Problem-solving effort and success in innovation ...

Journal of Operations Management 36 (2015) 187–200

Contents lists available at ScienceDirect

Journal of Operations Management journal homepage: www.elsevier.com/locate/jom

Problem-solving effort and success in innovation contests: The role of national wealth and national culture Jesse Bockstedt a,1 , Cheryl Druehl b,∗,1 , Anant Mishra b,1 a b

Eller College of Management, University of Arizona, United States School of Business, George Mason University, United States

a r t i c l e

i n f o

Article history: Received 12 March 2014 Received in revised form 15 December 2014 Accepted 17 December 2014 Available online 26 December 2014 Keywords: Innovation contests Problem solving National culture Crowdsourcing Econometric analysis

a b s t r a c t Innovation contests allow firms to harness specialized skills and services from globally dispersed participants for solutions to business problems. Such contests provide a rich setting for operations management (OM) scholars to explore problem solving in global labor markets as firms continue to unbundle their innovation value chains. In this study, we examine the implications of specific types of diversity in innovation contests on problem-solving effort and success. First, we conceptualize diversity among contestants in terms of national wealth (measured as gross domestic product per capita (GDPP) adjusted for purchasing power parity) and national culture (measured using the culture dimensions of performance orientation and uncertainty avoidance) and examine how such factors influence problem-solving effort. Next, we examine how differences between contestants and contest holders in terms of the above factors influence contest outcomes. Using data from a popular online innovation contest platform and countrylevel archival data, we find that contestants from countries with lower levels of GDPP are more likely to exert greater problem-solving effort compared to other contestants. With regard to national culture, we find that performance orientation and uncertainty avoidance have positive and negative effects, respectively, each of which weakens with increasing levels of GDPP. Finally, our analysis provides evidence of homophily effects indicating that contestants who share greater similarities with the contest holder in terms of national wealth and national culture are more likely to be successful in a contest. We discuss the implications of the study’s findings for contest holders and platform owners who organize innovation contests, and for emerging research on innovation contests. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Rapid growth in social media and the global reach of the Internet have fundamentally changed the way firms execute the various activities of their innovation value chains (Chesbrough, 2007; Billington and Davidson, 2013). Firms are seeking novel ways to collaborate and develop new products and services to meet the increasingly competitive pressures of the “flat world” (Friedman, 2005; Metters et al., 2010). To facilitate this, recent years have seen an emergence of cost-effective “innovation contests” that harness specialized skills and services from a crowd of globally distributed individuals to provide new creative ideas and solutions to

∗ Corresponding author. Tel.: +1 7039939760. E-mail addresses: [email protected] (J. Bockstedt), [email protected] (C. Druehl), [email protected] (A. Mishra). 1 Note: Authors are listed in alphabetical order. http://dx.doi.org/10.1016/j.jom.2014.12.002 0272-6963/© 2014 Elsevier B.V. All rights reserved.

challenging business problems (von Hippel, 2005; Terwiesch and Ulrich, 2009). While innovation contests, such as those conducted on InnoCentive.com, 99designs.com, and Logomyway.com, have seen significant growth in recent years, the notion of such contests per se is not new. A frequently cited historical example is the “Longitude Prize” contest that was held by the British Parliament in the 18th century (Jeppesen and Lakhani, 2010). Open to the general public and with a prize amount of up to £20,000, the goal of the contest was to find a practical method for accurately determining the longitude position of a ship in transoceanic voyages. As another example, in 1795 Napoleon launched a competition with a prize amount of 12,000 francs to invent a method of preserving food for his army (Wagner, 2011). Interestingly, both contests are based on important operational problems. Notwithstanding their existence over centuries, innovation contests today differ from traditional contests discussed above in a fundamental way. Specifically, online platforms decentralize the

188

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

problem-solving effort across a large, geographically dispersed group of participants from diverse economic and cultural backgrounds. Thus, today’s online innovation contests provide contest holders with unprecedented access to a global workforce. The “global” nature of this workforce can be gauged by looking at some recent statistics from popular innovation contest platforms. For example, Innocentive—a science-based innovation contest platform—has a registered base of about “300,000 [participating solvers] from nearly 200 countries” (Source: innocentive.com), Logomyway.com—a logo-design innovation contest platform—has over 15,000 participating designers from more than 100 countries (Source: logomyway.com), and TopCoder—a computer programming contest platform—has about 9000 participants from 61 different countries in its algorithm development contests (Source: topcoder.com). Although diversity of participants, including economic and cultural backgrounds, has been recognized as a defining feature of innovation contests (von Hippel, 2005; Daniel et al., 2013), we have a limited understanding from prior studies of its effects on participant engagement, and more specifically, on the problem-solving effort expended by participants. Prior research on innovation has frequently examined the role of “team” diversity (i.e., diversity among individuals within a team in terms of gender, age, educational background, functional background, etc.) in the effective functioning of new product development teams (e.g., Ancona and Caldwell, 1992; Sarin and McDermott, 2003; Narayanan et al., 2013). However, in innovation contests, the problem-solving effort is driven by individual participants who differ widely from one another in terms of economic and cultural backgrounds (Terwiesch and Ulrich, 2009; Billington and Davidson, 2013). How do these sources of diversity in innovation contests influence the problemsolving effort expended by participants? Additionally, many innovation contests are characterized by high levels of evaluation uncertainty and the selection of the winning solution is “taste-based,” depending upon the subjective preferences of a contest holder (Terwiesch and Xu, 2008; Erat and Krishnan, 2012). It remains to be understood how differences between contestants and the contest holder in terms of economic and cultural factors affect a contestant’s success. That is, do innovation contest platforms provide a “level playing field” for all participants (Howe, 2006; Belsky, 2010), or does homophily effect—the propensity of individuals to associate with others with similar social, cultural, economic and/or demographic characteristics (Milliken and Martins, 1996; McPherson et al., 2001)—exist in such settings? Our study attempts to address the above gaps in the prior literature. We conceptualize diversity in an innovation contest setting in terms of economic and cultural factors. Since these sources of diversity are inherently related to differences in nationalities among participants, we represent differences in economic and cultural factors among participants in terms of their national wealth and national culture. We measure national wealth using the gross domestic product per capita (or GDPP), adjusted for purchasing power parity (PPP) (e.g., Gefen and Carmel, 2008; Kull and Wacker, 2010). For national culture, we focus on two particular dimensions relevant to our study: performance orientation and uncertainty avoidance (House et al., 2004). Next, we develop and test hypotheses that examine the role of economic and cultural factors on problem-solving effort in innovation contests. Finally, we shed greater light on the role of homophily in such settings by examining how differences in national economic and cultural factors between contestants and contest holders influence contest outcomes. The empirical analysis is carried out using an integrated dataset that comprises detailed data from 1024 innovation contests and 2626 unique contestants (resulting in approximately 45,000 contest-contestant observations) from Logomyway.com. This data

is matched with country-level archival data on GDPP (adjusted for PPP) and national culture dimensions. Results indicate that both national wealth and national culture influence problem-solving effort and outcomes in innovation contests. We find that contestants from countries with lower levels of GDPP are more likely to make a larger number of submissions compared to other contestants. With regards to the culture dimensions, we find that performance orientation and uncertainty avoidance have opposing effects on the problem-solving effort of contestants. That is, increasing levels of performance orientation are associated with an increase in problem-solving effort, while increasing levels of uncertainty avoidance are associated with a decrease in problemsolving effort. However, both relationships become weaker as GDPP increases. Our analysis also provides evidence of homophily effects in innovation contests, indicating that contestants that share greater similarities with the contest holder in terms of national wealth and national culture are more likely to be successful compared to other contestants. To the best of our knowledge, our study is the first of its kind to highlight the strong links between macrolevel factors associated with the contest environment, individual problem-solving effort and innovation contest outcomes. These findings lead to valuable insights for the design of such contests and global labor markets.

2. Theoretical background and hypotheses The developing literature on innovation contests has its roots in the process model of innovation in operations management (e.g., Ha and Porteus, 1995; Dahan and Mendelson, 2001; Terwiesch and Xu, 2008) and research tournaments in economics (e.g., Lazear and Rosen, 1981; Nalebuff and Stiglitz, 1983). The process model of innovation conceptualizes problem solving as a search for the best solution from among a set of parallel experiments undertaken by the solver (Ha and Porteus, 1995; Dahan and Mendelson, 2001). The performance outcome from this search process is the highest realization from the set of parallel experiments. Building upon this idea, Terwiesch and Xu (2008) modeled problem solving in an innovation contest as a set of parallel experiments where the performance of a submitted solution (resulting from an experiment) is a function of the effort undertaken by the contestant and her prior expertise (i.e., experience and knowledge) in the problem domain. Lakhani et al. (2007) surveyed contestants on Innocentive, a science-focused innovation contest platform, and found that winning solvers put in twice as much problem-solving effort into their solutions, on average, compared to non-winners. The centrality of effort measurement in problem-solving research is also evident from the extensive economics literature on research tournaments (e.g., Taylor, 1995; Che and Gale, 2003; Moldovanu and Sela, 2006). This research has largely examined how variations in tournament design characteristics (e.g., the size and the number of incentives, the number of tournament stages, and the size of the participant pool) influence problem-solving effort by a participant. For instance, Taylor (1995) found that limiting the number of participants in a contest led to an increase in their problem-solving efforts due to a greater perceived chance of winning from reduced duplication. Fullerton and McAfee (1999) further argued that it is necessary to limit entry because the evaluation of problem-solving effort is not a costless exercise. The above studies are analytical in nature, where problem-solving effort is typically represented as a stochastic variable and contestants are assumed to submit only one solution to a contest. More recently, empirical studies on innovation contests have started to focus on contest environments that allow for multiple submissions per contestant (e.g., Yang et al., 2010; Bayus, 2013;

