Dual-Stock Structure, County Population [ln], and County High Education Rate). Two parent companies, Advance Publications and Knight–Ridder Newspapers, used the same website templates for all their member newspapers; therefore, I controlled for the members of these two parent companies (Advance Publications and Knight–Ridder). Finally, I created County Competitor, coded as “1” if there was more than one daily newspaper in the same county.
Annually updated Yearbook data from 1993 to 2007, except county data, were used to measure all control variables. Average circulation variables lagged 1 year behind, and Table 1 contains summary statistics and bivariate correlations.
Statistical analysis
I used the parametric event history model to analyze the launch dates of websites with time-varying covariates (Cleves et al., 2008). I divided the website launch data into 1-year spells, with year 1 being 1994 and year 10 being 2003. Each year begins in October of the previous calendar year and ends the following September, based on the definition of year by the Audit Bureau of Circulation, the most credible US circulation-auditing organization (now known as the Alliance for Audited Media or AAM). The basic hazard of launching a website initially increased with time but then decreased because certain newspapers persisted in adopting electronic news publication (Li, 2006), following a log-logistic distribution (one of the accelerated failure time metrics). Therefore, I estimate the factors that accelerate or decelerate a newspaper’s time to launch its website (Cox and Oakes, 1984). For the implementation analyses, I applied negative binomial regression models for panel data because the dependent variables were count variables and had high variances, compared to the means (Table 1) (Hilbe, 2007). I used random-effects models with year-fixed effects because the results of a Hausman specification test on the base model were not statistically significant (Hausman, 1978). I also used the seemingly unrelated regression model as error terms for different implementations could be correlated (Zellner, 1962). Applying the seemingly unrelated regression models was further corroborated; the results are available upon request.
Results
Table 2 reports the results of the event history analysis of the website launch date. It presents strong support for the first hypothesis, which states that middle-status newspapers launch their websites faster than other newspapers do. Model 1, including only control variables, shows that an electronic presence using a national online service and subscription to high-status news services negatively affect adoption time. Spanish newspapers, however, launch their websites slower than others. Finally, the number of prior adopters positively affects adoption time. Late adopters may choose not to launch their websites in order to differentiate themselves from competing newspapers that already had websites; this finding mirrors a previous finding (Lowrey, 2003). Model 2 adds the main independent variable and the squared term, and the result strongly supports the first hypothesis. There is a U-shaped relationship between the cumulative number of Pulitzer Prizes and the adoption time, with the inflection point at Pulitzer[ln] = 1.307 (0.0481/0.0184/2). That is, newspapers with 2.5 Pulitzer Prizes (or 5 nominations) launch their websites first
Event history analyses of adoption time.
HS: high-status.
Models are discrete time (yearly) log-logistic models; there were 200 firms, and 201 failure events occurred; one-tailed test where directional predictions are made (s.e.); in Model 3, the ranking measure and the consequent square measure are used instead of the number measures; if the square terms were removed from Models 2 to 6, the linear variables (e.g. Pulitzer [ln]) were not statistically significant.
p < 0.1; *p < 0.05; **p < 0.01; ***p < 0.001.
Tables 3 and 4 present the results of the negative binomial regression analyses of the implementation hypotheses. First, Table 3 offers strong support for Hypothesis 2, which predicts that middle-status newspapers develop websites with the most interactive features. Model 7, a baseline model that includes only control variables, shows that the existence of an online division has a negative effect: separating the online activity from the print newsroom has a negative impact on communication between readers and the newsroom. Moreover, the size and age of a newspaper and the subscription to high-status news services positively affect the level of interactivity on its website. A certain parent company also has a negative impact on interactive website development. Model 9 adds the independent variable and the squared term and indicates that newspapers with 4.5 Pulitzer Prizes (or 9 nominations) have the most interactive websites (an inflection point at Pulitzer[ln] = 1.7 and
Random-effect results of negative binomial regression on interactivity.
HS: high-status.
