Socio-emotional wealth separation and decision making quality in family
firm TMTs: The moderating role of psychological safety Link
Peer-reviewed author version
Made available by Hasselt University Library in Document Server@UHasselt
Reference (Published version):
Vandekerkhof, Pieter; Steijvers, Tensie; Hendriks, Walter & Voordeckers, Wim(2017)
Socio-emotional wealth separation and decision making quality in family firm TMTs: The
moderating role of psychological safety. In: JOURNAL OF MANAGEMENT STUDIES,
DOI: 10.1111/joms.12277
Handle: http://hdl.handle.net/1942/23462
Socio-emotional wealth separation and decision making quality in family firm TMTs:
the moderating role of psychological safety
Abstract
Socio-emotional wealth (SEW), defined as the firm’s non-financial aspects meeting the
family’s affective needs, has become the dominant paradigm in family firm research. Recent
debate acknowledges potential SEW heterogeneity within family firms. This study considers
the effect of polarizing opinions on SEW preservation among TMT members as a source of
separation in the TMT. More concretely, we study the effect of SEW separation on TMT
decision making quality, while taking into consideration behavioral integration as a team
process and psychological safety as a team context. Based on a unique multiple respondent
sample of 300 managers from 55 Belgian private family firms, we find that behavioral
integration mediates the negative effect of SEW separation on TMT decision making quality.
In addition, we find that the negative effect of SEW separation on behavioral integration is
mitigated by psychological safety and even turns into a positive effect at high levels of
psychological safety.
Keywords: behavioral integration, psychological safety, separation, socio-emotional wealth,
top management teams
INTRODUCTION
Many studies have been carried out to identify issues that are idiosyncratic to family firms
(Berrone et al., 2012). Generally, these studies have highlighted that family firms are
motivated by more than mere financial objectives (Miller and Le Breton-Miller, 2014).
Gomez-Mejia et al. (2007) capture these ‘other motives’ with the introduction of the concept
of socio-emotional wealth (hereafter SEW), defined as “the non-financial aspects of the firm
that meet the family’s affective needs, such as identity, the ability to exercise family
influence, and the perpetuation of family dynasty” (Gomez-Mejia et al., 2007, p.106). To
date, it is acknowledged that SEW preservation tends to represent the pivotal frame of
reference for strategic decision making in top management teams (hereafter TMT) in family
firms (Zellweger et al., 2013).
However, there is considerable variation in the extent to which family firms aim to
preserve their SEW (Berrone et al., 2012), and similarly there is also variation in the extent to
which members of one family firm want to preserve SEW (Miller and Le Breton-Miller,
2014), a point that has often been neglected in prior SEW research. In this study, we focus on
these individual differences, and study how such individual differences in the preservation of
SEW within family firm TMTs affect team outcomes. With a few exceptions (e.g. Cruz et al.,
2010; Ling and Kellermanns, 2010; Minichilli et al., 2010; Patel and Cooper, 2014), studies
on TMT diversity in a family firm context are scarce even though research on TMTs in the
upper echelons tradition (Hambrick and Mason, 1984) has clearly shown that TMTs exert a
strong impact on firm strategy and outcomes (for an overview see Finkelstein et al., 2009).
We build on the seminal paper of Harrison and Klein (2007) and posit that differences in the
extent to which individual TMT members want to preserve SEW trigger ‘separation’ within
TMTs. SEW separation refers to fundamental differences in opinions and positions regarding
the preservation of SEW (Harrison and Klein, 2007). Separation reduces cohesiveness,
1
increases interpersonal conflict and distrust between team members that negatively affect
team performance (Harrison and Klein, 2007; Williams and O'Reilly, 1998). However, we
propose that TMTs can mitigate these negative effects by providing the right team climate.
The main purpose of this study is to deepen our understanding as to how and when
variation in positions regarding SEW preservation between TMT members affects the
decision making quality of family firm TMTs (e.g. Amason, 1996; Carmeli et al., 2012;
Dooley and Fryxell, 1999; Olson et al., 2007). Decision making quality is important because
it determines the content of the firm’s strategies and commitment to implementation
(Mustakallio et al., 2002), and seems to be a critical factor in establishing long-term
sustainability of almost every organization including family firms (e.g. Dooley and Fryxell,
1999; Mustakallio et al., 2002).
Our major contribution is that we propose a model that not only shows how diversity
in positions regarding the preservation of SEW of family firm TMT members negatively
affects the quality of decision making, but also demonstrates that the provision of the right
team climate (‘when’) allows TMTs to mitigate these effects. The explicit focus on team
climate can increase the value of team composition research for managerial practice, because
it is amenable to managerial design (Priem et al., 1999). In order to fully explain the effect of
SEW separation, we build on the rich research tradition of the upper echelon theory that
shows that we need to include both a TMT mechanism (process) and the TMT context
(emergent state) influencing that process (Cannella and Holcomb, 2005; Certo et al., 2006;
Wei and Wu, 2013) to better understand how diverse TMTs work and when their work is
translated into TMT performance. More concretely, we present a model where the process of
behavioral integration acts as the mechanism (mediator), and the emergent state of
psychological safety as the context (moderator). We argue that SEW separation is negatively
related to TMT decision making quality because it complicates behavioral integration (e.g.
2
collaboration, communication and joint decision making) within the team. Moreover, we
assert that a climate of psychological safety can mitigate the negative effects of SEW
separation on TMT behavioral integration.
We have chosen behavioral integration as the appropriate mediator because the
opposing positions on preferred goals and strategic direction between team members that
believe in the preservation of SEW and those that do not, are likely to cause tensions within
the team which makes it more difficult to collaborate, to communicate openly (Jehn and
Mannix, 2001; Liang et al., 2012) and to create high levels of team behavioral integration
(e.g. Bell, 2007; van Knippenberg and Schippers, 2007). Hence, behavioral integration acts as
the mediating team process to explain how SEW separation affects TMT decision making
quality. Behavioral integration defines the degree to which the members of the TMT engage
in mutual and collective interaction (Hambrick, 1994). Studies show that high levels of
behavioral integration in TMTs positively affect team performance measures, as these teams
are better able to deal with complexity and to integrate diverging opinions or positions into
balanced decisions (Carmeli and Halevi, 2009; Carmeli and Schaubroeck, 2006; Carton and
Cummings, 2012).
Next, we argue that the detrimental effect of SEW separation on behavioral integration
can be addressed by creating an emergent state of psychological safety. Psychological safety
is a shared belief that the team is safe for interpersonal risk taking without fear of
embarrassment, rejection or punishment (Edmondson, 1999). It creates a team climate that
alleviates concerns about others’ reactions; people are being themselves through a sense of
interpersonal trust and mutual respect (Gibson and Vermeulen, 2003; Gibson and Gibbs,
2006; Joshi and Roh, 2009). The choice of psychological safety as a moderator in our research
model is justified as it is useful to address SEW differences between family firms’ TMT
members. Psychological safety allows for the expression of disagreements with respect to
3
SEW preferences, whereas more common emergent state measures such as cohesion (e.g.
Barrick et al., 2007; Marks et al., 2001) discourage the expression of disagreements (Burke et
al., 2006). Discouragement of disagreement can ultimately result in groupthink where
deviating opinions are not allowed, and team members are coerced to follow the communis
opinio (Janis, 1972). This reduces open communication between team members and puts
pressure on collaborative behavior and joint decision making (e.g. behavioral integration).
We use the commonly used subjective TMT performance measure decision making
quality as dependent variable (Amason, 1996; Carmeli and Schaubroeck, 2006; Carmeli et al.,
2012; Olson et al., 2007; Wooldridge and Floyd, 1990). TMT decision making quality is
defined in this study as a combined measure that consists of both the quality of decision
formulation and implementation (Dooley and Fryxell, 1999). The former is related to the
extent to which TMT decisions enhance the achievement of organizational goals (Amason,
1996; Dooley and Fryxell, 1999), while the latter refers to the amount of effort and tenacity
team members extend toward seeing TMT decisions successfully implemented (Dooley and
Fryxell, 1999; Wooldridge and Floyd, 1990).