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

Bockstedt et al., 2014). For instance, Yang et al. (2010) measured the problem-solving effort by a contestant in terms of the submission speed (i.e., the time taken by the contestant to submit a solution). In contrast, both Bayus (2013) and Bockstedt et al. (2014) captured problem-solving effort by a contestant in a given contest as the number of solutions submitted by the contestant. Such a characterization of problem-solving effort is consistent with existing studies on creativity and idea generation (e.g., Simonton, 2003; Girotra et al., 2010), where the quantity of solutions proposed has been found to be strongly correlated with the quality of the final solution. Our study utilizes data from an innovation contest setting that allows multiple submissions per contestant. Hence, consistent with prior studies, we conceptualize and measure problem-solving effort as the number of solutions submitted by a contestant in a given contest. Much of our understanding of the effects of diversity is derived from prior studies on R&D and new product development teams (e.g., Ancona and Caldwell, 1992; Sarin and McDermott, 2003; Narayanan et al., 2013; Daniel et al., 2013). This literature has largely conceptualized and measured diversity as a “within-team” phenomenon, wherein individuals comprising the team may differ from each other in terms of demographic attributes, such as gender (Etzkowitz et al., 2000), age (Lau and Murnighan, 1998), functional background (Ancona and Caldwell, 1992; Sarin and McDermott, 2003; Peters and Karren, 2009), and prior experience (Harrison and Klein, 2007; Narayanan et al., 2013). Building upon this literature, studies on innovation contests have examined how differences among contestants in terms of their technical backgrounds, social backgrounds, and prior success in the contest platform influence problem solving (Jeppesen and Lakhani, 2010; Boudreau et al., 2011; Bayus, 2013). However, given the global reach of innovation contests, with individuals participating from many countries (Brabham, 2008), these contests also experience diversity among contestants in terms of economic and cultural factors (Terwiesch and Ulrich, 2009; Billington and Davidson, 2013). We know little about the implications of such factors on problem-solving effort and outcomes in innovation contests. Since economic and cultural factors are inherently related to differences in nationalities among participants, we capture them in terms of national wealth and national culture, respectively. We discuss these factors in greater detail below and examine their implications for problem-solving effort and outcomes in an innovation contest setting. 2.1. National wealth The economic factor, national wealth, when associated with an individual, allows for the comparison of an individual’s pecuniary benefits derived from committing her time to a task with those derived in another nation. Prior studies in the innovation contest literature (e.g., Lakhani et al., 2007; Erat and Krishnan, 2012), and the open source literature (e.g., von Hippel, 2005; Daniel et al., 2013), have identified prizes and external incentives as important drivers for individuals to participate in such settings. A key assumption underlying the effects of prize amount is that its “value” is consistent across individuals. In reality, this assumption is not necessarily true. In innovation contests, where globally distributed individuals from many different countries participate, significant differences in the value of prize amounts can arise due to differences in economic prosperity levels of individuals across countries. To capture such differences, economists use gross domestic product per capita (GDPP)—which measures in US dollars the total market value of all the goods and services produced per person by a country in a given year (CIA World Factbook). The greater the GDPP, the greater is the average income level of citizens in that country. However, an average individual’s economic prosperity also depends upon the prevailing price levels of goods and services

189

in the country (Lee, 1976; Froot and Rogoff, 1995). Holding average income constant, higher (lower) prices can reduce (enhance) the economic prosperity of individuals in a country, compared to another country. Consider for example, two contestants, A and B, who have each won a contest with a US$200 prize amount. Contestant A is from China, while Contestant B is from the US. For this amount, and using 2014 price levels, Contestant A can purchase around 73 McDonald’s Big Mac burgers, while Contestant B can purchase around 43 Big Macs (Source: www.economist.com/content/big-mac-index). That is, a $200 prize amount provides greater purchasing power to Contestant A compared to Contestant B. Therefore, for comparability purpose, the GDPP for a country is adjusted by its purchasing power parity (PPP) with the US. 2.2. National culture With the increasing globalization and geographical distribution of firms’ value chain activities over the past decade, the role of national culture has come under increasing scrutiny in the OM literature across a variety of industry contexts (e.g., Gray and Massimino, 2014; Naor et al., 2010; Kull and Wacker, 2010; Handley and Benton, 2013; Hahn and Bunyaratavej, 2010; Cheung et al., 2010). Although many competing theoretical frameworks for studying national culture exist (e.g., Hall and Hall, 1990; Smith et al., 1995), Hofstede’s culture dimensions framework (Hofstede, 1980, 2001) and the GLOBE framework (House et al., 2004) remain by far the most widely applied frameworks used in studies of cross-cultural differences and effects in the OM literature and beyond. Hofstede (1980, p. 21) defined culture as “the collective programming of mind which distinguishes the members of one human group from another.” In his original study, carried out in the 1970s, Hofstede used data on more than 100,000 employees from IBM across 50 counties to conceptualize national culture as a multidimensional construct comprised of a variety of dimensions: power distance, individualism/collectivism, masculinity, and uncertainty avoidance. The list of countries included has since increased to approximately 100 and two new dimensions have been added to Hofstede’s framework, long-term orientation and indulgence. However, national data on these two dimensions is very limited. The GLOBE study was carried out in the mid-1990s and involved data collection from more than 17,000 middle managers in 951 organizations across three industries and 62 countries (House et al., 2004). This study built upon and extended the dimensions included in Hofstede’s framework to include nine dimensions: power distance, institutional collectivism, in-group collectivism, gender egalitarianism, uncertainty avoidance, performance orientation, assertiveness, future orientation, and humane orientation. While the various dimensions across the Hofstede and GLOBE frameworks capture multiple facets of national culture differences, not all dimensions may be salient in each research context (Leidner and Kayworth, 2006; Siegel et al., 2013). Consequently, for conceptual and empirical parsimony, we focus on a specific set of culture dimensions that seem to be directly relevant to our research context. Specifically, given that innovation contests involve unstructured problem-solving in a competitive environment, the national culture dimensions performance orientation and uncertainty avoidance are relevant and may provide insights on the behavior of contestants. The relevancy of the performance orientation dimension stems from the fact that innovation contests involve a large number of contestants competing for one or more monetary prizes and winning signals recognition and achievement. Societal views on competition may potentially influence contestants’ problem-solving efforts and resulting outcomes in an innovation contest setting. Further, the choice of the uncertainty avoidance dimension arises because innovation contests often involve

190

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

problems that are ambiguous and ill-specified with a high degree of evaluation uncertainty (Terwiesch and Xu, 2008; Erat and Krishnan, 2012), and which may require creative “out-of-the-box” solutions (Jeppesen and Lakhani, 2010). Therefore, understanding the implications of these cultural dimensions on contestants’ problemsolving efforts and outcomes is a relevant question to the emerging research on innovation contests. 2.3. Effects of national wealth and national culture on problem-solving effort Macro-level economic indicators, such as GDPP, have been frequently used in extant research to study relevant trends at the firm and individual levels of analysis (e.g., Zhu and Kraemer, 2005; Gefen and Carmel, 2008; Kull and Wacker, 2010; Hahn and Bunyaratavej, 2010). For instance, Zhu and Kraemer (2005) examined the antecedents of corporate e-business use across developed and developing countries using GDPP as an important dimension of the national environment. Similarly, Kull and Wacker (2010) used GDPP as a control for inter-country differences in development, market performance orientation, strength of auditing and standards in their study of the effectiveness of quality management practices between Asian and non-Asian countries. Focusing on the individual level, Gefen and Carmel (2008) examined the role of GDPP on “freelance” programmer behavior in an online programming platform, Rent A Coder, where programmers bid on very small IT projects (with median and average bid amounts of approximately $50 and $129, respectively). Highlighting differences between a Romanian programmer and a Canadian programmer, they noted that while both programmers may bid $100 on an IT project, the bid amount is worth significantly more to the Romanian programmer compared to the Canadian programmer. Further, the Canadian programmer is more likely to be attracted by alternative wage opportunities in the local marketplace where a higher wage structure prevails, compared to that for the Romanian programmer. As a result of differences in economic prosperity levels between Romania and Canada as denoted by GDPP levels, Romanian programmers are likely to be “hungrier” and expend greater problem-solving effort in completing the IT project compared to Canadian programmers. Taking the above studies into account, we build upon the extant literature on innovation contests that highlights the incentive effects of a prize on problem-solving effort (Terwiesch and Xu, 2008; Yang et al., 2010). Specifically, we argue that variations in GDPP levels across countries introduce asymmetries in the incentive effects of the prize amount. That is, the same prize in US dollars has greater incentivizing effects on contestants from countries with lower GDPP, compared to those with higher GDPP. Therefore, all else remaining constant, the marginal expected returns per unit of problem-solving effort is likely to be higher for contestants from lower GDPP countries compared to those from higher GDPP countries. In addition, the unit cost of problem-solving effort is also likely to be costlier for contestants from higher GDPP countries, given the potential for greater economic gains from alternative sources. For these reasons, we expect contestants from countries with lower GDPP to expend greater problem-solving effort in the contest compared to those from higher GDPP countries. Hypothesis 1. Contestants from countries with lower levels of GDPP will exert more problem-solving effort in innovation contests compared to those from countries with higher levels of GDPP. The national culture dimension of performance orientation represents a preference in a country’s society for achievement and material reward for success. Individuals from countries that rank high on the performance orientation dimension are therefore expected to expend greater effort in problem solving when