One-tailed test where directional predictions are made (s.e.); in Model 10, the ranking measure and the consequent square measure are used instead of the number measures; as in Models 8 and 9, Models 10 and 11 with square terms (e.g. Square of Pulitzer [ln]) improved the model fits significantly over the unreported models only with linear terms (e.g. Pulitzer [ln]).
p < 0.1; *p < 0.05; **p < 0.01; ***p < 0.001.
Random-effect results of negative binomial regression on multimedia capability and ad space.
HS: high-status.
Standard errors in parentheses; t-tests are one-tailed for hypothesized effects, two-tailed for control variables; in Models 15 and 19, the ranking measure and the consequent square measure are used instead of the number measure; adding square terms to Models 13 to 15 (e.g. Square of Pulitzer [ln]) did not statistically significantly improve the model fits, while Models 17 to 19 with square terms improved the model fits significantly over the unreported models only with linear terms.
p < 0.1; *p < 0.05; **p < 0.01; ***p < 0.001.
Finally, Table 4 provides the results for Hypothesis 3a (or 3b), which predicts a positive (or negative) relationship between status and multimedia capability (or ad space). Models 12 and 16 are base models for multimedia and ad space, respectively. The results show that parent companies differently influence whether newspapers may emphasize multimedia or advertisement on the web. Newspapers with direct competitors from the same county also focus less on developing multimedia-related websites. Most interestingly, firm size plays an important role in multimedia capability but not in advertisement space; indeed, more financial resources are required to develop multimedia-related reporting skills (Pavlik, 2008; Peng et al., 1999). This positive effect of size on multimedia capability could be the primary reason that the number of the Pulitzer Prizes has a marginally positive effect on multimedia capability (Model 13). If I focus on the top-100 sample (Model 14), the independent variable has a statistically positive effect on developing a multimedia-enriched website, providing overall support for Hypothesis 3a that predicts the positive effect of status on multimedia capability. In Model 15, I include Pulitzer Rank instead of Pulitzer [ln], and the result is consistent with Model 14: a one standard deviation increase in Pulitzer Rank increased Multimedia Capability Index by 20%. Interestingly, in Model 17, the result reveals a U-shaped relationship between status and ad space, with the inflection point at 1.235. In other words, newspapers with 2.5 Pulitzer Prizes have the fewest advertisement spaces on their websites
Discussion and conclusion
This study focused on the role of status in the adoption process of normatively uncertain innovations and examined how organizational status affects the strategic motivation to adopt and implement this type of innovation. I argued that middle-status incumbent firms perceive normatively uncertain innovations more as an opportunity for gain, whereas high- and low-status incumbents perceive them more as a threat of loss and competition, making middle-status firms likely to be the first to adopt these innovations. Furthermore, middle-status firms tend to emphasize the normatively uncertain aspect of the innovations in implementation to better free themselves from their existing constraints on innovation. High-status (or low-status) organizations, in contrast, focus on the normatively compatible (or incompatible) aspect that promises the best outcomes for themselves. The empirical results provide overall support for my arguments in the context of large US daily newspapers that have adopted a normatively uncertain innovation: digital media. Middle-status newspapers were the first to launch their websites and most actively allowed readers to participate in news-producing activities that could potentially change the identity of newspapers as the sole producers of news. High-status newspapers, in contrast, highlighted its normatively compatible aspect by developing multimedia-enriched websites. Low-status newspapers (along with high-status newspapers) focused on the normatively incompatible aspect of digital media, namely, advertising on their websites, to directly improve their performance without being excessively concerned with normative pressures.