In contrast to other more distant objective firm performance measures such as firm
profitability, TMT decision making quality is an accurate measure to verify whether the
decisions taken by the TMT enhance the achievement of the goals and direction of the firm,
taking into consideration the variation in positions regarding SEW preservation in the TMT,
and whether TMT members are committed to implementing these decisions (Colbert et al.,
2008; Dess, 1987; Rapert et al., 2002). Moreover, given the specific research context of
family firms, objective financial performance measures might not fully capture whether the
firm/TMT has made the best decisions. Family firm TMTs are likely caught in a dilemma
where they have to weigh the anticipated losses and gains (Bromiley, 2009, 2010) in terms of
their impact on both SEW endowment and financial wealth (Gomez‐Mejia et al., 2014). This
4
balancing act limits the use of objective measures such as firm profitability because it reflects
only financial wealth, and not the SEW preservation that many family firms consider an
important goal and objective (Gomez-Mejia et al., 2007).
We test our hypotheses based on data derived from detailed questionnaires, filled out
by each TMT member, resulting in a unique sample of 300 top managers out of 55 Belgian
private family firms. Collecting data from individual TMT members allows us to directly
measure the extent to which they want to preserve SEW. Previous research explained the
effect of the ratio family/nonfamily members within family firm TMTs by using SEW
arguments (e.g. Minichilli et al., 2010; Patel and Cooper, 2014). While it is correct to assume
that family executives attach on average a higher value to SEW as the point of reference than
nonfamily executives, this dichotomous diversity measure neglects the possibility that
individual family managers can have substantially different SEW preferences than other
family managers within the same team, and that nonfamily managers can also attach high
salience to affective outcomes as captured by SEW (Huybrechts et al., 2013; Miller and Le
Breton-Miller, 2014). Even though one could opt for an alternative sampling strategy with
only a selection of managers per team (e.g. Simons et al., 1999), this would increase sample
size but at the expense of the accuracy and reliability of the data (Chua et al., 1999; Homburg
et al., 1999; Podsakoff et al., 2003). In this respect, Allen et al. (2007) show that correlations
between team diversity scores and team outcomes based on partial rather than complete sets
of team member attributes are substantially attenuated.
5
THEORY AND HYPOTHESES DEVELOPMENT
SEW separation and decision making quality
A (family) firm can be considered as a reflection of its TMT, since the experiences, attitudes
and beliefs of TMT members are assumed to be the drivers of strategic and operational
decisions (Cyert and March, 1963; Hambrick and Mason, 1984). The central question in more
than 30 years of TMT research has been whether TMT diversity enhances organizational
effectiveness. Empirical evidence on this relationship however is equivocal at best (Carmeli
and Schaubroeck, 2006; Certo et al., 2006; Priem et al., 1999). The meta-analysis of Certo et
al. (2006) shows that the direct relations between TMT diversity and performance are modest
and do not hold across all indicators of TMT diversity. An important reason for these
inconsistent results is that researchers advocate that diversity can be regarded as a unitary
construct (van Knippenberg et al., 2004). However, there is increasing evidence that a
universal theory capturing the relationship between TMT diversity and performance is not
warranted (Harrison and Klein, 2007; Nielsen, 2010). As such, there is a need to better
understand the meaning of TMT diversity or within-team differences (Lawrence, 1997).
An important breakthrough has been provided by Harrison and Klein (2007) who
assert that diversity is a compositional construct, consisting of three types of diversity:
variety, disparity and separation. First, diversity as variety refers to categorical differences in
information, knowledge or experience. Second, diversity as disparity relates to inequality
within the team in terms of authority, prestige and status which triggers vertical distance
among group members. Third, diversity as separation represents differences in positions or
opinions on team- or firm-related issues that may create a horizontal distance among group
members.
Harrison and Klein (2007) stress the importance of shifting attention away from
considering isolated differences among team members to considering the pattern of these
6
differences as a whole. As such, they do not focus on individual differences between team
members but rather the ‘combination’ of differences between all team members within one
team. As such, minimal separation occurs when all team members occupy the same position
at any location on the continuum of a diversity indicator (Harrison and Klein, 2007). Maximal
separation occurs when there are two (and only two) staunchly divided but balanced blocks
within the TMT, each holding a position on a diversity indicator as far from the other as
possible (Harrison and Klein, 2007). In this study, we posit that differences in positions of
team members on the salience of SEW within the family firm TMT is consistent with
diversity as separation. Generally, team members that perceive SEW preservation to be
important are more likely to have a non-economic and emotional value towards the family
firm, and like to meet the affective needs of the family firm, such as the perpetuation of the
family dynasty and identity. In contrast, TMT members who perceive SEW preservation to be
less important are more likely to show less emotional attachment, tend to be more concerned
with the financial consequences of decisions and might prefer decisions that benefit the firm
in terms of financial indicators such as growth, profitability and market position (GomezMejia et al., 2007). These basic attitudinal differences between high and low SEW salience
are likely to influence how individuals analyze, interpret, feel and act upon the same decision
situation.
The use of different frames of reference of team members that hold high and low
positions on SEW likely results in fundamental differences of opinion about the course and
direction of the firm (Gomez-Mejia et al., 2007; Miller and Le Breton-Miller, 2014). The
impact of these differences on team outcomes is of great interest to diversity researchers (Bell,
2007; Carpenter et al., 2004). In general, polarizing effects through separation are assumed to
harm team outcomes such as TMT decision making quality (e.g. Bell, 2007; Harrison and
Klein, 2007). Dooley and Fryxell (1999) stress the importance of capturing a quality indicator
7
of both decision formulation and implementation, when studying decision making quality.
The quality of decision formulation refers to whether decisions are made based on valid
argumentation, the best available information, and whether the solutions that are generated fit
the organization’s goals and contribute to its overall effectiveness (Carmeli and Schaubroeck,
2006). A thorough understanding of the priorities of the firm is essential to a TMT’s
willingness to assume responsibility for the implementation of actions contributing to these
priorities (Kellermanns et al., 2005).
Authors like Eisenhardt and Bourgeois (1988) and Wiersema and Bantel (1992) have
already provided evidence that opposing opinions or positions lower the quality of decisions
as they can increase the chance that counterparts will engage in counter efforts to sabotage the
decision (Guth and MacMillan, 1986). Furthermore, the motivation of all TMT members to
ensure a successful implementation of decisions and thus be highly committed, is also
negatively affected when positions on team- or firm-related issues are too divergent (Bandura,
1986; Dooley and Fryxell, 1999; Wooldridge and Floyd, 1990). Based on these research
findings, we also expect SEW separation to be negatively related to TMT decision making
quality. Highly separated TMTs will have many more difficulties with the integration of
opposing perspectives and alternatives into meaningful and qualitative decisions which are
supported by all TMT members (Amason, 1996; Dooley and Fryxell, 1999). As such,
differences in opinions, beliefs and attitudes regarding SEW are expected to lower overall
TMT decision making quality by harming both the quality of TMT decisions as well as
managers’ commitment to these decisions. We therefore hypothesize:
Hypothesis 1 (H1): SEW separation has a negative effect on TMT decision making
quality.
8
The mediating role of behavioral integration
Next, we argue that high levels of SEW separation negatively affect TMT decision making
quality through low levels of behavioral integration. The comprehensive meta-construct
“behavioral integration” has been developed by Hambrick (1994) and has been acknowledged
as a core TMT process that captures the overall team factor of TMTs (Simsek et al., 2005).
Hambrick (1994) argued that TMT processes are distinct from group processes at other levels
in the organization, because TMT members face higher levels of responsibilities, both
individually and interdependently as members of a firm’s top decision making team.