there are clearly outlined monetary incentives and opportunities for achievement and recognition, as in innovation contests. The link between culture and individual effort has been reported across a variety of competitive environments in prior studies. For instance, in a study of sales force compensation schemes, Rouziès et al. (1999) found that sales people operating in cultures with a greater focus on competitiveness are more likely to be concerned with the financial results associated with their effort and explore greater opportunities for enhancing those results. Higher focus on performance orientation in certain national cultures can also provide individuals from such cultures with greater motivation to learn and improve one’s problem-solving skills for future performance (e.g., von Hippel, 2005). In an innovation contest, such motivation may manifest through an increasing number of submissions by a contestant, which allows greater opportunities to receive feedback from the contest environment and enables learning by doing (Argote, 1999). In sum, being from countries with greater levels of performance orientation should result in higher problem-solving effort, not only because of the potential for immediate financial and reputation gains, but also for the potential for these gains to occur in the future. The other dimension, uncertainty avoidance, expresses the degree to which the members of a society feel uncomfortable with uncertain and ambiguous situations. Countries exhibiting strong uncertainty avoidance maintain rigid codes of belief and behavior and are intolerant of unorthodox behavior and ideas. Creativity and problem solving in competitive environments are usually associated with experimentation and uncertainty, requiring acceptance of risk and change (Shane, 1993). These characteristics are seemingly counter to a culture characterized by high levels of uncertainty avoidance. Indeed, existing studies on new product development have found a positive association between risk-taking behavior in organizations and new product performance (e.g., Nakata and Sivakumar, 1996; Calantone et al., 2003). Furthermore, Shane (1993) found that uncertainty avoidance had the strongest influence among all culture dimensions on national levels of innovation. In the economics literature, Huang (2012) also linked effort to uncertainty avoidance, noting that the optimal effort by an individual decreases with increasing levels of uncertainty avoidance when the cost of effort is either monetary or non-monetary. Additionally, Alary et al. (2013) showed that individuals with high levels of uncertainty avoidance emphasize unfavorable outcomes over favorable outcomes and attempt to reduce costs by exerting lower levels of effort compared to others. Extrapolating these findings to an innovation contest setting, we expect higher levels of uncertainty avoidance to reduce an individual’s motivation to experiment and try alternative solutions. We therefore propose the following hypotheses. Hypothesis 2A. Contestants from countries with higher levels of performance orientation will exert more problem-solving effort in innovation contests compared to those from countries with lower levels of performance orientation. Hypothesis 2B. Contestants from countries with higher levels of uncertainty avoidance will exert less problem-solving effort in innovation contests compared to those from countries with lower levels of uncertainty avoidance. The above national culture factors represent the intrinsic set of values among individuals from a country that motivate problemsolving effort in innovation contests. Researchers have long noted that these inherent cultural tendencies among individuals are often moderated by the economic implications of such behavior (Deci, 1971; Kreps, 1997). Consistent with this theme, we examine the interplay between GDPP and national culture dimensions. Specifically, given that cultures with higher performance orientation

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

place greater emphasis on material gains and incentives, we anticipate the positive association between performance orientation and problem-solving effort to vary among contestants as the real value of the contest prize amount differs among them. As the value of the prize amount is significantly higher for contestants from countries with lower GDPP, we anticipate that performance orientation will have a stronger association with problem-solving effort for individuals in such countries compared to those from countries with higher GDPP. Hypothesis 3. GDPP will negatively moderate the relationship between performance orientation and problem-solving effort in innovation contests. The specific direction of the moderating effect of GDPP on the relationship between uncertainty avoidance and problem-solving effort is less clear. Prior research has shown that individuals from countries high on uncertainty avoidance are more likely to choose options with uncertain outcomes when they are associated with monetary gains (Harrison et al., 2009; Ladbury and Hinsz, 2009). For example, Harrison et al. (2009) showed that those from countries high on uncertainty avoidance were more likely to volunteer for treatment in a randomly assigned process when offered monetary compensation for showing up. That is, the inherent tendency to exert less effort in contexts characterized by high levels of uncertainty was significantly reduced when the potential for gaining monetary incentives was higher. In innovation contests, contestants from lower GDPP countries are more likely to gain economically from their increased problem-solving efforts in the contests relative to those from higher GDPP countries. Therefore, one might expect the effects of uncertainty avoidance on problemsolving effort to be lessened in lower GDPP countries. However, research has also shown that the effects of uncertainty avoidance are weakened when decision makers are better off economically (e.g., Williamson and Mathers, 2011), and when the economic costs of unsuccessful problem-solving efforts are lower. In an innovation contest, greater economic prosperity may weaken a contestant’s inherent aversion to trial and error learning and experimentation that is often necessary for developing solutions to uncertain problems. Given the limited prior theory and potential for countervailing arguments, it remains an open empirical question as to how GDPP moderates the relationship between uncertainty avoidance and problem-solving effort. We therefore do not hypothesize a specific direction for this moderating effect and present the subsequent empirical analysis as exploratory to uncover this potential relationship. 2.4. National culture distance, national wealth distance, and success In initiating an innovation contest, the contest holder develops the contest problem specification or contest brief, which outlines the specific context of the problem (e.g., the industry in which the problem is based, customer segment affiliated with the problem, the implications of the problem for the local economy), and the expected characteristics of a solution. Often the problem specification may only provide a brief overview of the contest holder’s requirements and may lack details needed for a contestant to solve the problem. Given this uncertainty, effective problem solving by contestants requires them to understand the specific context in which the problem is based and address the problem from the contest holder’s perspective. Since innovation contests allow for contest holders and contestants to be globally distributed, there is likelihood for significant differences in economic prosperity and culture between a contest holder and any given contestant. These differences may make it difficult for the contestant to fully understand the context of the problem and limit her ability to develop

191

an effective solution. More specifically, increased levels of culture distance between a contestant and the contest holder may cause linguistic subtleties and tacit knowledge embedded in the contest problem to be interpreted in divergent ways between the two sides (Nakata and Sivakumar, 1996; Gefen and Carmel, 2008). In addition, differences in economic background between a contestant and the contest holder may complicate the understanding of the business problem from the perspective of the contest holder’s local economy. For example, consumer characteristics, consumption patterns, consumer preferences, and situated business knowledge in the contestant’s local economy may differ dramatically from those of the contest holder’s target market. Thus, these differences may contribute to the inherent uncertainty that is present in the contest environment and result in solutions that fail to meet the tacit norms and standards of the target economy (Milner and Collins, 2000; Bang et al., 2005). Prior research has found evidence of homophily effects among organizational entities and individuals across various settings such as finance, international trade, and electronic commerce (e.g., Disdier and Head, 2008; Graham et al., 2009; Hortac¸su et al., 2009). As noted previously, homophily refers to the propensity of individuals to associate with other individuals who have similar social, cultural, economic and/or demographic characteristics (Milliken and Martins, 1996; McPherson et al., 2001). Studies have shown that firms are more likely to trade with other firms in the same country to reduce transaction costs (Wolf, 2000; Disdier and Head, 2008). In the e-commerce setting, Hortac¸su et al. (2009) found evidence of homophily in transactions on two large online auction sites wherein individuals from the same geographic location were more likely to enter into a trade agreement with one another compared to those from different locations. Recently, two studies of crowd-based online platforms revealed the presence of significant homophily effects. Specifically, in a crowdfunding setting, Lin and Viswanathan (2013) found that investors were more likely to provide funding to borrowers when they were located in the same state. Gefen and Carmel (2008), in their study of Rent A Coder, found that clients were more likely to choose programmers from the same country compared to those from a different country. Consistent with the above findings, we propose that greater similarity between a contestant and a contest holder in terms of economic and cultural factors will improve the contestant’s likelihood of success. Hypothesis 4. Contestants from countries with greater similarity in culture to the country of the contest holder are more likely to be successful in the contest. Hypothesis 5. Contestants from countries with greater similarity in GDPP levels to the country of the contest holder are more likely to be successful in the contest. 3. Research design and analysis 3.1. Data collection To test our hypotheses, we collected data on over one thousand logo-design contests from Logomyway.com, a popular innovation contest setting that matches graphic designers with organizations in need of new logos. The data was collected from the website (after obtaining permission from the website’s owner) using an automated HTML scraping tool. The US market for logo design is estimated to be valued at $3B in 2012 (Source: logoarena.com) and the online logo-design market represents a rapidly growing segment of innovation contests with notable economic implications. For example, Logomyway.com has successfully hosted more than 11,000 contests involving approximately 15,000 participants, with total awarded prize money estimated to be around $4 million since its launch in April 2009 (Source: logomyway.com). Similarly,

192

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

99designs.com has become one of the fastest growing platforms for logo-design contests with approximately 300,000 contests hosted to date involving more than 300,000 participants and total awarded prize money exceeding $75 million (Source: 99designs.com). The contest process on Logomyway.com begins when a contest holder specifies their preferences for a logo in a contest brief, sets the fixed-prize amount (between $200 and $1000), and sets the length of the contest (between one and 30 days). Once the contest has been launched, registered participants on Logomyway.com can openly make design submissions until the contest’s end time. All submissions are displayed publicly on the contest page. Contestants can submit an unlimited number of designs to a contest and feedback is provided by contest holders through rankings and labels during the contest. Contest holders rank submissions in the order of their preference (e.g., #1, #2, #3,. . .) and can re-rank as often or infrequently as desired. For those submissions that are considered promising, contest holders can label them with the phrase “elements we like” or if not promising, “not interested”. Additionally, as a secondary source of feedback, contest holders can post comments on specific designs or respond to contestants’ questions on an open discussion thread on the contest page. Our dataset followed a repeated cross-sectional structure in which we captured data from completed contests with many contestants participating in multiple contests. Data was collected on every contest that occurred on Logomyway.com between August 1, 2010 and February 12, 2011. Additionally, data was collected on each contestant that participated in any of the contests in our dataset. Our unit of analysis is the contest-contestant pair, and in total we have 44,551 observations of such pairs over 1024 contests and 2626 contestants. 3.2. Dependent and independent variables The dependent variable for testing Hypotheses 1–3 is the contestant’s problem-solving effort in a given contest. We measure a contestant’s problem-solving effort in terms of the number of submissions (Number of Submissions) made to a single contest by the contestant. Such a measure is particularly suitable in settings where contest requirements are ill-specified and require considerable experimentation on the part of the contestant in generating multiple submissions. The dependent variable for testing Hypotheses 4 and 5 is a contestant’s success in a given contest (PlacedThisContest). This variable measures whether a contestant had a submission that placed in the top three in a given contest. Although the prize amount is awarded only for the first-place submission (i.e., the winning submission), we chose a measure of placing as a dependent variable over a measure of winning for two reasons. First, the selection of the winning submission from among the top three submissions is highly subjective and idiosyncratic to the contest holder. Second, a measure of winning provides fewer observations of contestant performance in our sample compared to a measure of placing as a dependent variable.2 Our key independent variables were obtained from data sources on country-level economic and cultural factors. We collected GDPP values (adjusted for PPP) for each country represented by a contestant or contest holder in our sample from the CIA World Factbook archives (Source: https://www.cia.gov/library/publications/ the-world-factbook/rankorder/2004rank.html). All GDPP figures were in US$ from 2011, with the exception of three countries whose figures were from 2010. To measure national culture, we obtained data from the GLOBE Study on culture values for uncertainty avoidance and performance orientation for each country represented by a contestant or a contest holder in our sample (House et al.,