To corroborate the empirical findings that middle-status firms differently responded to digital media compared to other-status firms at the time of adoption, I collected and read the editorial statements made by 111 newspapers in the sample when they launched their websites. Consistent with the suggested mechanisms, middle-status newspapers often framed digital media as opportunities, while other-status firms often emphasized digital media as threats. For example, The News & Record in North Carolina, a middle-status newspaper with three finalists in 1994, stated that digital media is “a medium not bound by paper … and provides us with … some wonderful opportunities” (Ladd, 1994: B1). In contrast, The Boston Globe, a high-status newspaper, stated that it established “an early presence on line to help preserve” its current print business (Auerbach, 1995: 11). To further examine that different-status newspapers emphasized different elements of digital media in their editorials, I also coded whether the statements emphasized interactivity or advertisements on the web, using NVivo software (Bazeley and Jackson, 2013). 10 Simple ordinary least squares regression results revealed an inverted U-shaped relationship between Pulitzer Rank and how much of the statement is coded as interactivity, but a U-shaped relationship between Pulitzer Rank and how much of the statement is coded as advertisement. In other words, middle-status newspapers, such as The News & Record, The Plain Dealer in Ohio (one Pulitzer and two finalists in 1995), and The Denver Post in Colorado (four Pulitzers and one finalist in 1995), emphasized interactivity in their statements but never mentioned advertisement potential on the web. In contrast, frequent Pulitzer Prize–winning newspapers, such as The New York Times and The Boston Globe, and newspapers that had never received Pulitzer Prizes, such as The Morning Call in Pennsylvania and The Los Angeles Daily News, did not mention interactivity in their statements but often emphasized the advertisement potential on the web. The systematic differences in word choices across status further support that normative uncertainty indeed mattered to newspapers at the time of adoption.
This article makes several important contributions to research on status, on the adaptation of innovations, and on technology. First, it contributes to research on the role of status in markets and middle-status conformity. Current status research focuses on social psychological and structural explanations for why middle-status organizations tend to face the strongest conformity pressures, which often prevent them from innovating first (Menzel, 1960; Phillips and Zuckerman, 2001). I have identified the important boundary condition for the innovation, not the conformity, of middle-status organizations: normative uncertainty. I argued and provided empirical support that normative uncertainty is an underlying mechanism of relaxing innovation constraints faced by middle-status organizations. This emphasis on normative uncertainty as an underlying mechanism distinguishes my study from another study that theorized about middle-status deviance (Jensen, 2010). In the Danish film industry, middle-status film actors and actresses were more likely to deviate and participate in the legitimizing process of a new movie genre, the sex comedy genre, in the late 1960s and early 1970s. High-status actors could deviate from norms and expectations, but this ability did not in itself motivate them to do so, as they faced high opportunity costs of losing their high-status positions. This highlights a well-established economic concern as an underlying mechanism behind middle-status deviance. My study, in contrast, focuses on how lack of consensus regarding normative compatibility opens up innovation opportunities for middle-status firms with respect to when and how to adopt normatively uncertain innovations.
Second, my article contributes to research on the adaptation of innovations (e.g. Ansari et al., 2010; Chandler, 2014; Okhmatovskiy and David, 2012; Westphal et al., 1997). Adaptation refers to “the process by which an adopter strives to create a better fit between an external practice and the adopter’s particular needs … during implementation” (Ansari et al., 2010: 71). Building on this stream of research, I have shown that organizational status affects not only when firms adopt normatively uncertain innovations but also how they differently implement the innovations. Based on unbiased access to newspapers’ websites through the Internet Archive, I also empirically demonstrated that when facing a normatively uncertain innovation—digital media—different-status newspapers indeed envisioned and implemented alternative futures for the newspaper industry. Specifically, the different implementation approaches of the same innovation are associated with firms’ attempts to shape these innovations in a manner that creates a better fit between the innovations and their needs (cf. Ansari et al., 2010; Bijker et al., 1987). It is therefore important to account for how innovations can be differently interpreted by organizations in various social positions to better understand their adaptation process.