Hambrick (1995) found that truly integrated TMTs engage in several interrelated processes to
reflect the inherent complexity and dynamism of decision making that cannot be adequately
captured by any single process dimension. Behavioral integration consists of one social
dimension (TMT level of collaborative behavior) and two task dimensions (TMT information
exchange, and TMT joint decision making). As such, it encompasses several team process
elements that were previously represented as separate constructs like social integration or
group cohesion, frequency and quality of information exchange, and collaboration (e.g. Boone
and Hendriks, 2009; Buyl et al., 2011; Harrison et al., 2002; Wei and Wu, 2013; Woehr et al.,
2013). However, Hambrick (1994, 1995) argued that these mutually reinforcing processes,
when taken together, better capture a TMT’s level of wholeness and unity of effort than does
each
dimension
when
examined
separately.
Research
has
acknowledged
this
multidimensional origin of behavioral integration (Simsek et al., 2005).
SEW separation in family firm TMTs lowers the level of behavioral integration
because polarization prevents TMTs from working together effectively (e.g. Bell, 2007; van
Knippenberg and Schippers, 2007). Polarization hampers information exchange, reduces
collaboration behavior, and inhibits joint decision making. Polarization negatively affects
collaboration because one of the most robust and reliable findings in social psychology
9
studies is that people consciously and unconsciously prefer others who are similar to them
(for a review see Williams and O'Reilly, 1998). This phenomenon has also been supported in
small group research and in sociological research on homophily, defined as: “the tendency for
persons who affiliate with each other to be similar on various attributes" (Hogue and
Steinberg, 1995, p.897). Based on the satisfaction individuals obtain from working with
similar others, homogeneous groups are expected to be more cooperative and have less
conflict than heterogeneous groups, because of enhanced feelings of familiarity, attraction and
trust. These reinforcing effects of similarity, e.g. minimal or reduced separation, will then be
associated with more cooperative and cohesive group processes (e.g. Locke and Horowitz,
1990). People are also likely to communicate more readily with persons believed to be
similar, because of the role of similarity in interpersonal attraction (Byrne, 1971). Indeed,
social-psychological studies have shown that, if team members have diverging positions or
opinions, so-called process losses are likely to occur due to problematic communication
between the team members (Zenger and Lawrence, 1989). Furthermore, opposing, separate
factions in a TMT create fear and uncertainty about expressing differing attitudes in the TMT
(Harrison et al., 2002; Homberg and Bui, 2013), which stimulates political behavior and limits
joint decision making. Therefore, we posit that opposing positions and opinions on SEW
within a TMT are negatively related to behavioral integration.
Consequently, the negative impact of SEW separation on the level of behavioral
integration has important consequences for team outcomes. For instance, Hambrick (1995)
noted that TMTs with low levels of behavioral integration experience problems with adapting
in time to external challenges. In line with our research focus, Carmeli et al. (2006) argue that
less behavioral integration might hinder the TMT in attaining higher quality of decisions.
More specifically, the way TMT members interact and work together as a ‘team’ will not only
have an impact on the quality of its decisions, but also on how well such decisions are
10
implemented through team members’ commitment to them (Dooley and Fryxell, 1999;
Schweiger and Sandberg, 1991). Greater joint participation and shared decision making
increase the likelihood that strategic actions account for both emotional and financial goals.
Furthermore, shared decision making increases accountability for both members who are high
and low on SEW. To summarize, the polarizing impact of opposing positions and opinions on
SEW in a family firm TMT will negatively affect the level of behavioral integration, and
subsequently TMT decision making quality (Ellis et al., 2013; Kearney and Gebert, 2009;
Kearney et al., 2009; Lau and Murnighan, 2005). Therefore, we propose:
Hypothesis 2 (H2): The relationship between SEW separation and TMT decision
making quality is mediated by behavioral integration
The moderating role of psychological safety
We have already argued that SEW separation negatively influences TMT decision making
quality through lowering the level of behavioral integration in the team. We argue that
psychological safety provides a team context where high levels of behavioral integration can
be achieved, despite differing opinions on SEW. Psychological safety creates an atmosphere,
characterized by interpersonal trust and mutual respect, in which team members are
comfortable being themselves and do not fear negative judgments when expressing their
different opinions or positions on SEW (Edmondson, 1999; Kahn, 1990; West and Anderson,
1996). Psychological safety is an emergent state that influences the execution of team
processes such as team cooperation, communication and shared decision making (Marks et
al., 2001). Emergent states are not team processes in and of themselves since they do not
describe the nature of TMT members’ interactions such as behavioral integration.
Psychological safety as an emergent state is essential for facilitating processes that may give
rise to behaviors and outcomes (Carmeli et al., 2012). According to Carmeli and Schaubroeck
11
(2006), the willingness of TMT members to respect and accept differences avoids the risk of a
TMT becoming a group of individuals instead of a team. This willingness can be created by a
sense of psychological safety in the team (Gibson and Vermeulen, 2003; Gibson and Gibbs,
2006; Joshi and Roh, 2009). Hence, psychological safety is a contextual team factor (e.g.
emergent state) that facilitates the process of working together despite the presence of
different positions on SEW within the family firm TMT.
Psychological safety creates a sense of equality within the team. Although team
members have different standpoints, a climate of psychological safety creates an atmosphere
where each standpoint is respected, which creates a perception of equality between the team
members, irrespective of their viewpoints on the preservation of SEW. Patel and Cooper
(2014) argue that equality between members of family firm TMTs improves collaborative
behavior and shared decision making. Psychological safety can hence create a team climate
that allows TMTs with strong diverging positions on SEW, to engage in the constructive team
process of behavioral integration that in turn positively affects decision making quality
(Anderson and West, 1998; González-Romá et al., 2009). Therefore, we hypothesize:
Hypothesis 3 (H3): Psychological safety moderates the negative and indirect effect of
SEW separation on TMT decision making quality (through behavioral integration)
such that the relationship is less negative when psychological safety is higher.
In order to test the formulated hypotheses, this study will examine the effect of SEW
separation on TMT decision making quality. Behavioral integration as a team process is
expected to mediate this relationship whereas psychological safety as a team context is
expected to moderate the relationship between SEW separation and TMT decision making
quality. The research model is illustrated in Figure 1 below.
12
----------------------------------------------------------------------------------------------------------------Insert Figure 1 about here
-----------------------------------------------------------------------------------------------------------------
METHODS
Sample
The target sample consisted of Belgian private family firms, defined as firms perceived by the
CEO as being a family firm and where ownership is controlled by a single family (50% or
more of the shares) and at least two members of the same family significantly influence the
firm through positions in a governance mechanism (Chua et al., 1999; Tagiuri and Davis,
1996). Other conditions that had to be met by the firms are: (1) at least 20 employees active in
the firm because many small firms do not operate with a ‘real’ TMT, (2) at least three
managers in the TMT. A TMT was defined as the group of managers consisting of the CEO
and those managers who directly report to the CEO (Boeker, 1997).
The so-called snowball sampling method was chosen to select our sample cases because
of three major difficulties that we faced in data collection. Attaining reliable information and
a priori identification of private family firms is difficult (Daily and Dollinger, 1993; Schulze
et al., 2003). A comprehensive list of all Belgian family firms, based on a chosen definition of
family firms, is lacking. Moreover, the two conditions (based on firm and TMT size) that
simultaneously had to be met are difficult to detect a priori. The snowball sampling procedure
helps locate members of special hard-to-find populations via referral by network contacts
(Biernacki and Waldorf, 1981; Saunders et al., 2007) and has already been used in family
business studies that faced comparable constraints (e.g. Bettinelli, 2011; Björnberg and
Nicholson, 2012; Farrington et al., 2012; Fiegener et al., 1996; Van der Merwe, 2007; Venter
et al., 2003). The risk of sampling bias through this specific method (Lee, 1993), is less of a
problem in this study because the descriptive statistics (more details in Table I) provide
13
evidence of sufficient variation within the sample in terms of firms and respondents (e.g. firm
size, firm age, top management composition). Furthermore, we utilize multifaceted research
models (i.e. through inclusion of moderating and mediating effects) that lead to less problems
related to case selection bias (Simons et al., 1999).