2 As a robustness check, we use winning as the dependent variable and observe consistent results.

2004). For each dimension, a country is rated on a scale from 1 to 7, where higher scores represent a higher degree of that cultural dimension in the country. Given that cultural factors have been related to alternative national level indices, including GDPP, in the extant literature (e.g., Sapienza et al., 2006; Tang and Koveos, 2008), we examined the pairwise correlations between the GLOBE culture dimensions used in our study and the GDPP values for the countries represented in our sample. The pairwise correlation between GDPP and GLOBE uncertainty avoidance values is very high at 0.70 (p < 0.01), followed by a moderate correlation of 0.45 (p < 0.01) between GDPP and the performance orientation dimension. Given the lack of discriminant validity between GDPP and GLOBE uncertainty avoidance values, and the significant potential for multicollinearity issues, we use an alternative measure of uncertainty avoidance from Hofstede’s framework (Source: http://geert-hofstede.com/). Our review of the extant literature also indicates that the GLOBE dimension of uncertainty avoidance has been quite controversial and questions about its validity have been raised. For instance, Venaik and Brewer (2010) noted that the GLOBE dimension of uncertainty avoidance has come under considerably criticism for its ambiguity and empirically confounding measures, relative to Hofstede. Our final sample consists of data on economic and cultural factors from 48 countries represented by contestants and contest holders.3 To construct a measure of national culture distance (Culture Distance), we follow prior studies (e.g., Morosoni et al., 1998; Kogut and Singh, 1988) in representing culture distance as the Euclidean distance between the vectors of culture dimensions for the contest holder and contestant. Finally, national wealth distance (GDPP Distance) is measured as the absolute value of the difference between the national GDPP values (adjusted for PPP) of the contestant and contest holder.

3.3. Control variables In examining the role of economic and cultural factors on problem-solving effort (Hypotheses 1–3), we control for several contestant and contest specific factors in the analysis. First, we control for a contestant’s experience and use of the contest platform prior to participation in a given contest by including the following variables: Participation Experience (the number of previous contests a contestant participated in prior to a given contest), Placing Experience (the number of times the contestant placed in the top three in a given contest/total number of prior contests she participated in prior to the contest), and Submissions per Participation (the average number of submissions per contest across prior contests by a contestant). Next, we include control variables relating to a contest’s characteristics such as the contest prize amount in US$ (Prize Amount), the number of contestants participating in a given contest (Number of Contestants), the contest duration in days (Contest Duration), the length of a given contest brief in number of words (Words), the number of views the contest received (Contest Views), the month of the contest (September, October,. . .February), whether the contest required a valid login to be viewed (Private), and the number of concurrent contests with higher prize amounts (Concurrent Contests-Higher Prize). We also control for the feedback received by contestants in the contest environment using variables that capture the number of comments by the contest holder (Contest Holder Comments) and contestants (Contestant Comments) during the contest, and the total number of feedback events

3 The GLOBE study has data for 62 countries while the Hofstede study has data for over 100 countries.

2

= b0 + b1 (GDPPij ) + b2 (Performance Orientationij ) + b3 (Uncertainty Avoidanceij )

+ b5 (GDPPij × Uncertainty Avoidanceij ) + b6 (Controlsij ) + ui + εij

(1)

Table 3 presents the estimated coefficients and standard errors for Eq. (1). Model 1 is a control-variable only model, Model 2 includes the main effects, and Model 3 introduces the interaction effects. We use robust standard errors for all models. For ease of interpretation and comparability, all continuous predictor variables are standardized. We note here that the fixed effects in a negative binomial regression apply to the distribution of the dispersion parameter, not to the predicted values in the model. Therefore, covariates do not drop in this model as with other fixed effects

Table 1 Correlation matrix for variables in Eq. (1).

1

+ b4 (GDPPij × Performance Orientationij )

1.00 −0.11 1.00 −0.16 −0.08 1.00 −0.14 −0.19 −0.09 1.00 −0.17 −0.15 −0.20 −0.10 1.00 −0.19 −0.18 −0.16 -0.21 −0.10 1.00 0.00 −0.02 0.00 0.02 −0.02 −0.01 1.00 −0.04 0.08 −0.03 −0.03 −0.12 0.01 −0.07 1.00 −0.53 0.05 −0.07 0.07 −0.04 −0.04 0.00 −0.04 1.00 0.61 −0.16 −0.02 −0.09 0.12 −0.04 −0.01 0.06 −0.04 1.00 0.47 0.43 −0.06 0.02 −0.07 0.09 −0.04 0.04 0.01 −0.08 1.00 0.18 0.11 0.08 0.07 0.19 −0.03 0.00 −0.04 0.04 −0.04 −0.04 1.00 0.21 0.19 0.10 0.01 0.59 0.07 0.04 −0.07 −0.04 0.02 −0.07 −0.09 1.00 0.05 0.07 0.17 0.16 0.30 −0.13 0.08 0.05 −0.05 −0.04 −0.02 0.00 −0.03 1.00 0.05 0.43 0.24 0.50 0.68 0.63 −0.07 0.05 −0.05 0.07 −0.02 0.02 −0.03 −0.12 1.00 0.82 0.24 0.19 0.09 0.50 0.73 0.75 −0.28 −0.03 −0.04 0.04 −0.06 0.04 0.01 −0.10 1.00 0.04 0.04 0.02 0.01 0.02 0.02 0.03 0.05 −0.03 0.01 −0.01 −0.01 0.00 0.01 0.00 −0.01 1.00 0.36 −0.02 0.00 0.01 0.00 0.02 0.00 0.00 0.01 0.00 0.02 0.00 0.00 −0.01 −0.01 −0.01 −0.01 1.00 0.11 0.14 −0.03 −0.03 0.00 0.01 0.01 −0.02 −0.03 −0.03 0.00 0.00 0.00 0.00 0.02 0.02 0.01 0.00 1.00 −0.05 −0.03 0.09 −0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.01 0.00 0.04 0.02 −0.02 −0.03 −0.01 −0.02 || ≥ 0.01 significant at 0.05 level.

3

log(E(Number of Submissionsij |x))

1.00 0.00 −0.01 0.08 0.00 0.00 0.02 0.00 0.02 0.01 0.00 0.01 0.01 0.00 0.02 0.02 0.00 0.01 −0.02 −0.03 0.00

4

5

6

7

8

9

10

11

12

The econometric analysis in this study is performed using Stata 12. Eq. (1) specifies the econometric model that relates national wealth and national culture values associated with contestant j to her problem-solving effort in innovation contest i. Since the dependent variable, Number of Submissions, is measured as an overdispersed count variable (sample variance = 37.95, sample mean = 4.47), a negative binomial estimator was used, such that E(Number of Submissionsij |x) = ef(x,ˇ) , where x represents the vector of independent variables and ˇ represents the vector of estimated coefficients (thus the log representation in Eq. (1)). Also, note that the composite error term (ui + εij ) includes the contest-specific fixed effect, ui , and the error term, εij . The contest-specific effect was included to control for unobservable differences in contests that could be driving a contestant’s problem-solving effort (e.g., problem specificity, contest holder tastes). The appropriateness of this model is also tested by running a Hausman’s test which rejects the null hypothesis (2 = 1847.15, p < 0.001) that the preferred specification is a random-effects estimator in favor of the alternative hypothesis that the preferred specification is a fixed-effects estimator.

1.00 0.39 0.00 −0.04 0.17 0.03 −0.01 0.02 0.00 0.03 0.02 0.01 0.00 0.01 0.02 0.02 0.00 0.01 −0.01 −0.03 0.00 0.01

13

4.1. National culture, national wealth, and problem-solving effort

Number of submissions 1.00 GDPP 0.01 0.01 Performance orientation Uncertainty avoidance 0.03 Participation experience 0.00 Placing experience 0.17 Submissions per participation 0.36 Number of contestants 0.09 Contest views 0.13 0.04 Private Contest duration 0.05 Contestant comments 0.10 Contest holder comments 0.08 Total feedback events 0.14 0.10 Prize amount Concurrent contests—higher prize 0.00 Words 0.05 September −0.02 October 0.00 November 0.00 December −0.01 January 0.00 February −0.01

14

4. Model specification, analysis and results

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

15

16

17

18

19

20

21

21

23

(solution rankings and labels) provided by the contest holder in a contest (Total Feedback Events). In examining the role of national culture and national wealth distance on likelihood of success in a given contest, we control for a contestant’s prior Participation Experience and Placing Experience as discussed above. In addition, since prior research has shown that submission behavior of contestants can affect their likelihood of success in innovation contests (e.g., Yang et al., 2010; Bockstedt et al., 2014), we include control variables related to such behavior in a given contest. These variables relate to the position of a contestant’s first submission in a given contest (Position of First Submission), the number of submissions by a contestant in a given contest (Number of Submissions), the length of participation of a contestant in a given contest (Length of Participation), and the skewness in terms of submission pattern of a contestant in a given contest (Skewness of Submissions). Table A1 in the Appendix provides a description of each variable and summary statistics. Tables 1 and 2 present correlation matrices for all variables.

193 1.00

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

194

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

Table 2 Correlation matrix for variables in Eq. (2).

