1
IFD reconceptualized: a study on the effects of IFD and its characteristics on boundary spanning behaviours in open source teams
Master thesis, Msc HRM
University of Groningen, Faculty of Economics and Business
Peter W. Vos Student number: 2515695
Friesestraatweg 16 C, 9718 NG, Groningen Email: peterwvos@gmail.com
Supervisor: Thom de Vries
2 ABSTRACT
For a team to perform well, it is of critical importance that the team members work across the team boundaries to communicate with other teams and external experts. This process is known as boundary spanning. Intrapersonal Functional Diversity (IFD) is an important construct when considering an individual's effectiveness in boundary spanning. IFD refers to the number of functional domains an individual has worked in during their career. Research has been done on the effects of Intrapersonal Functional Diversity (IFD) on boundary spanning behaviours. This study expands upon that research and reconceptualizes IFD by using the characteristics of IFD as moderators of the relationship between IFD and boundary spanning. The moderators used were the distribution of IFD (did an individual spend equal time working in their functional domains or is the distribution skewed?) and the IFD recency (how long ago did the individual work in another functional domain?). The study was conducted using data gathered from open source software teams active in the cryptocurrency industry. I expected a more skewed distribution to have a positive effect on the relationship between IFD and boundary spanning, and I expected the same positive effect on that relationship with a more recent IFD recency. The data showed significant support for the positive effect of the moderator IFD recency on the relationship between IFD and
boundary spanning. Future research can use these findings to conceptualize IFD in a more accurate way.
3 INTRODUCTION
The performance of a team is not only dependent on the processes that happen within the team, but also on the extent that team members create and manage relationships and networks outside of their team (Gladstein, 1984). This process is known as boundary spanning and is defined as “a team’s efforts to establish and manage external linkages that can occur within an organization or across organizational boundaries” (Marrone, 2010). Boundary spanning activities that an individual or team can engage in include representing the team to stakeholders, coordinating tasks with other teams in the organization and consulting with outside experts (Ancona & Caldwell, 1992). Boundary spanning in
organizations has received much attention for the past two decades (Marrone, 2010; Ancona & Caldwell, 1992; Marks et al, 2005). Gladstein (1984) was one of the first to introduce the concept of boundary spanning as an important predictor of team and organizational
performance. Since then, research found that not only the amount but also the type of boundary spanning is important and that specifically in multi-team systems (defined by Mathieu, Marks & Zacarro (2001) as “two or more teams that interface directly and
interdependently in response to environmental contingencies toward the accomplishment of collective goals”), coordination between teams is of critical importance. (Ancona & Caldwell, 1992; Marks et al., 2005) In other words, boundary spanning is an important determinator of team and organizational performance.
4
(Bunderson, 2003). In other words, IFD looks at how many different fields an individual has worked in. For example, someone who has worked in marketing, accounting, recruiting and academia throughout their career has a higher IFD than someone who has only worked in marketing for their entire career. A high IFD can have positive effects on the amount and effectiveness of boundary spanning through the expertise effect (combinations of functional knowledge) and the similarity effect (individuals are able to find common ground with others through shared functional experiences)(Bunderson, 2003). Other research, however, suggests that people with a broad functional experience are a “jack off all trades, master of none” (Buyl, Boone, Hendriks, & Matthysens, 2011) and that the limited amount of deep specific knowledge hampers the effectiveness of engaging in complex communication processes between teams. De Vries et al. (2014) looked at the effects of IFD on individual boundary spanning and report both positive and negative effects, depending on moderating factors.
While researchers agree that IFD is an important antecedent of boundary spanning, the literature remains inconclusive about the direction and nature of the relationship between these factors. De Vries et al. (2014) considered the moderating factor organizational
5
understand the mechanisms of the relationship between IFD and boundary spanning, it is important to understand what parts of IFD are important in that relationship.
To solve this deficiency in the field, I will be building on the research of De Vries et al. (2014), answering their call for future research into moderating factors of the relationship between IFD and boundary spanning. Furthermore, I will build upon past research into IFD (Bunderson, 2003) and examine the different aspects that make up the popular
conceptualizations of IFD and how they interact with boundary spanning. I will be
conducting research using two characteristics of IFD as moderators: the distribution of IFD and the recency of IFD. These moderators were chosen because they are an important part of IFD, but are rarely measured separately. By using these moderators, researchers can better understand under what conditions IFD has a positive effect on boundary spanning. These moderators will be tested within a moderated-mediation model that includes team
6 Figure 1. conceptual model
The practical contribution of this research is that it will help managers to know the optimal conditions for boundary spanning. It will shed light on how the advantage of high IFD in relation to boundary spanning can erode depending on the characteristics of the IFD. Possible effects are that managers who desire high levels of boundary spanning in their teams or organization will need to achieve a high level of turnover (internal or external). It can also show the benefits of teams in matrix organizations, where the composition of teams is constantly changing.