Sample selection started by visiting a small group of family firm CEOs in our network
whose firms matched our sampling criteria. After sending an introduction letter with a brief
description of our research and the procedures, we conducted a structured interview with the
CEO to find out if he/she was willing to collaborate. The CEO was asked to define the TMT
with respect to identifying its fellow top managers (Pitcher and Smith, 2001) in order to give
the CEO and each TMT member self-administrated structured questionnaires. To ensure
confidentiality, we added a return envelope for each TMT member, indicating only the firm
name, in which they could put the completed questionnaire. Hereby we also guaranteed each
participant that no individual information would be reported back to the CEO and/or TMT
members. After a few weeks, we personally went back to pick up all return envelopes. We
further used our contacts with the CEOs of these family firms to get in touch with CEOs of
other family firms. This sampling procedure led to 68 structured interviews with CEOs of
Belgian private family firms that were interested in collaborating. Our final sample consisted
of 55 Belgian private family firms from which we received complete information (i.e. from
the CEO and each TMT member) while 13 cases were excluded (8 firms decided not to
collaborate in the end and 5 firms did not manage to retrieve information from the whole
team). Within our sample of 55 Belgian private family firms, a total of 300 individual
respondents filled in the questionnaire.
Measures
Decision making quality. As in previous team-level studies (e.g. Amason, 1996; Carmeli et
al., 2012; Dooley and Fryxell, 1999; Mustakallio et al., 2002; Olson et al., 2007), we use
14
perceptual measures of decision making quality. We measure decision making quality as a
combined measure of both decision quality and decision commitment (Mustakallio et al.,
2002). Decision quality relates to the extent to which a decision enhances the achievement of
organizational goals. Items for decision quality are based on Amason (1996), and evaluated
by the CEO and all team members on a five point Likert scale. Items include: (i) ‘The
decisions made by the TMT generally have a positive effect on achieving the goals of the
firm’; (ii) ‘Generally, the decisions made by the TMT meet their expectations’; and (iii)
‘Generally, the TMT is satisfied with the quality of their decisions’. Decision commitment
refers to the extent to which TMT members accept and commit to their decisions (Dooley and
Fryxell, 1999; Korsgaard et al., 1995). Hence, decision commitment reflects TMT members’
attitudes toward TMT decisions, after they have been made. The items of decision
commitment are based on Wooldridge and Floyd (1990), and evaluated by all TMT
executives on a five point Likert scale. Sample items include: (i) ‘The team members put in a
lot of effort to adequately implement the decisions made by the TMT’; (ii) ‘The decisions
made by the TMT meet the priorities of the individual team members’; and (iii) ‘The team
members generally agree that the decisions made contribute to firm performance’. A
principal component factor analysis revealed a single decision making quality factor. All
factor loadings are higher than .572 with an eigenvalue of 5.56 and explain 46.37% of the
variance among the items. The Cronbach’s alpha for this scale was 0.89. The overall measure
of decision making quality will vary from 1 (low level of decision making quality) to 5 (high
level of decision making quality).
Socio-emotional wealth separation. We use a direct measure of SEW based on the FIBER
model of Berrone et al. (2012). The FIBER model distinguishes between five major
dimensions of SEW: (i) family control and influence; (ii) family members’ identification with
the firm (iii) building social ties; (iv) emotional attachment; and (v) renewal of family bonds
15
to the firm through dynastic succession. In this study, we use one item per dimension to
capture the position of each individual TMT member on the preservation of SEW. We asked
team members to indicate their position on statements such as: (i) ‘It is essential to preserve
family control and independence of the family firm’; (ii) ‘The family members have a strong
sense of belonging to the family firm’; (iii) ‘Nonfamily members are treated as part of the
family’; (iv) ‘Emotional bonds between family members are strong’; (v) ‘Successful business
transfer to the next family generation is an important goal of the family firm’. A principal
component factor analysis revealed a single SEW factor. All factor loadings are higher than
.607 with an eigenvalue of 2.39 and explain 47.74% of the variance among the items. The
Cronbach’s alpha for this 5-item scale was 0.72. The overall SEW construct will vary from 1
(low level of SEW) to 5 (high level of SEW). As the purpose of our study is to express the
distribution of differences among team members on their positions on SEW within the TMT
of the family firm, we use the operationalization of Harrison and Klein (2007). The withinunit standard deviation will be used to express the cumulative distances in SEW that capture
separation based on positions and opinions within the family firm TMTs. It should be
emphasized that it is the extent to which team members are similar or different and thus to
which extent the TMT is polarized that matters, not the fact whether team members are high
or low on SEW (Bell et al., 2011; Harrison and Klein, 2007).
Psychological safety. We used the 7-item measure of Edmondson (1999) to measure the
psychological safety climate in the TMT. Sample items include: (i) ‘Members of this team are
able to bring up problems and tough issues’; (ii) ‘People on this team sometimes reject others
for being different’ (negatively formulated); (iii) ‘No one in this team would deliberately act
in a way that undermines my efforts’; and (iv) ‘It is safe to take a risk on this team’. Hereby,
we capture the shared belief with regard to the extent to which managers feel psychologically
safe in taking interpersonal risks, speaking openly and making mistakes (Carmeli and Gittell,
16
2009). A principal component factor analysis revealed a single psychological safety factor.
All factor loadings are higher than .454 with an eigenvalue of 2.80 and explain 39.88% of the
variance among the items. The Cronbach’s alpha for this 7-item scale was 0.72. The construct
psychological safety will vary from 1 (low level of psychological safety) to 5 (high level of
psychological safety).
Behavioral integration. Hambrick (1994) divided the meta-construct behavioral integration
into three interrelated and mutually reinforcing team processes: collaborative behavior,
information exchange, and joint decision making. In our study, we use specific measures for
each dimension that capture the process itself before assessing all items together to express
the meta-construct TMT behavioral integration. In line with Boone and Hendriks (2009), we
build on Hambrick (1994) to measure collaborative behavior by the following three items on a
5-point Likert scale: ‘There is a fruitful, rewarding cooperation within this team’; ‘It is easy
to ask advice from any member of this team’; and ‘This TMT operates as a “real” team’.
With regard to information exchange, we follow the reasoning of Buyl et al. (2011) by
adapting the following 2 items on a 5-point Likert scale: ‘The communication in this team
normally goes without hidden agendas’; and ‘In general, differences of opinions with respect
to task execution are discussed openly and thoroughly’. These items are derived from the
‘perceived communication openness’ scale of O’Reilly and Roberts (1976) that closely
resembles the degree to which information within the TMT is exchanged and integrated in an
open way (Buyl et al., 2011; Dahlin et al., 2005). Building on Hambrick (1994, 1995), joint
decision making was measured by the next two items on a 5-point Likert scale: ‘In decision
making, usually every team member’s input is used’; and ‘Most team members only have
limited influence on the decision making process’. A principal component factor analysis
reveals that the 7 items load together on one factor with factor loadings higher than .546 with
an eigenvalue of 3.39, explaining 48.40% of the variance among the items. Cronbach’s alphas
17
for collaborative behavior, information exchange and joint decision making are respectively
0.70, 0.60 and 0.64, while the overall reliability of the meta-construct was 0.81. The metaconstruct was set on a 5-point Likert scale ranging from 1 (low level of behavioral
integration) to 5 (high level of behavioral integration).