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Placed this contest GDPP distance Culture distance Uncertainty avoidance—distance Performance orientation—distance Position of first submission Number of submissions Length of participation Skewness of submissions Participation experience Placing experience

1

2

3

4

5

6

7

8

9

10

1.00 −0.04 −0.03 −0.02 −0.02 −0.07 0.29 0.09 0.00 0.04 0.13

1.00 0.11 −0.06 0.26 −0.02 −0.02 0.00 0.01 0.06 −0.10

1.00 0.86 0.68 0.02 0.02 0.00 0.00 −0.04 −0.11

1.00 0.25 0.02 0.03 −0.01 −0.01 −0.07 −0.11

1.00 0.01 0.00 0.01 0.01 0.00 −0.06

1.00 0.00 0.05 −0.04 −0.02 −0.01

1.00 0.48 0.01 0.00 0.10

1.00 0.11 0.07 0.06

1.00 0.03 0.03

1.00 0.73

11

1.00

|| ≥ 0.01 significant at 0.05 level. Table 3 Regression results for Eq. (1). Dependent variable: Number of submissions Fixed effects—Negative Binomial regression Model 1 Independent variables GDPP Performance orientation Uncertainty avoidance GDPP × Performance orientation GDPP × Uncertainty avoidance

Fixed effects—Poisson regression

Model 2

Model 3

Model 4

−0.023 (0.004)** 0.016 (0.005)** −0.010 (0.004)†

0.001 (0.006) 0.012 (0.005)** −0.021 (0.006)** −0.011 (0.005)* 0.047 (0.009)**

0.016 (0.011) 0.006 (0.008) −0.044 (0.010)** −0.017 (0.009)* 0.074 (0.016)**

Control variables Participation experience Placing experience Submissions per participation Number of contestants Contest views Private Contest duration Contest holder comments Contestant comments Total feedback events Prize amount Concurrent contests—Higher prize ln(Words) September October November December January February Constant

−0.035 (0.004)** 0.042 (0.003)** 0.222 (0.003)** 0.262 (0.021)** −0.197 (0.020)** −0.069 (0.023)** −0.004 (0.015) −0.065 (0.010)** 0.041 (0.011)** −0.172 (0.013)** −0.095 (0.018)** −0.030 (0.016)† −0.074 (0.010)** 0.117 (0.032)** 0.222 (0.033)** 0.030 (0.033) 0.136 (0.038)† 0.056 (0.031)† 0.056 (0.049) 0.688 (0.022)**

−0.026 (0.004)** 0.043 (0.003)** 0.221 (0.003)** 0.269 (0.022)** −0.190 (0.021)** −0.065 (0.024)** −0.012 (0.016) −0.062 (0.010)** 0.042 (0.012)** −0.176 (0.014)** −0.105 (0.019)** −0.021 (0.017)† −0.073 (0.011)** 0.143 (0.034)** 0.232 (0.036)** 0.031 (0.035) 0.156 (0.040)† 0.048 (0.033)† 0.078 (0.052) 0.687 (0.024)**

−0.026 (0.004)** 0.043 (0.003)** 0.223 (0.003)** 0.270 (0.022)** −0.193 (0.021)** −0.067 (0.024)** −0.011 (0.016) −0.062 (0.010)** 0.041 (0.012)** −0.175 (0.014)** −0.105 (0.019)** −0.021 (0.017)† −0.073 (0.011)** 0.144 (0.034)** 0.227 (0.036)** 0.029 (0.035) 0.153 (0.040)** 0.047 (0.033) 0.076 (0.052) 0.692 (0.024)**

−0.023 (0.008)** 0.051 (0.007)** 0.279 (0.009)**

-Log-likelihood 2 Df Number of observations Number of contests VIFmax

104,598.94 9864.25 19 44,551 1024 5.64

93,886.86 9218.24 22 40,149 1024 5.65

93,871.01 9174.46 24 40,149 1024 5.65

122,486.32 40,745.79 8 44,551 1024 5.65

Standard errors included in parentheses. † p < 0.1. * p < 0.05. ** p < 0.01.

models (e.g., fixed effects linear regression or fixed effects Poisson regression).4 In Model 2, we observe significant negative effects of GDPP on problem-solving effort (ˇ = −0.023, p < 0.01), indicating that contestants from countries with lower GDPP tend to make more submissions in innovation contests, which supports Hypothesis 1. The regression coefficients in our negative binomial regression can

4 For more information, see the Stata 12 manual page on xtnbreg (http://www.stata.com/help11.cgi?xtnbreg).

be interpreted as follows: for a one unit change in the predictor variable, the regression coefficient represents the expected change in difference in logs of the number of submissions by contestant j in contest i. More intuitively, the dependent variable is expected to change by 100*(coefficient) percentage for one unit increase in the corresponding independent variable. Since our independent variables are standardized, a one standard deviation increase in GDPP in Model 2 would be expected to result in a decrease of about 2% in problem-solving effort for a contestant. The contests in our sample had on average more than 45 contestants and the contest platform has hosted more than 11,000 contests to date.

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

Therefore, the aggregate effects of national wealth on problemsolving effort within contests and the contest platform would be substantial, significantly affecting the effectiveness of the contest platform as a business model. We also observe a significant positive effect of performance orientation (ˇ = 0.016, p < 0.01) and a marginally significant negative effect of uncertainty avoidance (ˇ = −0.010, p < 0.1). These results provide support to our predictions that, holding all else constant, contestants from countries with higher levels of performance orientation exert more problemsolving effort and those from countries with higher levels of uncertainty avoidance exert less problem-solving effort in innovation contests. In sum, the above results provide support for Hypotheses 2A and 2B. In Model 3, we introduce interaction effects between the national culture dimensions and GDPP. The interaction between GDPP and performance orientation is negative and significant (ˇ = −0.011, p < 0.05) suggesting that performance orientation has a stronger positive association with problem-solving effort in countries with lower levels of GDPP compared to those with higher levels of GDPP. This provides support for Hypothesis 3. Additionally, although not hypothesized, we examined the interaction between GDPP and uncertainty avoidance on problem-solving effort. The results provide support for a positive and significant interaction (ˇ = 0.047, p < 0.01) between GDPP and uncertainty avoidance, suggesting that uncertainty avoidance has a weaker negative association with problem-solving effort in countries with higher levels of GDPP compared to those with lower levels of GDPP. The overarching theme from these results is that the effects of national culture on problem-solving effort become weaker for contestants from countries with higher levels of national wealth. Since Eq. (1) is nonlinear and involves a count measure as a dependent variable, standard interaction plots based on the Aiken and West (1991) approach may result in erroneous interpretation of the interaction term coefficients in Model 3 (Ai and Norton, 2003). Using the margins command in Stata, we verified the marginal effects of the interactions and confirmed that the true effects are in the same direction as the reported coefficients (see Fig. 1). In addition to the above models, we also carried out additional analyses to check for the robustness of our results. First, in Model 4, we present an alternative Poisson estimator for robustness, which provides consistent results to Model 3. Second, we replaced the measure of performance orientation from the GLOBE framework with the measure of masculinity from Hofstede’s framework. This dimension measures the extent to which a society stresses achievement, ambition, and acquisition of wealth, among other factors. The more focused a society is on competition and striving for personal success, the greater is its masculinity score (Hofstede, 2001; Leidner and Kayworth, 2006). The analysis results based on this measure (available upon request) were highly consistent with our main results, which used the GLOBE measure of performance orientation. Third, we included additional national level indices as controls in the analysis to rule out alternative explanations associated with such differences among contestants. These include: (i) human capital and research index—which represents the level and standard of education and research activity in a country (often considered as prime determinants of the innovation capacity of a nation), and (ii) creative outputs index—which represents the extent to which creative outputs are generated as part of a nation’s innovative activity. Data on these indices was collected from the “Global Innovation Index 2012: Stronger Innovation Linkages for Global Growth” study involving collaboration between INSEAD and the World Intellectual Property Organization (WIPO) (Dutta, 2012). The results from the robustness checks (available upon request) were consistent with the main analysis. Finally, to better understand the practical significance of examining national level differences among contestants as predictors

195

Table 4 Partitioning of variance based on hierarchical linear modeling analysis. Random-effects parameters

Estimate of variance

Level 1—contest Level 2—country var(GDPP) var(Performance orientation) var(Uncertainty avoidance) var(Constant) Level 3—contestant Error

0.60 8.64 1.48 2.20 3.01 1.96 30.75 1.98

41.95 Total LR test vs. linear regression: chi2(6) = 604.29

% of total variance 1.42 20.58 3.52 5.23 7.16 4.67 73.30 4.72 100

of problem-solving effort, we carry out a variance partitioning analysis based on the hierarchical linear modeling framework (Rabe-Hesketh and Skrondal, 2008; Mollick, 2012). The data suggests a nesting structure wherein contest-level (Level 1) is the overarching level within which country level (Level 2) differences and subsequently contestant level (Level 3) differences are observed. Based on this structure, the analysis results included in Table 4 reveal some interesting insights. First, we find that between-contestant differences account for a significant proportion of the variance in problem-solving effort (73%). While this result is expected, the magnitude of the effect on problem-solving effort is interesting. What is surprising, however, is how the remaining variance is apportioned across contest and the country levels. Nearly 21% of the variance in problem-solving effort is explained by between-country differences across contestants, and only 1.42% of the variance is explained by between-contest differences. A subsequent partitioning of the variance explained by between-country differences indicates that nearly 16% (out of 21%) is explained by variables representing national wealth and national culture. This finding reinforces the notion that economic and cultural differences across contestants (based on country-level differences) can significantly impact problem-solving effort in innovation contests, above and beyond a focus on contest and contestant level factors. 4.2. National culture distance, national wealth distance, and success We have hypothesized that contestants from countries with national culture and national wealth values similar to the country of the contest holder should be more likely to be successful in a contest, since the output of the contest—a logo design for a company, product or service—represents both a cultural and an economic artifact to be used in the contest holder’s country. As an initial analysis for these hypotheses, we calculated general statistics related to participation and success in the logo-design contests of our data set. Table 5 summarizes these statistics, which show a distinct trend toward “homophily” in the selection of winners at the continent level. The first row in Table 5 shows the number of contests that originated in each continent (i.e., the contest holder’s home continent). The second row provides the total number of contestants in these contests and the third row shows the percentage of contestants in these contests that were from the same continent as the contest holder. Interestingly, the final row shows that across all continents, a larger percentage of contest winners were from the same continent than the percentage of participants, suggesting a homophily effect in winner selection at the continent level. We note, however, that continent is not necessarily the best level for measuring similarities in national culture and national wealth as large variability often exists across countries within a continent. This initial analysis does provide motivation for further analysis using country-level data.