Finally, my study provides interesting insights into the role of status in incumbents’ responses to technological discontinuities. The previous literature on status in general has focused on how status influences the diffusion of incremental innovations (Podolny, 2005). Technological discontinuities, defined as significant breaks in technological variation from previously dominant technologies or designs (Tushman and Anderson, 1986), tend to be relatively unexplored from the status perspective (Podolny, 2005: 134). This lack of academic attention is consistent with the technology literature that generally argues that incumbents often fail to successfully respond to technological discontinuities (Tushman and Anderson, 1986). In particular, this stream of research mainly emphasizes incumbents’ lower incentive and/or their inability to cope with discontinuous technologies. Technological discontinuities, however, are likely to create normative uncertainty because a significant break in technological variation can be differently interpreted by various incumbents, based on their histories, motivations, and needs (Tushman and Anderson, 1986). My study suggests that contextual factors, that is, incumbents’ social positions, are also important to consider when understanding variance in the incumbents’ strategic responses to technological discontinuities. A subgroup of incumbents, that is, middle-status firms, is most likely to experiment with a technological discontinuity first if this significant break in technological variation, which is not clearly defined under the existing norms and expectations, helps them to free themselves from existing conformity pressures.
My article has some limitations that should encourage future research. First, the specificity of the newspaper industry could result in a generalizability concern, especially because most large daily newspapers in the United States are regional monopolies (Meyer, 2009). One might argue that regional newspapers do not directly compete with newspapers in different regions; therefore, they need not be concerned with national-level awards or competitors in other markets. To address this concern, I re-estimated the models in unreported analyses without nationwide newspapers and certain high-status newspapers that often had a nationwide distribution, such as The New York Times. Even when I excluded these nationwide newspapers, the results generally supported my hypotheses. These findings suggest that regional newspapers are also concerned with their status levels, even when they do not directly compete with other regional newspapers. Another potential limitation is that by focusing on the motivation and behavior of firms with respect to innovations, this study does not address the adoption consequences of different attempts. Distinctively theorizing a new innovation is known to be less effective in established fields than in emerging fields (David et al., 2013). According to my hypotheses, different-status incumbents attempt to shape normatively uncertain innovations to become more beneficial to their own groups: middle-status incumbents attempt to shape these innovations in a normatively uncertain way, which can be a potential source of an emerging market. In contrast, high- and low-status firms are not strongly motivated to change the established market as long as they reasonably maintain their current businesses even with the innovations. Exploring the subsequent relationships between adoption behavior and performance is important for understanding the diffusion process of normatively uncertain innovations from the status perspective and how this process affects the nature of the innovations in question and the locus of potential institutional change.
In conclusion, my study makes important contributions to research on status and the diffusion of innovations. Middle-status and other-status organizations perceive normatively uncertain innovations differently, influencing the adoption and implementation decisions of these innovations. My study shows that the status of incumbents can either facilitate organizations’ movements beyond “traditional” boundaries or constrain their movements within these boundaries. Specifically, it suggests that middle-status organizations, often confronted by the strongest constraints on innovation, may be most eager to cross the traditional boundaries surrounding them when there is lack of consensus regarding whether an innovation is consistent with existing market norms.
Footnotes
Appendix 1
A table of contemporaneous quotes for the three elements of digital media.
| Normatively uncertain interactivity |
|---|
| Reporters at The San Jose Mercury News are urged to respond to electronic mail from readers. Several reporters have conducted their own on-line sessions to take questions or comments. In interviews, sentiment appeared to be divided on how significant that communication was. Some reporters said the level of discourse varied from “babble” and petty complaints to insights about what kinds of coverage might interest readers and even news tips that might never have reached the newsroom otherwise. (Glaberson, 1994: D6) |
| Anybody can post anything on the Net, which is good in that it creates a limitless variety of material. It’s bad in that much of what one encounters in cyberspace is garbage and a huge waste of time. (Fancher, 1995: A2) |