Control variables. For an organizational level control variable, we use firm size, measured by
the number of full-time employees, since firm size affects the framework of organizational
decision making: levels of debate and disagreement in decision making increase in larger
firms (Iaquinto and Fredrickson, 1997; Papadakis et al., 1998; Shepherd and Rudd, 2014). We
use the natural logarithm of the number of employees to account for its skewed distribution
(Gujarati, 1995). For a team level control variable, we use TMT size, measured by the number
of TMT members (CEO and those managers directly reporting to the CEO). TMT size is
considered to possibly have an effect on the team and decision making processes and
outcomes (Kearney and Gebert, 2009; Simsek et al., 2005; West and Anderson, 1996). We
again use the natural logarithm of team size to account for its skewed distribution (Gujarati,
1995). Finally, to control for generally used diversity measures that reflect distributed
differences within a team, we also include a measure for age and gender diversity (Harrison
and Klein, 2007).
Data Reduction & Common Method Variance
Data aggregation. This study is focused on examining the impact of separation within the
TMT on team processes and outcomes. Our analyses are performed at the team level, while
our key variables are measured at the individual level by all TMT members of each firm.
Multiple respondents have been considered as being more reliable and less superficial
(Bowman and Ambrosini, 1997). Therefore, we aggregated all individual measures into teamlevel scores. However, assessment of the consistency of responses within a team and
suitability of data aggregation are required (Smith et al., 1994). Therefore, we assessed the
18
suitability of data aggregation by using interteam-member agreement (Rwg), one-way analysis
of variance, and intra-class correlation coefficients ICC(1) and ICC(2) (Bliese, 2000; James et
al., 1993). A median interrater agreement score of 0.87 for decision making quality, 0.80 for
psychological safety and 0.75 for behavioral integration were all well above the acceptance
value of 0.70 (James et al., 1993). Furthermore, a significant one-way analysis of variance for
the three variables (p < .000 for decision making quality, psychological safety and behavioral
integration) and an ICC(1) of 0.37 for decision making quality, 0.15 for psychological safety
and 0.32 for behavioral integration revealed that the between-team variances were larger than
within-team variances. Finally, an ICC(2) of 0.89 for decision making quality, 0.72 for
psychological safety and 0.81 for behavioral integration indicate sufficient reliability of
average team perceptions to obtain team-level scores (Bliese, 2000).
Confirmatory factor analyses. We performed confirmatory factor analyses to test the
construct validity of our key variables (SEW separation, behavioral integration, psychological
safety and TMT decision making quality). We start with examining several fit indicators of
our baseline four-factor model. Based on the overall model’s chi-squared, CFI, RMSEA,
SRMR and TLI, we can conclude that the baseline four-factor model fitted the data properly
with χ ² = 675.65 (424), p < 0.01; CFI = 0.93; RMSEA = 0.044; SRMR = 0.049; and TLI =
0.92. All factor loadings were also significant, demonstrating convergent validity. Finally, the
AVE scores of each factor in the four-factor solution are above the threshold of 0.50
(behavioral integration = 0.585; psychological safety = 0.765; SEW = 0.528; and decision
making quality = 0.742), which further confirms the validity of the four-factor model (Fornell
and Larcker, 1981). Next, we compared our baseline four-factor model with alternate models
to test the discriminant validity of the four-factor model. More concretely, we compared the
baseline model with several three-factor models (behavioral integration and psychological
safety combined, behavioral integration and decision making quality combined, psychological
19
safety and decision making quality combined); a two-factor model (behavioral integration,
psychological safety and decision making quality combined); and a one-factor model. The
three-factor models had the following goodness-of-fit statistics: χ² = 725.51 (429), p < 0.01;
CFI = 0.90; RMSEA = 0.050; SRMR =0.051; and TLI = 0.90 for the model in which
behavioral integration and psychological safety are combined, χ² = 766.92 (429), p < 0.01;
CFI = 0.90; RMSEA = 0.051; SRMR =0.052; and TLI = 0.89 for the model in which
behavioral integration and decision making quality are combined, χ² = 765.22 (429), p < 0.01;
CFI = 0.90; RMSEA = 0.051; SRMR =0.053; and TLI = 0.89 for the model in which
psychological safety and decision making quality are combined. The goodness-of-fit statistics
for the two-factor model are: χ² = 807.14 (431), p < 0.01; CFI = 0.89; RMSEA = 0.054;
SRMR =0.054; and TLI = 0.88, while those for the one-factor model are: χ ² = 984.08 (431), p
< 0.01; CFI = 0.84; RMSEA = 0.065; SRMR = 0.065; and TLI = 0.82. The comparison
provides evidence of construct distinctiveness for SEW separation, behavioral integration,
psychological safety and TMT decision making quality (Bentler and Bonett, 1980; Browne et
al., 1993; Cheung and Rensvold, 2002). Since the four-factor model fitted the data better than
the alternate models, we can also confirm the discriminant validity of the constructs.
Common method variance. Because our predictor and criterion variables are obtained from a
single source, any observed covariance may be due to the fact that the variables in our model
share the same method of measurement (Podsakoff et al., 2012). Therefore, we first tried to
reduce the potential for common method variance (CMV) by several procedural remedies as
suggested by Podsakoff et al. (2003; 2012) such as trying to camouflage interest in criterion
and predictor variables in the cover story of the questionnaire and avoiding asking for
sensitive data. In addition, we guaranteed confidentiality of the data and personally collected
the surveys. For each TMT member, we received the questionnaire in a separate sealed
envelope to ensure that individual responses were not accessible by other team members. This
20
procedure makes respondents less likely to answer in a socially desirable way (Podsakoff et
al., 2012). Finally, we eliminated any ambiguous and unfamiliar scale items by performing a
pretest of the questionnaire in which several CEOs of family firms critically pretested all
questions. Based on the comments of this pretest, any ambiguities in scale items were
corrected.
Second, the specific methodology that we used in this study further eases common
method variance concerns. We calculated the independent variable (SEW separation) in a
totally different way than our dependent and mediator variables. More specifically, SEW
separation is a diversity measure calculated as the distribution of differences among team
members on their positions towards SEW while for the other variables in the model, we
aggregated all individual measures into team-level scores, which is a different way of using
the aggregate data. Next, our moderated mediation model can be considered to be a complex
model which is very likely to reduce common method variance. Indeed, the specific
relationships in our model are unlikely to be part of the individual respondents’ cognitive
maps and theories-in-use which substantially eases common method concerns (Chang et al.,
2010). In addition, Siemsen, Roth and Oliveira (2010, p.472) found that “interaction effects
cannot be artificially created through Common Method Variance”. As CMV is found to
deflate regression estimates of interaction effects, a significant interaction effect of SEW
separation and psychological safety on behavioral integration should be considered as strong
evidence that the effect exists (Siemsen et al., 2010).
Third, we also performed three ex post CMV tests. As a first preliminary test, we
performed a Harman one-factor test on the individual data of the four main variables in our
research model (Podsakoff et al., 2003) which resulted in a five factor solution accounting for
52.58% of the total variance and a first factor accounting for 32.31% of the total variance.
(Note that this procedure implies the use of the straight scores of the respondents on the SEW
21
scale and thus does not measure the SEW separation construct of our research model). As a
second test, we estimated an unmeasured latent method factor model on the three variables of
our research model for which CMV could be a potential problem (behavioral integration,
psychological safety and decision making quality) (Podsakoff et al., 2003; Podsakoff et al.,
2012). We performed this test on the individual level data as well as the aggregate level data.
This analysis showed a common factor value of 0.16 which represents a common variance of
(0.13)² = 0.0169 or 1.69% (7.84% for the aggregate data). As a third test, we used a common
marker variable technique (Lindell and Whitney, 2001). We identified a variable in our
database that could serve as a viable marker variable for this test: we also asked our
respondents six questions concerning the six executive value dimensions defined by
Hambrick and Brandon (1988) (collectivism, rationality, novelty, duty, materialism and
power) and composed an executive value variable with these six items. These items were not
(behavioral integration, psychological safety, SEW) or only weakly (decision making quality)
correlated with our main variables and are expected to share potential common rater, common
item method and socially desirability bias with them (Podsakoff et al., 2012). This analysis
shows a common factor value of 0.20 and a common variance of 4% (2.56% for the aggregate
data). In sum, these CMV statistical tests further suggest that the likelihood that our results are
the result of common method bias is low.