The theoretical contribution of this research is that it will investigate the moderating factors of the relationship between IFD and boundary spanning, and how the characteristics of IFD can determine the effect of IFD on team boundary spanning. A better understanding of the characteristics of IFD will aid future research efforts into the area by allowing researchers to conceptualize IFD on a more detailed level and possibly changing the definition of IFD to include its characteristics. With these insights, future models on team effectiveness that include IFD can be more refined and advanced.
7 THEORY AND HYPOTHESIS DEVELOPMENT
Boundary spanning refers to a team’s efforts to establish and manage external
linkages that can occur within an organization or across organizational boundaries (Marrone, 2010). Ever since boundary spanning and its effects on team performance have gotten attention in the literature, researchers have agreed that the amount and nature of boundary spanning activities being conducted in a team is a critical determinant of team performance (Gladstein 1984; Ancona & Caldwell, 1992; Choi, 2002; Marks et al, 2005). Typical activities a team can undertake to engage in boundary spanning behaviour are mapping (creating a picture of the outside environment), coordinating tasks and exchanging technical knowledge with other teams in the organization, influencing the external environment and consulting with outside experts (Ancona & Caldwell 1992). Marrone, Tesluk & Carson (2007) emphasize that to classify as boundary spanning behaviour, activities should assist the team in meeting its overall objectives. This emphasis is important because it separates
activities that do not aim to work towards the team’s objectives from those that do, and only classifies the latter as boundary spanning activities. To effectively engage in boundary spanning activities, team members must be knowledgeable about the workings of multiple teams and adjust themselves to both the internal and external environment (DeChurch & Marks, 2006).
IFD and boundary spanning
8
the nine functional domains individuals have worked in, but also the amount of time they have worked in each domain. However, because this research looks at the different characteristics of IFD, including the distribution of IFD, I will be using different parts of Bunderson’s definition for the different constructs in the conceptual model. Thus, for the independent variable IFD, I will only look at the number of functional domains in an
individual’s work history. Using this conceptualization, someone who has only worked in one functional domain (human resources for example) has the lowest possible IFD. An individual who has worked in multiple domains throughout their career (for example: started out in marketing, then worked in recruitment, then switched to research and development and now works as general manager) has a high IFD.
Effectively engaging in boundary spanning activities requires team members to have knowledge of the workings of multiple teams and environments (DeChurch & Marks, 2006). This is because in order to effectively communicate in a team or environment, individuals will need to be able to understand the jargon and communication norms of that team or environment. For example, someone who works in a marketing team and needs to
communicate with web development team to implement a new part of the website will have an easier time doing so if he understands the technical terms and the limitations of the web development team. This also goes the other way around: if the members of the web
development team can explain their process in a way that people from the marketing team will be able to easily understand it, communication will be more efficient and effective.
Individuals with a high IFD have been active in many functional environments and teams operating in those environments. Due to this, they will have been exposed to
9
high IFD individuals are more likely to possess the necessary knowledge of the workings of teams and environments beyond the boundary of their own team in order to effectively engage in boundary spanning. Summarized, high IFD individuals will have a greater capability to effectively engage in boundary spanning.
Individuals with greater capability to perform an activity are more likely to engage in those activities. This is because when individuals know that they are good at something, they are less afraid of failing and are therefore more likely to engage in those activities (Bandura, 1997). The same effect is true for boundary spanning activities, theorists propose that team members with greater boundary spanning capabilities are more likely to engage in boundary spanning behaviours (Marrone, Tesluk, & Carson, 2007; Bandura, 1997). Teams whose members have a higher IFD will be more effective in boundary spanning activities. Because they are more effective in it, they will engage in more boundary spanning behaviours than people with lower IFD. This relationship is summarized in hypothesis 1a.
Hypothesis 1a: Intrapersonal functional diversity has a positive relationship with the amount of boundary spanning a team engages in.