| In one example of how newspapers and their Web sites can link with the community, Marian Anderfuren, The Virginian-Pilot’s editor for the fast-growing suburb of Suffolk (VA), got the city’s mayor to answer residents’ questions about development policies on a message board. At times, Anderfuren has joined the discussion to pose questions. … Of course, it is possible to include too many Web enhancements. When overused, these components can create clutter and confuse readers. (South, 1999) |
| Compared to a newspaper, the Internet allows far greater participation in many-to-many communication. … [So,] the Internet was described as a “democratic cyberspace,” the ideal place of virtual democratic community building and maintaining (244-245). … [However,] some journalists appeared horrified at the idea that readers want to engage in a discussion with them. Many traditional journalists see the online world as a place filled with unverified rumors, a place where traditional journalism standards count for nothing. (Ye and Li, 2006: 255) |
| Normatively compatible multimedia capability |
| We try to make things a little funkier, more graphically driven … and more fun without compromising our commitment to serious, responsible journalism. The [Chicago] Tribune has successfully used this approach on several occasions. One was … a series about a murder victim whose body parts were carried home by two farm dogs; readers could choose to experience the story from a variety of viewpoints: the victim’s, the investigators,’ and even the dogs …’ The paper’s online staff … takes digital still photos and collects Real Audio clips by recording phone interviews with the source’s permission. (Lasica, 1996: 30) |
| The most impressive aspect of [the newspaper’s Web site is] … full-text versions of almost every significant story from each day’s printed edition of The [Las Vegas] Sun … frequently accompanied by special features incorporating video and sound to take advantage of the Web’s multimedia capability. (Las Vegas Sun, 1996: NEWS2) |
| Internet provides new choice for local news; Newscoast.com extends the multimedia capabilities of The Herald-Tribune. … The intention of this company is to be a full multimedia provider of news and other important information. (Matrullo, 1997: 1A) |
| The Washington Post enhanced [a story] online by embedding small video clips in the text. By mouse-clicking one of the videos, readers could see-and hear-exactly what Weingarten [the reporter of the article] described. [The article] quickly became one of the most-viewed stories ever to appear on washingtonpost.com. The … probably most important reason [of its popularity] was that there was video, so this became quickly viral. The story … was more effective online than it was in print … [and] it happens more and more now as we figure out how to use video as an adjunct to a story. (Layton, 2007: 26) |
| Normatively incompatible advertising practices |
| The transition to electronic journalism is bringing with it some challenges to old rules of the trade. Mercury Center [the name of The San Jose Mercury News’ online experiment], for instance, includes news and business employees, breaching the traditional wall that separates the two at most newspapers. Its executives say they are committed to maintaining editorial independence, but contend that news and business employees must work together to shape the new venture into a money-making enterprise. (Glaberson, 1994: D1, D6) |
| Read a book review at nytimes.com , click a Barnes & Noble link, and order the book. You get quick service, the giant bookseller makes a sale, and the Times get a commission. Is it just good business-or an inappropriate melding of a respected news site’s advertising and editorial content? (Lynch, 1998: 45) |
| [N]ot every publication shares The New York Times’ high journalistic standards, and we can expect the ethical Maginot line to be breached repeatedly in the coming years. Already, advertisers are banking on the absence of the traditional “Chinese wall” between editorial and advertising in many online publications. … “An advertiser asked us to create an interactive game based on the news of the day and put it at the top of our home page. To enter and win prizes, the user had to download some of their software. The revenue was attractive, but we decided we’d be selling our soul a little bit. So, we told them they could run an ad, but we couldn’t dress up their promotion to look like content.” (Lasica, 1997) |
| Most U.S. newspapers do not run display advertising on front pages, citing the need to mark the independence of articles from the newspaper’s advertisers. … Online, however, banner advertising appears prominently on the home pages of most online newspapers. … “In time, we in the media and the journalism profession will begin to realize that we have to take back that real estate on the front page of our Web sites.” (Gaw, 1997: 1D) |
Appendix 2
Acknowledgements
The author is very grateful to Michael Jensen, Heeyon Kim, Jinkyung Na, Donghoon Shin, Jim Westphal, Gautam Ahuja, Jerry Davis, Mark Mizruchi, and seminar participants at the University of Michigan, Southern Methodist University, Northwestern University, the University of California at Berkeley, HEC Paris, Hong Kong University of Science and Technology, Yonsei University, and Academy of Management Annual Conference for their excellent and helpful comments on earlier drafts of this manuscript.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/ or publication of this article: The study was supported (in part) by the 2017 Research Fellowship Fund of the Sangnam Institute of Management and the 2017 Management Field Research Fund of the Yonsei School of Business.