RESULTS
Prior to hypotheses testing, descriptive statistics and correlations are summarized in Table I.
A family firm in our sample has on average 374 employees and a management team of about
5 members (including the CEO). The mean levels are for behavioral integration 3.80, 4.08 for
decision making quality, and 3.95 for psychological safety, which are comparable to prior
research (e.g. Carmeli et al., 2012; Dooley and Fryxell, 1999; Mustakallio et al., 2002; Simsek
22
et al., 2005). With regard to TMT members’ SEW preservation, mean level is 3.88 and an
important range of SEW values was found for family as well as nonfamily managers in our
sample (not reported in the tables). 11.2% of the family managers have a SEW value lower or
equal to 3 with a minimum value of 2.20, 27.9% of the family managers have a value between
3 and 4, while 60.9% of the family managers hold a SEW ranging between 4 and 5 with a
maximum value of 5. In comparison, 15.6% of the nonfamily managers have a SEW value
lower or equal to 3 with a minimum value of 1.20, 50.6% of the nonfamily managers have an
average value between 3 and 4, while 33.8% of the nonfamily managers indicate a high value
of SEW salience ranging between 4 and 5 with a maximum value of 5.
The correlations show a significant (univariate) positive relationship between
psychological safety as well as behavioral integration and the quality of decision making in
the TMTs. There is also a significant negative relationship between SEW separation and TMT
decision making quality. Furthermore, a negative relationship between SEW separation and
both behavioral integration and psychological safety was found. Moreover, the team process,
behavioral integration, and the emergent team state, psychological safety, appear to be
positively related. To finalize the univariate analysis, we check for the presence of
multicollinearity. Since the correlation values are lower than 0.8 and the variance inflation
factor (VIF) of each variable is lower than the recommended cutoff of 10 (highest value of
VIF is 2.38), multicollinearity is not a problem in our study (Gujarati, 1995; Mansfield and
Helms, 1982).
-------------------------------------------------------------------------------------------------------------Insert Table I about here
------------------------------------------------------------------------------------------------------------Prior to testing the moderated mediation model, we test hypothesis H1. Results in Table II
confirm that SEW separation has a significant negative effect on TMT decision making
quality (ẞ = -0.440, p < .05). In this study, the main focus is on how and when SEW
23
separation has an effect on TMT decision making quality. Therefore, we estimate a simple
mediation model to test H2 followed by a moderated mediation model to test H3. For both
steps, we apply the PROCESS codes of Hayes (2013). These codes test for statistically
significant effects through the use of bootstrapping methods to avoid power problems that
result from asymmetric and other non-normal sampling distributions of an indirect effect,
while also being able to probe the significance of conditional indirect effects at different
values of our moderator variable.
----------------------------------------------------------------------------------------------------------------Insert Table II about here
----------------------------------------------------------------------------------------------------------------The results of the simple mediation model to test H2 are shown in Table III. H2 states that
behavioral integration mediates the relationship between SEW separation and TMT decision
making quality. Table III shows that H2 can be supported. The indirect effect of SEW
separation on decision making quality through behavioral integration is confirmed by the
bootstrap results as the bootstrapped 95% confidence interval around the indirect effect does
not contain zero (-0.625, -0.0099).
----------------------------------------------------------------------------------------------------------------Insert Table III about here
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Insert Table IV about here
----------------------------------------------------------------------------------------------------------------The results of the moderated mediation model to test H3 are presented in Table IV. We mean
centered the interaction variables (SEW separation, behavioral integration and psychological
safety). This would facilitate substantive interpretation of the interaction effects (Kam and
Franzese, 2007). Table IV reveals that the interaction term obtained by multiplying SEW
separation and psychological safety is positive and significant (ẞ = 2.126, p < .01). We
further examined the conditional indirect effect of SEW separation on decision making
24
quality through behavioral integration at three values of psychological safety: the mean
(3.9471) as well as one standard deviation above (4.3264) and below (3.5678) the mean.
Bootstrap results at a 95% confidence interval around the indirect effect do not contain zero (0.951, -0.307) at a value of psychological safety of 3.5678 (one below the mean). This implies
a significant conditional indirect effect of psychological safety. At the other two values of
psychological safety of 3.9471 (mean) and 4.3264 (one above the mean), both intervals do
contain zero which indicate insignificant conditional indirect effects of psychological safety.
Alternatively, as a robustness check, we also took the alternative operationalization of
separation, namely the mean Euclidean distance, into consideration (Harrison and Klein,
2007). The use of this separation measure confirms all results presented in Table II, III, and
IV (results not reported). These findings confirm the freedom of choice that Harrison and
Klein (2007) emphasize in choosing a separation measure due to conceptual and mathematical
similarities between standard deviation and mean Euclidean distance measures.
In order to complete the analysis and formulate a final conclusion about the moderated
mediation, we explore the conditional indirect effect through the use of the Johnson and
Neyman technique (Hayes, 2013) to detect the range of values of psychological safety for
which conditional indirect effects are statistically significant at a .05 level. Figure 2
graphically represents the conditional indirect effect as well as the upper and lower level 95%
confidence interval. The conditional indirect effect of SEW separation on decision making
quality through behavioral integration is significant when both upper and lower bounds of the
confidence interval are above (or below) the zero line. The figure shows that SEW separation
has a significant negative effect on decision making quality through behavioral integration
when the level of psychological safety is situated between 2.943 and 3.917. Furthermore,
SEW separation appears to have a significant and positive effect on TMT decision making
quality through behavioral integration when the level of psychological safety is situated
25
between 4.369 and 4.714. Looking at our sample, we see that 51% of the family firm TMTs is
characterized by a level of psychological safety situated in these ranges, with 40% in the
range between 2.943 and 3.917 and 10.91% in the 4.369 to 4.714 range. Within the range of
2.943 to 3.917, the negative effect is lessened when the level of psychological safety
increases. Within the range of 4.369 to 4.714, the positive effect increases when psychological
safety increases. For average values of psychological safety (3.917 to 4.369), it seems that
SEW separation no longer affects decision making quality through behavioral integration.
This means that when the level of psychological safety in the TMT is average, the sense of
feeling psychologically safe is sufficient to prevent value dissimilarities to have a negative
impact on TMT decision making quality through behavioral integration. Since 49% of the
TMTs in our sample are characterized by psychological safety values between 3.917 and
4.369, our hypothesis H3 is fully supported.
Taken all together, our results indicate that the marginal effect of SEW separation on
TMT decision making quality decreases when TMT members feel psychologically safer
within the team which provides evidence for our hypothesis. Furthermore, the moderating role
of psychological safety becomes stronger for average levels of psychological safety because
our results indicate that more SEW separation is no longer translated into lower decision
making quality through behavioral integration when the sense of psychological safety further
increases. Finally, the moderating role of psychological safety becomes even more dominant
for extremely large values of psychological safety because our results indicate that SEW
separation is translated into higher-quality decisions when the psychological safety climate is
very strong within the team.
----------------------------------------------------------------------------------------------------------------Insert Figure 2 about here
-----------------------------------------------------------------------------------------------------------------
26
DISCUSSION
This study examines the effect of socio-emotional wealth (SEW) separation of family firm top
management teams (TMT) on decision making quality. SEW separation is expected to
negatively influence decision making quality of family firms. In our study, we investigate
both how and when SEW separation affects the decision making quality of family firms. We
argued that SEW separation negatively affects the level of behavioral integration and decision
making quality. This negative separation effect can be mitigated by a climate of psychological
safety, where it is safe for team members to freely express their feelings and beliefs. By using
a moderated mediation model on a unique sample of 300 managers from 55 family firms, we
indeed found that TMT behavioral integration mediates the negative relation between SEW
separation and TMT decision making quality, which shows that opposing positions on SEW
between team members complicate cooperation and communication in the TMT. In addition,
our results reveal that the negative effect of SEW separation on behavioral integration, and
ultimately on decision making quality, is mitigated by psychological safety and even becomes
positive for high values of psychological safety.