IFD recency as a moderator
10
experiences those work activities and positions, and the order in which they developed their skills. Johns (1996) identified key elements of careers and added that a career involves moving on a path over time. Combining the career path sequence with IFD, an individual can have a very functionally diverse start of their career, but they might have worked in the same field for the rest of their career. Alternatively, an individual can have a functionally diverse career for the entire duration of their career. An individual that has developed a diverse skill set early in their career but has not developed skills relevant to new functional domains for a long time might have different boundary spanning capabilities than an individual who has continued gaining experience in different functional domains. Therefore, it is important to pay attention to the career path sequence when considering the effect of IFD on individual boundary spanning. In this research, I conceptualize this moderating factor as “IFD recency”. It refers to how long ago an individual has worked in a domain other than their current domain, relative to the length of their career.
For an individual to be able to retain and later retrieve knowledge they acquired, it is important that they retrieve and use this knowledge periodically (Richardson-Klavehn, Bjork, 1988; Bjork, 1988). Therefore, an individual with a functionally diverse career, but a low IFD recency (meaning that it has been a long time since they worked in another domain), might have lost the ability to retrieve his acquired knowledge of the functional domains he has worked in. This can occur because he has not been active in those functional domains for a long time and therefore has not used and accessed the acquired knowledge periodically, thus creating difficulty in retrieving it. In other words, the knowledge created by being active in multiple functional domains (a high IFD) can be lost if high IFD individuals have not worked in another functional domain since early in their career.
11
When individuals lose the ability to effectively access the knowledge gained in other functional domains, the boundary spanning capabilities of an individual will decrease. This will, in turn, decrease the amount of boundary spanning the individual engages in (see: DeChurch & Marks, 2006). Summarized, when individuals have not gained new experiences and knowledge to broaden their IFD in a long time, the positive relationship between IFD and individual boundary spanning will be less positive or even neutral.
When an individual has a high IFD recency (they have kept acquiring functionally diverse experiences and knowledge throughout their entire career), they will be able to retrieve this knowledge better, because it was acquired more recently (Bjork, 1988). When engaging in boundary spanning activities, and thus communicating in other teams and environments, they can use this knowledge to facilitate better boundary spanning outcomes and be more effective in boundary spanning. Following that, individuals who have greater boundary spanning capabilities engage in more boundary spanning behaviours (see:
DeChurch & Marks, 2006). Therefore, for teams whose members have a high IFD recency, IFD will have a greater positive relationship with boundary spanning. This reasoning is summarized in hypothesis 1b.
Hypothesis 1b: The positive relationship between intrapersonal functional diversity and team boundary spanning is moderated by IFD recency, in a way that higher IFD recency leads to a more positive relationship.
Distribution of IFD as a moderator
12
divided among the different functional experiences they have, but the distribution can also be skewed; the individual has spent a small part of their career in diverse functional
environments and the majority of their career in the same functional environment. I will refer to individuals with a skewed distribution as “specialists” and individuals with an even
distribution as “generalists”. It is important to include IFD distribution as a moderating factor, because it considers one of the potential pitfalls of IFD; that the individual may become too much of a generalist, a “jack of all trades, but master of none”, having limited expertise in many functional areas, but no deep expertise in any of them (Buyl, Boone, Henriks, & Matthysens, 2011).
In order to develop thorough knowledge on a topic and be able to retrieve it easily, it is important that enough time is spent developing that knowledge and that it is retrieved periodically (Bjork, 1988). A generalist may experience that they have not spent enough time in any given field to develop extensive domain-specific knowledge and to engage in complex tasks in that domain (De Vries et al., 2016). One of the mechanisms of the relationship between IFD and boundary spanning is “expert power” where people use their unique knowledge to influence others (French & Raven, 1959; Bunderson, 2003). For a generalist, this unique knowledge might not be sufficiently developed to use their expert power and therefore, they will be less effective in their boundary spanning activities. Thus, when the moderating factor IFD distribution is such that the individual’s functional experiences are evenly distributed across their career, IFD will have a less positive relationship with boundary spanning.
13
not be true for individuals whose IFD distribution is uneven (and the individual is thus a specialist). The individual then has the diverse functional experience to effectively
communicate in multiple functional environments and use the referent power base but also has enough functional experience in one field to have a thorough knowledge of it and use the expert power base. In other words, the specialist can use both power bases to aid their
boundary spanning efforts, whereas the generalist will have less deep knowledge required to use the expert power base. Thus, for individuals who are specialists, IFD will have a more positive relationship with boundary spanning than for individuals who are generalists. The moderating effect of IFD distribution on boundary spanning is summarized in hypothesis 1c.