196

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

Fig. 1. Marginal effects plots for the moderating effects of GDPP.

Eq. (2) specifies our econometric model for the hypothesized effects of national wealth distance and national culture distance on the likelihood of success in a contest. Specifically, the likelihood of contestant j placing in the top three finalists of contest i is represented as a function of the national culture distance (Culture Distance) and national wealth distance (GDPP Distance) between the contestant and contest holder. We also include control variables for a contestant’s prior experience and success on the platform, as well as controls for her submission behavior in contest i. As in Eq. (1), we use a composite error term (ui + εij ), which includes the contest-specific fixed effect, ui , and the stochastic error term, εij . We use a fixed-effects logit estimator to estimate the coefficients and standard errors of Eq. (2). Robust standard errors are used for this estimation.

+ b2 (GDPP Distanceij ) + b3 (Controlsj )

uncertainty avoidance had a significant negative effect on the likelihood of success (ˇ = −0.068, p < 0.05). This indicates that an increase in the uncertainty avoidance between a contestant’s and the contest holder’s countries by one standard deviation significantly decreases the odds of a contestant’s success by 6.4% (i.e., 1 − exp−0.068 = 0.064). We do not see any significant relationship between the distance measure for performance orientation and the likelihood of success. Taken together, the above results provide partial support for Hypothesis 5. As a robustness check we estimate Model 4, which includes a binary measure of winning (WonThisContestij ) as the dependent variable. As can be seen, the results in Model 4 are consistent with our prior findings. Finally, we replicated the analysis using Hofstede’s masculinity dimension as a substitute for the GLOBE performance orientation dimension. These results (available upon request) were also consistent with those from our main analysis, highlighting the robustness of our results.

+ ui + εij

5. Discussion

logit(Placed This Contestij ) = b0 + b1 (Culture Distanceij )

(2)

Table 6 presents the results for Eq. (2). Model 1 is a control variable only model. Model 2 includes the measures of national culture distance and national wealth distance between contestant and contest holder. The results from Model 2 provide strong evidence for a significant negative effect of GDPP distance (ˇ = −0.076, p < 0.01) on the likelihood of success, indicating that an increase in the GDPP distance between a contestant and the contest holder by one standard deviation decreases the odds of the contestant’s success by 7.3% (i.e., 1 − exp−0.076 = 0.073). Hypothesis 4 is therefore supported in our analysis. However, we do not see significant results for the effect of culture distance on the likelihood of success. We delve deeper into this finding by carrying out subsequent analysis in Model 3, which uses separate distance measures for each of the culture dimensions. Specifically, we use the absolute difference between contestant’s and contest holder’s countries values for both performance orientation and uncertainty avoidance. The results from this model indicate that culture distance relating to the measure of

5.1. Summary of findings The purpose of this study is to develop and test a theoretical framework of how national economic and cultural diversity among participants influences problem-solving effort and outcomes in online global innovation contests. We capture differences in economic and cultural factors among participants in terms of their national wealth, measured using GDPP of the participant’s country (adjusted for PPP), and national culture, measured using the Hofstede and GLOBE culture frameworks. Our findings lend initial empirical support to the arguments that both national wealth and national culture influence problem-solving effort and outcomes in global innovation contests. First, we find that contestants from countries with lower levels of GDPP are more likely to make a larger number of submissions compared to other contestants. Next, with regards to the culture dimensions, we find that the performance orientation and uncertainty avoidance dimensions of culture have opposing effects on problem-solving effort. That is, higher

Table 5 Homophily at the continent level.

Number of contests Number of contestants Percentage of contestants from same continent as contest holder Percentage of winners from same continent as contest holder

Australia

Asia

Europe

58 2424 2.85% 3.45%

30 1595 61.69% 63.33%

144 6481 24.12% 28.47%

Note: Only 3 contests in the dataset originated in Africa and 46 contests had no final winner announced.

North America 729 33,215 19.07% 32.51%

South America 14 975 6.05% 14.29%

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

197

Table 6 Regression results for Eq. (2). Dependent variable: Placed this contest

Dependent variable: Won this contest

Fixed effects logit model Model 1 Independent variables GDPP distance Culture distance—two dimensions Culture distance—performance orientation Culture distance—uncertainty avoidance

Model 2

Model 3

Model 4

−0.076 (0.028)** 0.035 (0.028)

−0.085 (0.031)**

−0.144 (0.044)** 0.070 (0.044)

−0.003 (0.032) −0.068 (0.039)*

Control variables Participation experience Placing experience Timing of first submission Number of submissions Length of participation Skewness of submissions

0.196 (0.022)** 0.635 (0.026)** 1.038 (0.135)** 0.496 (0.038)** 1.543 (0.133)** −0.091 (0.042)*

0.181 (0.024)** 0.609 (0.027)** 0.992 (0.138)** 0.496 (0.040)** 1.510 (0.138)** −0.088 (0.044)*

0.173 (0.024)** 0.604 (0.027)** 0.999 (0.138)** 0.500 (0.040)** 1.508 (0.138)** −0.087 (0.044)*

0.280 (0.030)** 0.428 (0.037)** 1.223 (0.200)** 0.413 (0.040)** 1.573 (0.160)** −0.123 (0.064)*

-Log-likelihood 2 Df Number of observations Number of contests VIFmax

5144.85 1332.65 6 44,186 1016 1.34

4554.28 1251.50 8 37,750 965 1.35

4553.30 1254.94 9 37,750 965 1.34

2342.10 606.40 8 35,059 888 1.35

Standard errors included in parentheses. † p < 0.1. * p < 0.05. ** p < 0.01.

levels of performance orientation are associated with an increase in problem-solving effort, while higher levels of uncertainty avoidance are associated with a reduction in problem-solving effort. Further, we find that both of these relationships become weaker as GDPP increases. Our results also provide evidence of homophily effects in the context of innovation contests. Specifically, we find that contestants who share greater country-level similarities with the contest holder in terms of GDPP and the culture dimension of uncertainty avoidance are more likely to be successful in a contest compared to other contestants.

5.2. Contributions to theory Taken together, the above findings make three key contributions to the growing literature on innovation contests. First, our study sheds much needed light on the effects of national wealth and cultural differences among participants in innovation contest settings. Despite the acknowledgment in the popular press and research about participant diversity as a defining feature of innovation contests (Brabham, 2008; Daniel et al., 2013), prior empirical studies have often treated participant diversity as a “fixed effect” that needs to be controlled to examine problem solving and outcomes (Bockstedt et al., 2014). As a result, the effects of specific diversity factors in such settings remain largely under researched. In this study, we conceptualize diversity among contestants in terms of two key factors, economic and cultural, and examine its effect on problem-solving effort and contest outcomes. Second, our study highlights the presence of homophily effects in innovation contests. As noted earlier, recent studies on crowdbased settings (e.g., Gefen and Carmel, 2008; Lin and Viswanathan, 2013) have shown that the information asymmetries in such settings can often influence a buyer to contract with a service provider from a similar cultural background. Our study builds upon such findings by indicating that such effects are important predictors of a contestant’s success in an innovation contest, above and beyond their problem-solving skills, prior experience and submission behavior. Specifically, we find that innovation contests are not

a level playing field as has been assumed in the literature. Although the Internet has enabled global communication and participation in online platforms, the world has not necessarily become as “flat” in terms of commerce as suggested by Friedman (2005). This study provides strong evidence that local cultural and economic conditions are still important factors in global transactions for online crowd-based initiatives (Gefen and Carmel, 2008; Terwiesch and Ulrich, 2009). A third contribution of this study is that it demonstrates the use of innovation contest platforms as a rich setting to explore problem solving in global labor markets, particularly those pertaining to online service operations (Metters et al., 2010). In particular, innovation contest settings such as Logomyway.com can potentially be used by OM scholars as microcosms for the global labor market in research on the dynamics of global competition and problem solving.

5.3. Contributions to practice Findings from our study have important practical implications for design changes in innovation contests, which can be made both by contest holders and contest platform owners to extract greater benefits (vis-à-vis greater problem-solving effort) from contestants. Our results suggest that contestants from countries with high levels of performance orientation are more likely to submit more entries. One aspect of performance orientation is the desire to improve from learning (House et al., 2004). Recognizing this aspect of performance orientation, we recommend that contest holders could: (i) explicitly note in the contest description that feedback and responses to specific questions will be provided to contestants in a timely manner, and (ii) provide timely feedback and responses to contestants. Such actions on the part of the contest holder are likely to encourage contestants from countries with higher performance orientation to exert greater problem-solving effort as it enables them to learn more about their submissions through feedback and improve upon them (Bockstedt et al., 2014; Wooten and

198

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

Ulrich, 2012).5 Further, since innovation contests are typically characterized by high levels of evaluation uncertainty, adjustments to the design of contest briefs and problem specifications may be useful. For example, information on local market conditions and provision of timely feedback from the contest holder can potential reduce uncertainty about contest expectations and requirements. This, in turn, can encourage participants from high uncertainty avoidance cultures to exert greater problem-solving effort. We further recommend that, given the increases in the challenges associated with communication, problem interpretation and contextualization for global participants, contest holders should clearly communicate local cultural and economic conditions to contestants. Current innovation contest settings do not provide recommendations along these lines, which limits the contest holder from benefiting from the unique skills of participants from diverse economic and cultural backgrounds. In other words, contest holders may deliberately “tweak” their contest briefs to encourage greater participation and problem-solving effort from contestants. Finally, the finding that contestants who share greater similarity in economic and cultural factors with a contest holder (i.e., homophily effect) are more likely to succeed in a contest suggests that innovation contests may not necessarily provide a “level playing field” for all contestants. Given the taste based nature of creative innovation contests, similarity in cultural and economic factors may allow contestants to better understand contest holder requirements and improve their chances of winning. However, the potential for “home bias” (i.e., the potential for contest holders to deliberately select submissions similar to their own economic and cultural background) may also be a mechanism underlying these results. We recognize that the contests in our sample represent a “fully” unblind setting where the participant identities (as well as their submissions) are visible to the contest holder and all other contestants (Bockstedt et al., 2014). We therefore recommend that information relating to participant identify should be made invisible to contest holders during a contest. If home bias exists, such a modification to the contest setting would eliminate home bias and result in a more level playing field for all contestants. 5.4. Limitations and future research Our study has the following limitations, some of which can serve as avenues for future research. First, our study does not control whether a participant joins the contest platform and becomes a contestant in logo-design contests. This would require data on people in each country that opted not to join the platform. Therefore, the effects estimated represent the “intensive margin”, i.e., they represent the effects of national wealth and national culture on a contestant’s effort and success given that they have already decided to participate in the platform. It is likely that national wealth and