This study is one of the few studies that focuses on family firm TMTs. Traditionally,
family firm research is concentrated on boards, where the agency perspective prevails
(Bammens et al., 2011). The limited attention to TMTs in family firm research is an important
gap in the field because many studies in the upper echelons tradition show that TMTs exert a
strong effect on firm decision making processes and outcomes (Finkelstein et al., 2009). In
this paper, we have focused on TMT members’ positions on the extent to which the goal of
the firm is to preserve socio-emotional wealth (SEW). The preservation of SEW has become a
dominant theme in family firm research as it captures an essential trait of family firms. We
build on the recent developments in SEW research that there exists ‘within firm variation’ in
the salience of SEW preservation (Berrone et al., 2012). This variation is not accurately
27
captured by the family-nonfamily member ratio of the TMT, because family members differ
in the extent to which they want to preserve SEW, and nonfamily members can also indicate
strong positions towards SEW (Miller et al., 2014). This assumption is further confirmed by
some preliminary evidence, based on our data, indicating no significant relationship between
the percentage of nonfamily members in the TMT and SEW separation (results not reported).
Overall, our study contributes to the vivid SEW debates in family firm research (e.g. Berrone
et al., 2012; Chua et al., 2015; Cruz and Arredondo, 2016; Debicki et al., 2016; Miller and Le
Breton-Miller, 2014) in two ways. First, we are one of the first studies to use a direct SEW
measure, based on the FIBER model of Berrone et al. (2012). Our results indicate that family
managers express a stronger SEW preference on average than nonfamily managers. These
findings give some support to the use of family-related proxy measures of SEW preservation
(e.g. family ties or ownership), even though the explanatory power of direct SEW measures is
much larger. Second, we shed empirical light to the conceptual debate whether different SEW
dimensions may have different effects on decision making processes of family firms (e.g.
Berrone et al., 2012; Miller and Le Breton-Miller, 2014; Schulze and Kellermanns, 2015).
More concrete, our statistical tests suggest that SEW is a unitary construct composed of
different dimensions when measured in the context of family firm TMTs. The fact that we
find a different outcome than the multidimensional SEW operationalizations found in studies
focusing on family owners (Debicki et al., 2016; Hauck et al., 2016) suggests that context
matters in SEW operationalizations. Hence, this finding provides interesting insights to the
SEW conceptualization debate which calls for more research in a variety of contexts and
samples (e.g. Debicki et al., 2016).
We build on the diversity framework of Harrison and Klein (2007) and argue that
TMT diversity with respect to positions on SEW creates separation within the TMT that
lowers TMT behavioral integration and in turn the quality of TMT decision making. In
28
addition, we argue that creating a team climate of psychological safety can reduce the
negative effects of SEW dissimilarities. To our knowledge, this is the first study that
systematically analyses how and when differences in SEW salience within the family firm
TMT affect important team outcomes such as decision making quality. Including internal
team moderators such as the climate of psychological safety is valuable because it is
amenable for managerial design (Priem et al., 1999). Creating the right team climate can
prevent the firm from costly interventions such as changing TMT composition. Furthermore,
psychological safety may determine the nature of the fiduciary relationship between TMT
members within family firms. Cruz et al. (2010) mentioned that the interplay between trust
and behavioral uncertainty in family firm TMTs explains the nature of control mechanisms
used in family firm TMTs, with lower levels of trust leading to behavioral uncertainty and
consequently more explicit controlling features in TMT contracts. We complement these
findings by introducing psychological safety as a trust-based climate (Edmondson, 1999) that
may influence the level of behavioral uncertainty in the team and in turn the incentives and
mechanisms of control installed in the family firm TMT. While Cruz et al. (2010) stressed the
importance of trust between the CEO and the top managers at an individual level, we
introduce psychological safety as team-level feature to influence the level of protective or
controlling features in TMT contracts. Figure 2 gives us an indication of the powerful effect
of creating a conducive team climate to address SEW differences in family firm TMTs. This
figure shows that at very high levels of psychological safety, the effect of SEW separation on
team outcomes becomes positive, which indicates that with the right team climate, SEW
dissimilarities may not be considered as being negative for team functioning but can even
enhance team outcomes.
29
Practical implications
Top management teams can face all sorts of problems and one of them is differing positions
or opinions on team- or firm-related issues among team members and their consequences. In
the specific family firm context, it is important to notice that the opinion concerning SEW
salience can be different for each individual, family or nonfamily member. In order to cope
with the negative effects of SEW dissimilarities, (family) firms can focus on homogeneity in
opinions on SEW among team members when recruiting new managers. However, it takes a
while to discover these traits which makes this solution challenging to achieve. In our study,
we show that another attempt to tackle the disrupting forces of SEW differences may be the
creation of a psychologically safe climate in the TMT. This sense of psychological safety can
be created or improved by a set of team structural features (team size, clear team goals and
adequate resources, information and rewards), and a leader that focuses on aspects like
coaching and interpersonal relationships among team members (Edmondson, 1999;
Edmondson and Lei, 2014). Literature indicates that important CEO traits influence
psychological safety (e.g. Carmeli et al., 2009; Kearney and Gebert, 2009). CEO dominance
is negative for psychological safety because a dominant CEO will be rather individualistic
without taking into account the different views and opinions of fellow team members in the
decision making process. CEO relational leadership however will have a positive impact on
the creation of psychological safety in the TMT. Instead of being dominant, CEOs have to
focus on building and nurturing social bonds and promote sincere team behavior such that
different opinions and uniqueness of each team member are accepted and respected (Carmeli
et al., 2009; Kearney and Gebert, 2009). Furthermore, important team traits can also predict
psychological safety (Edmondson and Lei, 2014). First, TMT size is negatively related to a
psychologically safe TMT climate such that large teams have more difficulties in achieving
and maintaining this specific team context. Second, TMT tenure differences are also
30
negatively correlated with psychological safety such that too many differences in team tenure
through a highly diversified mixture of senior and junior managers leads to risks of, for
instance, conflict or dominance by the seniors which again lowers the psychological safety
within the TMT.
Limitations and future research
Our research has some limitations that also provide interesting avenues for future research.
First, Harrison and Klein (2007) differentiate between three types of diversity. In this paper,
we focus on diversity as separation. Future research on family firm TMTs could also
investigate diversity as disparity. Common features within family firm TMTs that create
disparity are related to ownership concentration or status differences between family and
nonfamily members. Typically, ownership in family firms is often concentrated in the hands
of one or a few family members. Additionally, family firm TMTs are composed of managers
with family ties as well as managers without. Disparity creates power hierarchies where
formal ownership power or familial status overrules expertise during TMT decision making
(Ibarra, 1992). We assert that family firm research will profit from studies that take into
account such power/status differences in family firm TMTs. For instance, Patel and Cooper
(2014) found that greater equality in structural power across family and nonfamily members
increases performance of family firms. Future studies could focus on how disparity in
ownership or status in family firm TMTs affects team performance through the negative
impact on team processes such as behavioral integration and trust (e.g. Cruz et al., 2010;
Minichilli et al., 2010). Furthermore, it could be interesting to search for contextual variables
that can mitigate the negative effect of disparity on team processes and performance.