Hypothesis 1c: The positive relationship between IFD and team boundary spanning is moderated by the IFD distribution, in a way that the more a team’s IFD distribution is specialistic, the more positive the relationship is.
IFD and team performance
Boundary spanning has been identified as an important predictor of team performance (Gladstein, 1984; Ancona & Caldwell, 1992). Teams that engage in more and in more
effective boundary spanning have greater access to resources and knowledge than teams who do not (Bunderson, 2003). These teams can then use these extra resources and knowledge to enhance their performance relative to teams that do not have access to these extra resources and knowledge. Building on hypothesis 1a, in which I propose that IFD has a positive relationship with team boundary spanning, I propose that a higher IFD leads to an increased team performance through team boundary spanning. This relationship is summarized in hypothesis 2.
14
Hypothesis 2: IFD has an indirect positive relationship with team performance through boundary spanning. The relationship between IFD and boundary spanning is moderated by IFD recency and IFD distribution.
15 METHODS
Sample and procedure
This research was conducted using data from 200 teams who all created and managed their own cryptocurrency and blockchain technology. Cryptocurrencies are digital currencies based on blockchains and have existed since around 2008, with Bitcoin being the first and still the most famous of them. Summarized in a few words, a cryptocurrency is a
decentralized digital cash system (Narayanan, Bonneau, Felten, Miller, & Goldfeder, 2016). Narayanan et al. explain that a cryptocurrency is essentially a network of peers, who all have a complete history of all transactions and records of the amounts of currency all accounts have. Once someone makes a transaction, this transaction needs to be confirmed by every peer in the network and it is then confirmed. After it is confirmed, it cannot be
reversed. The list that all peers have access to and that is continually updated with new transactions is called a blockchain. When a transaction has been confirmed by every peer on the network it becomes part of the blockchain. The peers who confirm these transactions are called miners, and for their efforts, they are rewarded with tokens of the currency they are validating. In order to validate a transaction, a miner has to solve a cryptologic puzzle with their computer. This ensures that no one can engage in fraudulent transactions, since they would need to solve all the cryptologic puzzles of all the peers in order for their transaction to become part of the blockchain. This is also where the “crypto” in cryptocurrencies comes from.
16
network has a list of transaction and who owns which amount of digital currency. Summarizing, cryptocurrencies building on blockchain technology can be described as “limited entries in a database no one can change without fulfilling specific conditions”.
Cryptocurrencies come in two forms: coins and tokens. A coin is a cryptocurrency which has their own blockchain. Bitcoin is the most famous example, but Ethereum and Litecoin are other examples. Tokens build on the blockchain of a coin. For example, a token can have Ethereum as their platform, which means that every time there is a transaction with this token, this is done through the Ethereum blockchain. This token can then be bought or sold with Ethereum, and is also dependent on the exchange rate of Ethereum to other currencies.
For this research, I considered both tokens and coins. Every token and coin has its own team behind it, and has their own exchange rate. I randomly selected 200
cryptocurrencies from an online database1 of 1578 cryptocurrencies. For those 200
currencies, I gathered performance data over a period of three months (January 1st, 2018 until April 1st, 2018). Apart from the variables from the conceptual model and hypothesis, the control variables gender, career length, the age of the team and team size were measured and added to the dataset. All measures, except team size, team performance and team boundary spanning were first measured on the individual level and then averaged to gain team-level data. If a team had more than 8 members, only data from 8 randomly selected team members was used to calculate a team average. For all IFD related measures, the Linkedin pages from the team members were used. These were found on the websites of their respective
cryptocurrencies. For all the random selections, a random number generator was used. Currencies that were selected, but did not have sufficient data on their performance or team
17
members were excluded. Ultimately, 90 out of the 200 selected cryptocurrencies had enough data available and were used in the analysis.
For these 90 teams, the team size ranged from 1 to 50 and the average team size was 13.73. The standard deviation was 10.15. Gender was measured for 0 = female, 1 = male. The highest score was 1 (meaning the team was all male) and the lowest score was 0.50. The average gender was 0.88 and the standard deviation for gender was 0.12. The highest average career length in a team was 304 months and the lowest one was 30 months. The average career length was 144.28 months and the standard deviation was 55.84. Months. The age of the team was measured by how long ago its cryptocurrency was first available. The average was 11.40 months, the standard deviation was 13.38 months, the lowest was 1 months and the highest was 57 months.