5 Initial evidence in this regard has been found by Wooten and Ulrich (2012) who observe that directed feedback in creative contest settings often resulted in multiple entries from a contestant.

national culture might impact an individual’s decision to participate in an innovation contest in the first place. For example, citizens of countries with lower levels of GDPP are less likely to have access to the Internet. Therefore, when we observe significant effects of national wealth and national culture on a contestant’s effort in an innovation contest, these effects are above and beyond any effects driving initial participation, thus highlighting a strong impact of cultural and economic factors. We also acknowledge that the results reported in this study are conditional on contestants participating in a given contest; therefore, our findings may not be generalizable to the entire sample of contestants on the contest platform. Nonetheless, we did compare summary statistics between participants and non-participants (those that were active on the contest platform, but did not participate in a given contest) on variables representing prior experience (i.e., average prior experience) and success (average prior success on the contest platform). Using a Wilcoxon matched-pairs signed rank test, we found no significant differences between the average populations of non-participants and participants, highlighting the generalizability of the results to the overall contest platform. Another limitation of our study is that we cannot determine the specific causal mechanisms resulting in the success of a given contestant. We show that reducing national culture and national wealth distance results in a higher likelihood of success; while we argue that this success is primarily because contestants who are from countries that are similar in national culture and national wealth produce designs that better match requirements, it is also plausible that contest holder may be biased toward rewarding such contestants, because of the above-mentioned similarities. There is potential to address this issue in future research by conducting a natural experiment in which information about the contestants’ home country is randomly made visible to the contest holder. Currently on Logomyway.com, country information is completely visible to all participants on the platform. If the same effects of culture and wealth can be observed when the contestant’s country information is removed, then this would provide strong evidence that contestants from countries with similar culture and wealth as the contest holder are better at producing designs to meet the contest holder’s requirements. In closing, as global competition continues to increase, research on innovation contests needs to progress. Increased understanding of cultural and economic factors has the potential to generate deeper insights on the effective design of innovation contests and inform our understanding of global labor markets. While our research makes an initial foray into understanding the implications of these important factors, we encourage OM and innovation scholars to continue pursuit of this line of inquiry.

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

199

Appendix.

Table A1 Variable descriptions and summary statistics. Description Dependent variables Effort: Number of submissions Success: Placed this contest Key predictor variables GDPP Performance orientation Uncertainty avoidance GDPP difference Culture distance (two dimensions) Uncertainty avoidance—distance Performance orientation—distance Control variables Participation experience Placing experience Submissions per participation Contest designers Contest views Private Contest duration Contestant comments Contest holder comments Total feedback events Prize amount Concurrent contests—higher prize Words Month (September, October. . .February) Position of first submission Length of participation Skewness of submissions

Mean

Number of logo designs submitted by a contestant in a given contest. Binary variable = 1 if contestant placed in top three of the given contest and 0 otherwise.

S.D.

4.47 0.05

6.16 0.22

GDP per capita adjusted for purchasing power parity of contestant’s home country. Performance orientation value for the contestant’s home country (based on GLOBE framework). Uncertainty avoidance value for the contestant’s home country (based on Hofstede’s framework). The absolute value of the difference between GDPP of contestant’s and contest holder’s countries. Euclidean distance between contestant’s and contest holder’s countries along performance orientation and uncertainty avoidance cultural dimensions. The absolute value of the difference between uncertainty avoidance scores for contestant’s and contest holder’s countries. The absolute value of the difference between performance orientation values of contestant’s and contest holder’s countries.

18,225.55

17,517.02

52.11

12.88

53.88

18.93

27,469.07

16,725.97

21.17

17.76

13.48

17.30

13.53

12.32

Number of contests in which a contestant participated prior to a given contest. Number of contests in which a contestant placed in top three finalists/participation experience prior to a given contest. Average number of submissions per contest for a contestant in contests prior to a given contest. Number of designers (contestants) who participated in a given contest. Number of views a contest page received. Binary variable = 1 if contest required valid login for platform to be viewed and 0 otherwise. Length in days of a given contest. Number of comments posted by contestants in a given contest. Number of comments posted by the contest holder in a given contest. Sum of all feedback events in a given contest (i.e., entries ranked, labeled “elements we like”, or labeled “not interested”). Prize amount in US $ awarded to the winner of a given contest. Number of concurrent contests with a higher prize amount than the given contest. Natural log of the length of the contest brief (specifications) in words. A series of dummy variables indicating the month in which a given contest started (August–February). The position (number) of the first submission by the given contestant in the given contest. The total number of submissions made to a given contest between the first and last entry of a given contestant. The mean/median ratio of the submission position of a given contestant in a given contest.

190.48 0.06

218.66 0.07

4.38

2.61

61.59 1855.82 0.33

39.13 1301.49 0.47

12.01 1.47 0.60 47.78

9.68 2.77 1.16 76.95

315.44 37.23

162.72 36.64

4.85 –

0.67 –

121.90

172.90

44.41

125.82

1.04

0.60

Note: Unit of analysis was the contest–contestant pair; therefore the reported summary statistics for contest-level variables are weighted by the number of observations per contest.

References 99designs.com (accessed November 2013). Ai, C., Norton, E.C., 2003. Interaction terms in logit and probit models. Econ. Lett. 80 (1), 123–129. Aiken, L., West, S., 1991. Multiple Regression: Testing and Interpreting Interactions. Sage Publications, London. Alary, D., Gollier, C., Treich, N., 2013. The effect of ambiguity aversion on insurance and self-protection. Econ. J. 123, 1188–1202. Ancona, D.G., Caldwell, D.F., 1992. Demography and design: Predictors of new product team performance. Organ. Sci. 3 (3), 321–341. Argote, L., 1999. Organizational Learning: Creating, Retaining and Transferring Knowledge. Kluwer, Norwell, MA. Bang, H.-K., Raymond, M.A., Taylor, C.R., Moon, Y.S., 2005. A comparison of service quality dimensions conveyed in advertisements for service providers in the USA and Korea: A content analysis. Int. Market. Rev. 22 (3), 309–326. Bayus, B.L., 2013. Crowdsourcing new product ideas over time: An analysis of Dell Ideastorm community. Manag. Sci. 59 (1), 226–244. Belsky, S., 2010. Crowdsourcing is broken: how to fix it. Bus. Week Online, http://www.businessweek.com/innovate/content/jan2010/id20100122 047502.htm

Billington, C., Davidson, R., 2013. Leveraging open innovation using intermediary networks. Prod. Oper. Manag. 22 (6), 1464–1477. Bockstedt, J., Druehl, C., Mishra, A., 2014. Implications of Heterogeneous Submission Behavior in Unblind Online Crowdsourcing Contests. Working Paper. Boudreau, K.J., Lacetera, N., Lakhani, K.R., 2011. Incentives and problem uncertainty in innovation contests: An empirical analysis. Manag. Sci. 57 (5), 843–863. Brabham, D.C., 2008. Crowdsourcing as a model for problem solving: an introduction and cases. Converg.: Int. J. Res. New Media Technol. 14 (1), 75–90. Calantone, R., Garcia, R., Droge, C., 2003. The effects of environmental turbulence on new product development strategy planning. J. Prod. Innov. Manag. 20 (2), 90–103. Che, Y.-K., Gale, I., 2003. Optimal design of research contests. Am. Econ. Rev. 93 (3), 646–671. Chesbrough, H.W., 2007. Why companies should have open business models. MIT Sloan Manag. Rev. 48 (2), 22–28. Cheung, M.-S., Myers, M.B., Mentzer, J.T., 2010. Does relationship learning lead to relationship value? A cross-national supply chain investigation. J. Oper. Manag. 28, 472–487. CIA, The World Factbook. https://www.cia.gov/library/publications/the-worldfactbook/ (accessed 12.06.13).