Measures Performance
Performance was measured over a three month period (January 1st, 2018 until April 1st, 2018) by looking at the change in market share from a cryptocurrency during that time period. Using the performance data from the online database2, the daily value for a currency was multiplied with the number of coins or tokens in circulation, and then divided by the total value of all cryptocurrencies combined. This market share was then averaged for every week, so that thirteen data points were created, each of the data points being the average market share of a coin in one week. Using these data points, a performance score was calculated during the analysis using linear latent growth curve modeling. This approach was chosen as opposed to taking the market share at a singular point in time. This was done to prevent smaller and younger teams that are rapidly growing from scoring very low on the
18
performance measure compared to large teams that see their market share shrinking. Performance was measured using market share because market share is a key factor for profitability (Buzzell, Gate, Sultan, 1975) and thus performance. In addition, market share data is objective and well measurable in the dataset I worked with.
Boundary Spanning
The boundary spanning of the team behind a cryptocurrency was measured by counting number of strategic partners the team listed on their site or in their currency
documentation. This measure was chosen because creating and maintaining partnerships is a form of boundary spanning (Williams, 2002; Piercy, 2009), and this partnership data was available in the database I worked with. A team’s boundary spanning score is equal to their number of strategic partners. If the strategic partners were not mentioned on the website, I looked in the whitepaper and tried to find other online sources that mentioned partnerships. If nothing was found, I assumed that the team had no strategic partnerships, and therefore they would receive a boundary spanning score of 0.
IFD
In most research considering IFD, IFD is measured using a formula that accounts for both the distribution and the number of different functional domains a person has worked in (Bunderson, 2002; de Vries et al., 2014). Because this research considers the distribution of IFD and IFD itself as separate moderators, IFD was measured by simply counting the number of functional domains an individual has been active in during their career. To identify
different functional domains, I used the nine functional domains Bunderson (2002)
19
Each individual then received a score that is equal to the number of domains they have worked in. These scores were then averaged to gain a team score for IFD.
IFD Distribution
To measure the distribution of IFD, the normalised Hirschman-Herfindahl Index (Herfindahl, 1950; Hirshman, 1964; Hirschman, 1980) was used. The Hirschman-Herfindahl Index is commonly used to measure the distribution of market share between companies and to determine if a company has a monopoly, but can be used to measure other distributions as well. The index (H) is calculated in the following way:
Where N is the number of functional domains a person has worked in and Si is the share of an individual's career that a domain has.
A flaw of the Hirschman-Herfindahl Index is that the outcome is always relative to the number of data points used. In this research, that means that an individual who has a perfectly even distributed IFD and has worked in five areas would receive a different score than an individual with an even distribution who has worked in three areas. To compensate for this, a normalised index was used, which scores each individual on a scale from 0 to 1, with a perfectly even distribution and zero being the most uneven distribution. The
normalised index (H*) is calculated the following way:
20
In practice, the more even an individual's distribution is, the more of a generalist, or jack of all trades they will be. The more uneven an individual's score is, the more of a specialist they will be. Therefore, an individual who only had experience in one functional domain received a score of 0, since they are a specialist instead of a generalist. After each individual received a score, these scores were averaged to gain a team-level score for IFD distribution.
IFD Recency
IFD Recency was calculated by measuring how long ago an individual had experience in another functional domain relative to the length of their career. It was calculated as the percentage of their career they have worked in their current functional domain. It was
calculated using the following formula: 1 - (Ld/Lc), where Ld is the amount of time in months an individual has worked in their current domain and Lc is the length of their entire career. Scores can range from 0 to 1, with higher scores reflecting that an individual has had experience in a different functional domain a relative short time ago. All of the individual scores in a team were averaged in order to gain a team level score.
Analysis
21
22 RESULTS
Descriptive statistics
Table 1
Pearson zero-order correlations among the study variables
Variable 1 2 3 4 5 6 7 8 9 1. Average performance 2. Boundary spanning ,257* 3. Team IFD ,042 0,025 4. IFD recency -,080 -,065 -,728** 5. IFD Distribution ,168 ,015 -,706** -,679**
6. Team career length ,036 -,165 -,289** -,295** ,290**
7. Team gender -,233* -,168 -,070 -,005 ,099 ,024
8. Team size ,006 ,198 ,016 ,009 -,058 -,171 -,219*
9. Coin age ,125 -,110 -,113 ,176 ,015 ,050 ,060 -,064
23 Table 2
Descriptives among the study variables
Variable M SD Max Min
Average performance 1429,393803 4091,045173 228822,06692 0,0315384615 Boundary spanning -0,519444444 5,272700758 16,02500000 -4,97500000 Team IFD -0,64055556 0,6644299933 1,635500000 -1,36450000 IFD recency 1,017194444 0,2212014916 1,503083333 0,5430833333 IFD distribution -0,003861111 0,2188161600 0,4929166667 -0,507083333 Team gender 0,0034166667 0,1228889149 0,1150833333 -0,384916667 Coin age 3,011111111 14,18805832 45,60000000 -8,40000000 Team size -1,68056 8,220160 33,275 -12,725
Team career length -0,129305556 56,17455362 159,7209167 -114,279083
Note. N = 90.
24
outcomes on average performance, an additional analysis was run using the intercept of performance instead of the slope as the dependent variable, but the model was not found to be significant in that analysis.
Because the moderators IFD recency and IFD distribution have a high correlation with IFD, a multicollinearity analysis was done, using the Variance Inflation Factor (VIF) (Allison, 1999). The VIF is calculated in the following way:
where 𝑅𝑗2 is the regression coefficient of the independent variable on the moderators. Tolerance values of .20 or lower and VIF values of 5 or higher present a multicollinearity problem (O’Brien, 2007). The results from the analysis are presented in table 3. The tolerance and VIF values indicate that multicollinearity does not pose a problem in the analysis.
Table 3
Results of multicollinearity analysis
Tolerance VIF
IFD and IFD recency 0,476 2,100
25 Model fit and regression results
Table 4
Values of the measuring models of fit and additional explained variance per added study variable
Model R2adj ∆R2adj p-value AIC NFI CFI RMSEA 𝜒² ∆𝜒² p-value df
Model 1 0.04 - .34 2594.11 0.63 0.64 0.44 2528.11 - .00 137
Model 2 0.04 0.00 .35 2618.85 0.63 0.64 0.42 2554.85 29.74 .00 157
Model 3 0.10 0.06 .10 2686.41 0.62 0.64 0.40 2616.41 61.56 .00 174
Model 4 0.12 0.02 .49 2876.87 0.60 0.63 0.37 2794.87 187.46 .00 211
Model 5 0.13 0.01 .08 3354.63 0.57 0.59 0.37 3258.63 463.76 .00 251
Note. Model 1 = Effect of control variables on performance. Model 2 = Main effect of
26
The model fit analysis was run using five models, each one adding another part of the conceptual framework. Model 5, which includes the entire conceptual framework was found significant (R² = .13, p < .1, N= 90). Model 3, which includes the main effects of the control variables, boundary spanning and IFD was also found significant (R² = .10, p < .1, N= 90).
The regression and PROCESS results can be seen in Table 5 and 6. Hypothesis 1a: (Intrapersonal functional diversity has a positive relationship with the amount of boundary spanning a team engages in.) was not proven significant, and in fact there was a significant relationship between team IFD and team Boundary Spanning in the other direction than predicted by the hypothesis (B = -4.129, SE = .000, p < .05).
Hypothesis 1b: (The positive relationship between intrapersonal functional diversity and team boundary spanning is moderated by IFD recency, in a way that higher IFD recency leads to a more positive relationship.) was proven significant (B = 4.110, SE = .760, p < .05).
Hypothesis 1c: (The positive relationship between IFD and team boundary spanning is moderated by the IFD distribution, in a way that the more a team’s IFD distribution is specialist, the more positive the relationship is.) was not proven significant (B = 0.278, SE = 3.171, p > .01).
Hypothesis 2: (IFD has an indirect positive relationship with team performance through boundary spanning. The relationship between IFD and boundary spanning is moderated by IFD recency and IFD distribution.) was not proven significant. Instead, there was a significant outcome suggesting that there is no correlation between the two. (B = 0.000, SE = 0.000, p < .05).