200

J. Bockstedt et al. / Journal of Operations Management 36 (2015) 187–200

Dahan, E., Mendelson, H., 2001. An extreme value model of concept testing. Manag. Sci. 47 (1), 102–116. Daniel, S., Agarwal, R., Stewart, K.J., 2013. The effects of diversity in global, distributed collectives: a study of open source project success. Inf. Syst. Res. 24 (2), 312–333. Deci, E., 1971. The effects of externally mediated rewards on intrinsic motivation. J. Pers. Soc. Psychol. 18 (1), 105–115. Disdier, A.-C., Head, K., 2008. The puzzling persistence of the distance effect on bilateral trade. Rev. Econ. Stat. 90 (1), 37–48. Dutta, S., 2012. The Global Innovation Index 2012: Stronger Innovation Linkages for Global Growth. INSEAD. Erat, S., Krishnan, V., 2012. Managing delegated search over design spaces. Manag. Sci. 58 (3), 606–623. Etzkowitz, H.C., Kemlgor, C., Uzzi, B., 2000. Athena Unbound: The Advancement of Women in Science and Technology. Cambridge University Press, Cambridge, UK. Friedman, T., 2005. The World is Flat: A Brief History of the 21st Century. Farrar, Straus, and Giroux, NY. Froot, K.A., Rogoff, K., 1995. Perspectives on PPP and Long-run Real Exchange Rates. NBER Working Paper No. 4952, http://scholar.harvard.edu/files/rogoff/files/ perspectives on ppp and long-run real exchange rates.pdf Fullerton, R.L., McAfee, R.P., 1999. Auctioning entry into tournaments. Am. Econ. Rev. 107 (3), 573–605. Gefen, D., Carmel, E., 2008. Is the world really flat? A look at offshoring at an online programming marketplace. MIS Q. 32 (2), 367–384. Girotra, K., Terwiesch, C., Ulrich, K.T., 2010. Idea generation and the quality of the best idea. Manag. Sci. 56 (4), 591–605. Graham, J.R., Harvey, C.R., Huang, H., 2009. Investor competence, trading frequency, and home bias. Manag. Sci. 55 (7), 1094–1106. Gray, J.V., Massimino, B., 2014. The effect of language differences and national culture on operational process compliance. Prod. Oper. Manag. 23 (6), 1042–1056. Ha, A.Y., Porteus, E.L., 1995. Optimal timing of reviews in concurrent design for manufacturability. Manag. Sci. 41 (9), 1431–1447. Hahn, E.D., Bunyaratavej, K., 2010. Services cultural alignment in offshoring: The impact of cultural dimensions on offshoring location choices. J. Oper. Manag. 28 (3), 186–193. Hall, E.T., Hall, M.R., 1990. Understanding Cultural Differences. Intercultural Press, Yarmouth, ME. Handley, S.M., Benton Jr., W.C., 2013. The influence of task- and location-specific complexity on the control and coordination costs in global outsourcing relationships. J. Oper. Manag. 31, 109–128. Harrison, D.A., Klein, K.J., 2007. What’s the difference? Diversity constructs as separation, variety, or disparity in organizations. Acad. Manag. Rev. 32, 1199–1228. Harrison, G.W., Lau, M.I., Rutström, E.E., 2009. Risk attitudes, randomization to treatment, and self-selection into experiments. J. Econ. Behav. Organ. 70 (3), 498–507. Hofstede, G., 1980. Culture’s Consequences: International Differences in Work Related Values. Sage Publications, Newbury Park, CA, http://geert-hofstede.com/ (accessed October 2013). Hofstede, G., 2001. Culture’s Consequences: Comparing Values, Behaviors, Institutions and Organizations across Nations, second ed. Sage Publications, Thousand Oaks, CA. Hortac¸su, A., Asís Martínez-Jerez, F., Douglas, J., 2009. The geography of trade in online transactions: evidence from eBay and MercadoLibre. Am. Econ. J.: Microecon. 1 (1), 53–74. House, R.J., Hanges, P.J., Javidan, M., Dorfman, P.W., Gupta, V. (Eds.), 2004. Culture, Leadership, and Organizations: The GLOBE Study of 62 Societies. Sage Publications, Thousand Oaks, CA. Howe, J., 2006. The rise of crowdsourcing. Wired Mag. 14 (6), 1–4. Huang, R.J., 2012. Ambiguity aversion, higher-order risk attitude and optimal effort. Insur.: Math. Econ. 50, 338–345. Innocentive.com (accessed October 2013). Jeppesen, L.B., Lakhani, K.R., 2010. Marginality and problem-solving effectiveness in broadcast search. Organ. Sci. 21 (5), 1016–1033. Kogut, B., Singh, H., 1988. The effect of national culture on the choice of entry mode. J. Int. Bus. Stud. 19 (3), 411–432. Kreps, D.M., 1997. Intrinsic motivation and extrinsic incentives. Am. Econ. Rev. 87 (2), 359–364. Kull, T.J., Wacker, J.G., 2010. Quality management effectiveness in Asia: The influence of culture. J. Oper. Manag. 28, 223–239. Ladbury, J.L., Hinsz, V.B., 2009. Uncertainty avoidance influences choices for potential gains but not losses. Curr. Psychol. 28 (3), 187–193. Lakhani, K.R., Jeppesen, L.B., Lohse, P.A., Panetta, J.A., 2007. The Value of Openness in Scientific Problem Solving. Harvard Business School Working Paper No. 07-050, http://www.hbs.edu/research/pdf/07-050.pdf Lin, M., Viswanathan, S., 2013. Home Bias in Online Investments: An Empirical Study of an Online Crowd Funding Market. SSRN Working Paper, http://papers.ssrn.com/sol3/papers.cfm?abstract id=2219546 Lau, D.C., Murnighan, J.K., 1998. Demographic diversity and faultlines: The compositional dynamics of organizational groups. Acad. Manag. Rev. 23 (2), 325–340. Lazear, E.P., Rosen, S., 1981. Rank-order tournaments as optimum labor contracts. J. Polit. Econ. 89 (5), 841–864. Lee, M., 1976. Purchasing Power Parity. Marcel Dekker, New York.

Leidner, D.E., Kayworth, T., 2006. A review of culture in information systems research: toward a theory of information technology culture conflict. MIS Q. 30 (2), 357–399. Logoarena.com, May 2011. Fun Facts About Logo Design Market in USA, http://www.logoarena.com/blog/fun-facts-about-logo-design-market-in-usa/ Logomyway.com (accessed November 2013). McPherson, M., Smith-Lovin, L., Cook, J.M., 2001. Birds of a feather: Homophily in social networks. Annu. Rev. Sociol. 27, 415–444. Metters, R., Zhao, X., Bendoly, E., Jiang, B., 2010. “The way that can be told of is not an unvarying way”: Cultural impacts on operations management in Asia. J. Oper. Manag. 28 (3), 177–185. Milliken, F.J., Martins, L.L., 1996. Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Acad. Manag. Rev. 21 (2), 402–433. Milner, L.M., Collins, J.M., 2000. Sex-role portrayals and the gender of nations. J. Advert. 29 (1), 67–79. Moldovanu, B., Sela, A., 2006. Contest architecture. J. Econ. Theory 126 (1), 70–96. Mollick, E., 2012. People and process, suits and innovators: The role of individuals in firm performance. Strateg. Manag. J. 33, 1001–1015. Morosoni, P., Shane, S., Singh, H., 1998. National cultural distance and cross-border acquisition performance. J. Int. Bus. Stud. 29 (1), 137–158. Nakata, C., Sivakumar, K., 1996. National culture and new product development: An integrative review. J. Market. 60 (1), 61–72. Nalebuff, B.J., Stiglitz, J.E., 1983. Prizes and incentives: towards a general theory of compensation and competition. Bell J. Econ. 14 (1), 21–43. Naor, M., Linderman, K., Schroeder, R., 2010. The globalization of operations in Eastern and Western countries: Unpacking the relationship between national and organizational culture and its impact on manufacturing performance. J. Oper. Manag. 28, 194–205. Narayanan, S., Swaminathan, J.M., Talluri, S., 2014. Knowledge diversity, turnover, and organizational-unit productivity: An empirical analysis in a knowledgeintensive context. Prod. Oper. Manag. 23 (8), 1332–1351. Peters, L., Karren, R., 2009. An examination of the roles of trust and functional diversity on virtual team performance ratings. Group Organ. Manag. 34 (4), 479–504. Rabe-Hesketh, S., Skrondal, A., 2008. Multilevel and Longitudinal Modeling Using Stata, second ed. Stata Press, College Station, TX. Rouziès, D., Segalla, M., Besson, M., 1999. The Impact of Cultural Dimensions on Sales Force Compensation. Working Paper No. 670. HEC Business School, www.hec.edu/heccontent/download/5001/133466/version/file/CR670.pdf Sapienza, P., Zingales, L., Guiso, L., 2006. Does Culture Affect Economic Outcomes? (No. w11999). National Bureau of Economic Research, Chicago. Sarin, S., McDermott, C., 2003. The effect of team leader characteristics on learning, knowledge application, and performance of cross-functional new product development teams. Decis. Sci. 34 (4), 707–739. Siegel, J.I., Licht, A.N., Schwartz, S.H., 2013. Egalitarianism, culture distance, and foreign direct investment: A new approach. Organ. Sci. 24 (4), 1174–1194. Shane, S., 1993. Cultural influences on national rates of innovation. J. Bus. Ventur. 8, 59–73. Simonton, D., 2003. Scientific creativity as a constrained stochastic behavior: The integration of product, person and process perspectives. Psychol. Bull. 129 (4), 475–494. Smith, P., Trompenaars, F., Dugan, S., 1995. The Rotter locus of control scale in 43 countries: A test of cultural relativity. Int. J. Psychol. 30 (3), 377–400. Tang, L., Koveos, P.E., 2008. A framework to update Hofstede’s cultural value indices: Economic dynamics and institutional stability. J. Int. Bus. Stud. 39 (6), 1045–1063. Taylor, C.R., 1995. Digging for golden carrots: an analysis of research tournaments. Am. Econ. Rev. 85 (4), 872–890. Terwiesch, C., Xu, Y., 2008. Crowdsourcing contests, open innovation, and multiagent problem solving. Manag. Sci. 54 (9), 1529–1543. Terwiesch, C., Ulrich, K.T., 2009. Innovation Tournaments. Harvard Business Press, Boston, MA. TopCoder.com (accessed November 2013). Venaik, S., Brewer, P., 2010. Avoiding uncertainty in Hofstede and GLOBE. J. Int. Bus. Stud. 41, 1294–1315. von Hippel, E., 2005. Democratizing Innovation. MIT Press, Cambridge, MA. Wagner, E.B., 2011. Why prize? The surprising resurgence of prizes to stimulate innovation. Res. Technol. Manag. 54 (6), 32–36. Williamson, C., Mathers, R., 2011. Economic freedom, culture, and growth. Public Choice 148, 313–335. Wolf, H.C., 2000. Intranational home bias in trade. Rev. Econ. Stat. 82 (4), 555–563. Wooten, J.O., Ulrich, K.T., 2012. Idea Generation and the Role of Feedback: Evidence from Field Experiments with Innovation Tournaments. SSRN Working Paper, http://ssrn.com/abstract=1838733 Yang, Y., Chen, P., Pavlou, P., 2010. Managing Open Innovation Contests in Online Market. Working Paper. Temple University. Zhu, K., Kraemer, K.L., 2005. Post-adoption variations in usage and value of ebusiness by organizations: Cross-country evidence from the retail industry. Inf. Syst. Res. 16 (1), 61–84.