27 Table 5
Summary of Latent Growth Regression Analysis (N = 90)
28 Table 6
Effect values when the moderated effect is kept constant in three different categories
Moderator variable Moderator effect B p-value
Low 4.447 .199
IFD Recency Medium 1.471 .239
High -1.534 .197
Low 1.514 .302
IFD Distribution Medium 1.471 .239
High 1.427 .164
Direct 0.559 .808
29 DISCUSSION
Theoretical Contributions
The theoretical contributions of this research are limited to parts of the conceptual model and include some unexpected results. Hypothesis 1c was supported in the analysis, suggesting that IFD recency moderates the relationship between team IFD and team boundary spanning. This means that more recent experience in another functional domain will lead to a more positive relationship between team IFD and team boundary spanning. This is an important contribution to existing literature, since most of the existing research in IFD (see: Bunderson, 2003, De Vries et al., 2014), does account for the number of functional experiences, and the length of those experiences relative to each other, but does not account for how long ago an individual has worked in a different domain. These results do not discredit earlier research on IFD, but rather add another potentially interesting dimension to it. While this research
concerned the relationship between IFD and boundary spanning, this relationship may very well hold true in the relationship between IFD and other constructs.
An unexpected finding was that the relationship between team IFD and team
boundary spanning as a whole was negative, meaning that a higher team IFD actually led to a lower amount of team boundary spanning. This finding is not in line with earlier research (Ancona & Caldwell, 1992; Goldstein, 1984; Marrone, 2010). However, this research was done in an emerging and fairly unique industry, where teams are often small and need an extremely high concentration of specialized knowledge to succeed. This might lead to
30
spanning, while boundary spanning behaviours can include many more activities (see: Ancona & Caldwell, 1992).
Finally, no other significant relationships were found. This could be a side effect of the lack of generalizability of this dataset, or the accuracy of the measures, as will also be discussed in the limitations section. However, the possibility that, due to the unique nature of the cryptocurrency industry, commonly used theory does not apply in the same way should not be underestimated. In that regard, this research sheds new light on the differences between industries concerning boundary spanning and IFD. It is possible that in volatile industries where highly specialized knowledge is of paramount importance, the commonly accepted theories on boundary spanning and IFD do not apply. More research will need to be done on this topic, which is discussed in the future research section.
Practical Implications
Although more research will need to be done to further study the moderating factors around IFD and boundary spanning, this research has some useful practical implications. The most important one is that when hiring or assembling a team, recent experience in another functional domain is an important factor to consider when aiming for better boundary
spanning performance. When evaluating resumes, this could therefore be a factor to consider. Even when not hiring, managers can rotate employees through multiple positions to keep IFD recency high, as is already policy in some companies. Furthermore, this research could help managers consider the nature of the industry and the nature of the objectives of a team when hiring or assembling. For example, in a team or unit where specialized knowledge is
31 Limitations and Future Research
This study has some important limitations. Firstly, the data was collected using teams in a very new industry that operates in an uncertain and rapidly changing environment.
Furthermore, most cryptocurrency projects are at the stage in their lifetime where they require a high amount of very specialized knowledge to build the product but are not mature enough to have a fully-fledged project team. Therefore this data might be quite specific to the
cryptocurrency industry and might not be very generalizable to a broader context. Further research that aims to create generalizable results should consider investigating a mix of mature industries in slower changing environments and younger industries in uncertain environments. Not only can more generalizable results be found that way, researches can also look at the industry characteristics and how it moderates the relationships between boundary spanning and other team or individual characteristics. New perspectives on the relationship between IFD and boundary spanning and perhaps team science in a broader context can be gained if research accounts for differences in industry. This is particularly true for recently emerged industry types like the cryptocurrency industry, which are young, volatile and mostly digital in nature. Concretely, future researchers could investigate if the relationship between boundary spanning and team performance is different in younger industries than in mature industries, or in certain or uncertain environments.
32
Thirdly and finally, during the analysis, the latent growth curve modelling was done using a linear analysis. Since the performance of most teams was not strictly linear, a quadratic or even cubic model might be a better fit for such data. Alternatively, performance could be measured in another way entirely, such as looking at the average absolute market share over a period in time. Future research could explore those possibilities and possibly gain more valuable insights.
Finally, this study was done on the team level. Other studies into IFD have often been conducted on the individual or the team level, or a mix of both levels. To investigate if this model holds up on those levels, future research could investigate a conceptual model similar to this, but in other industries and on the individual level.
CONCLUSION
This research sheds more light on the effects of the characteristics of IFD, how they moderate the relationship between IFD and boundary spanning, and the nature of the
33 REFERENCES
Allison, P. D. (1999). Multiple regression: A primer. Pine Forge Press
Ancona, D. G., & Caldwell, D. F. (1992). Bridging the boundary: External activity and performance in organizational teams. Administrative Science Quarterly, 37: 634-665.
Arbuckle, J. L. (2014). Amos (Version 23.0) [Computer Program]. Chicago: IBM SPSS.
Bandura, A. 1997. Self-efficacy: The exercise of control. New York: W.H. Freeman.
Bjork, R. A. (1988). Retrieval practice and the maintenance of knowledge. Practical aspects of memory: Current research and issues, 1, 396-401.
Bunderson, J. S. (2003). Team member functional background and involvement in management teams: direct effects and the moderating role of power centralization. Academy Of Management Journal, 46(4), 458-474. doi:10.2307/30040638
Bunderson, J. S., & Sutcliffe, K. M. (2002). Comparing alternative conceptualizations of functional diversity in management teams: Process and performance effects. Academy of management journal, 45(5), 875-893.
34
Buzzell, R. D., Gale, B. T., & Sultan, R. G. (1975). Market share-a key to profitability. Harvard business review, 53(1), 97-106.
Chan, D. (1998). The conceptualization and analysis of change over time: An integrative approach incorporating longitudinal mean and covariance structures analysis (LMACS) and multiple indicator latent growth modeling (MLGM). Organizational Research Methods, 1(4), 421-483.
Choi, J. N. (2002). External activities and team effectiveness: Review and theoretical development. Small Group Research, 33(2), 181-208.
DeChurch, L. A., & Marks, M. A. (2006). Leadership in multiteam systems. Journal of Applied Psychology, 91(2), 311.
El-Sabaa, S. (2001). The skills and career path of an effective project manager. International journal of project management, 19(1), 1-7.
French, J. R., Raven, B., & Cartwright, D. (1959). The bases of social power. Classics of organization theory, 7.
Gladstein, D. (1984). Groups in context: A model of task group effectiveness. Administrative Science Quarterly, 29, 499–517.
35
Griffin, A., & Hauser, J. R. (1996). Integrating R&D and marketing: a review and analysis of the literature. Journal of product innovation management, 13(3), 191-215.
Hayes, A. F. (2013). Mediation, moderation, and conditional process analysis. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach edn. New York, NY: Guilford Publications.
Herfindahl, O. C. (1950). Concentration in the US steel industry. Unpublished PhD. Dissertation, Columbia University.
Hirschman, A. O. (1980). National power and the structure of foreign trade (Vol. 105). Univ of California Press.
Hirschman, A. O. (1964). The paternity of an index. The American Economic Review, 54(5), 761-762.
Johns G. Organizational behavior: understanding and managing life at work. 4th edition. Harper Collins, 1996.
Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (Vol. 2). New York: Wiley.
36
Marrone, J. A., Tesluk, P. E., & Carson, J. B. (2007). A multilevel investigation of antecedents and consequences of team member boundary-spanning behavior. Academy of Management Journal, 50(6), 1423-1439.
Marrone, J. A. (2010). Team boundary spanning: A multilevel review of past research and proposals for the future. Journal of Management, 36: 911–940.
Marks, M. A., Dechurch, L. A., Mathieu, J. E., Panzer, F. J., & Alonso, A. (2005). Teamwork in Multiteam Systems. Journal Of Applied Psychology, 90(5), 964-971.
Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Academy of management review, 26(3), 356-376.
Narayanan, A., Bonneau, J., Felten, E., Miller, A., & Goldfeder, S. (2016). Bitcoin
and Cryptocurrency Technologies: A Comprehensive Introduction. Princeton University
Press.
O’brien, R. M. (2007). A caution regarding rules of thumb for variance inflation factors. Quality & Quantity, 41(5), 673-690.
37
Richardson-Klavehn, A., & Bjork, R. A. (1988). Measures of memory. Annual review of psychology, 39(1), 475-543.
Tushman, M. L., & Scanlan, T. J. (1981). Boundary spanning individuals: Their role in information transfer and their antecedents. Academy of management journal, 24(2), 289-305.
de Vries, T. A., Walter, F., Van der Vegt, G. S., & Essens, P. J. (2014). Antecedents of individuals' interteam coordination: Broad functional experiences as a mixed blessing. Academy of Management Journal, 57(5), 1334-1359.
de Vries, T. A., Hollenbeck, J. R., Davison, R. B., Walter, F., & Van Der Vegt, G. S. (2016). Managing coordination in multiteam systems: Integrating micro and macro
perspectives. Academy of Management Journal, 59(5), 1823-1844.