Second, this study stresses the importance of psychological safety as a contextual
factor that moderates the relation between SEW separation and team performance. The
creation of a climate of psychological safety is crucial to address the potential integration
31
problems between team members, and their subsequent negative effect on decision making
quality. It is therefore interesting to study the determinants of a psychologically safe climate
in TMTs. Edmondson (1999) has already indicated that psychological safety can be created or
improved by a set of team features such as a (team) leader that values and focuses on
interpersonal relationships among team members. In addition to our research, one can thus
examine specific CEO personality traits that are believed to affect the climate in a TMT. This
line of research fits in the emerging research stream of the ‘CEO-TMT interface’ relationship
(e.g. Klimoski and Koles, 2001; Peterson et al., 2003), where successful TMT performance
jointly depends on team and leader dynamics and their interactions (Ling et al., 2008).
Research on family firms will benefit from more studies that address CEO personality
characteristics, and how this affects CEO behavior towards TMTs. In general, family firm
studies that focus on determinants of team climate (emergent state) and the moderating effect
of such emergent states on TMT behavior (process) and performance seem a very promising
trajectory of future family firm research.
Third, our results show that SEW is an important trait of family firms, and that within
team differences in positions on SEW affect team processes and team outcomes. However,
there remains much work to do on capturing the consequences of SEW diversity on family
firm behavior. For example, a useful extension of this paper would be the consideration
whether SEW separation helps or hinders financial performance by examining the relationship
between TMT decision making quality and firm performance of family firms. When
investigating this relationship, future researchers should consider exogenous factors such as
industrial, strategic, and environmental factors (Delaney and Huselid, 1996; Hough and
White, 2003).
Fourth, another important feature to take into account is the existence of extreme
situations such as badly performing family firms. Several studies on SEW have revealed that
32
SEW preferences can change when individuals are confronted with extreme situations
(Chrisman and Patel, 2012; Gomez-Mejia et al., 2011; Gomez‐Mejia et al., 2014). We
speculate that SEW separation may change in family firm TMTs that encounter adverse
financial results. Since the family firms in our sample were not confronted with such extreme
situations, future studies should take these extreme situations into account when investigating
the relationship between TMT decision processes and firm performance.
CONCLUSION
To summarize, our study shows that the negative influence of SEW separation on team
outcomes in family firm TMTs can be tackled by creating and maintaining a psychologically
safe team climate. The dark side of separation can even become a bright side at high levels of
psychological safety. This implies that dissimilar positions on SEW can become an asset for a
family firm’s TMT if the right team climate is created. We conclude this by testing a
moderated mediation model using a unique sample of 300 managers working in 55 private
family firms. Adding to recent discussions in both family firm and TMT literature, our study
provides interesting implications for theory and practice as well as offering future researchers
some promising research avenues.
33
REFERENCES
Allen, N. J., Stanley, D. J., Williams, H. M. and Ross, S. J. (2007). Assessing the impact of
nonresponse on work group diversity effects. Organizational Research Methods, 10,
262-286.
Amason, A. C. (1996). Distinguishing The Effects Of Functional and Dysfunctional Conflict
On Strategic Decision Making: Resolving A Paradox For Top Management Teams
Academy of Management Journal, 39, 123-148.
Anderson, N. R. and West, M. A. (1998). Meauring climate for work group innovation:
Development and validation of the team climate inventory. Journal of Organizational
Behavior, 19, 235.
Bammens, Y., Voordeckers, W. and Van Gils, A. (2011). Boards of Directors in Family
Businesses: A Literature Review and Research Agenda. International Journal of
Management Reviews, 13, 134-152.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Prentice-Hall, Inc.
Barrick, M. R., Bradley, B. H., Kristof-Brown, A. L. and Colbert, A. E. (2007). The
Moderating Role of Top Management Team Interdependence: Implications for Real
Teams and Working Groups. Academy of Management Journal, 50, 544-557.
Bell, S. T. (2007). Deep-Level Composition Variables as Predictors of Team Performance: A
Meta-Analysis. Journal of Applied Psychology, 92, 595-615.
Bell, S. T., Villado, A. J., Lukasik, M. A., Belau, L. and Briggs, A. L. (2011). Getting
Specific about Demographic Diversity Variable and Team Performance Relationships:
A Meta-Analysis. Journal of Management, 37, 709-743.
Bentler, P. M. and Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis
of covariance structures. Psychological bulletin, 88, 588.
Berrone, P., Cruz, C. and Gomez-Mejia, L. R. (2012). Socioemotional Wealth in Family
Firms: Theoretical Dimensions, Assessment Approaches, and Agenda for Future
Research. Family Business Review, 25, 258-279.
Bettinelli, C. (2011). Boards of Directors in Family Firms: An Exploratory Study of Structure
and Group Process. Family Business Review, 24, 151-169.
Biernacki, P. and Waldorf, D. (1981). Snowball sampling: Problems and techniques of chain
referral sampling. Sociological methods & research, 10, 141-163.
Björnberg, Å. and Nicholson, N. (2012). Emotional Ownership: The Next Generation’s
Relationship With the Family Firm. Family Business Review, 25, 374-390.
Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications
for data aggregation and analysis.
Boeker, W. (1997). Strategic Change: The Influence Of Managerial Characteristics And
Organizational Growth. Academy of Management Journal, 40, 152-170.
Boone, C. and Hendriks, W. (2009). Top Management Team Diversity and Firm
Performance: Moderators of Functional-Background and Locus-of-Control Diversity.
Management Science, 55, 165-180.
Bowman, C. and Ambrosini, V. (1997). Using single respondents in strategy research. British
Journal of Management, 8, 119-131.
Bromiley, P. (2009). A prospect theory model of resource allocation. Decision Analysis, 6,
124-138.
Bromiley, P. (2010). Looking at prospect theory. Strategic Management Journal, 31, 13571370.
Browne, M. W., Cudeck, R., Bollen, K. A. and Long, J. S. (1993). Alternative ways of
assessing model fit. Sage focus editions, 154, 136-136.
34
Burke, C. S., Stagl, K. C., Salas, E., Pierce, L. and Kendall, D. (2006). Understanding team
adaptation: a conceptual analysis and model. Journal of Applied Psychology, 91, 1189.
Buyl, T., Boone, C., Hendriks, W. and Matthyssens, P. (2011). Top Management Team
Functional Diversity and Firm Performance: The Moderating Role of CEO
Characteristics. Journal of Management Studies, 48, 151-177.
Byrne, D. (1971). The attraction paradigm. New York: Academic press.
Cannella, A. A. and Holcomb, T. R. (2005). A multi-level analysis of the upper-echelons
model. Research in multi-level issues, 4, 197-237.
Carmeli, A., Ben-Hador, B., Waldman, D. A. and Rupp, D. E. (2009). How Leaders Cultivate
Social Capital and Nurture Employee Vigor: Implications for Job Performance.
Journal of Applied Psychology, 94, 1553-1561.
Carmeli, A. and Gittell, J. H. (2009). High-quality relationships, psychological safety, and
learning from failures in work organizations. Journal of Organizational Behavior, 30,
709-729.
Carmeli, A. and Halevi, M. Y. (2009). How top management team behavioral integration and
behavioral complexity enable organizational ambidexterity: The moderating role of
contextual ambidexterity. Leadership Quarterly, 20, 207-218.
Carmeli, A. and Schaubroeck, J. (2006). Top management team behavioral integration,
decision quality, and organizational decline. Leadership Quarterly, 17, 441-453.
Carmeli, A., Tishler, A. and Edmondson, A. C. (2012). CEO relational leadership and
strategic decision quality in top management teams: The role of team trust and
learning from failure. Strategic Organization, 10, 31-54.
Carpenter, M. M. A. C., Geletkanycz, M. M. A. G. and Sanders, W. W. G. S. (2004). Upper
Echelons Research Revisited: Antecedents, Elements, and Consequences of Top
Management Team Composition. Journal of Management, 30, 749-778.
Carton, A. M. and Cummings, J. N. (2012). A Theory of Subgroups in Work Teams.
Academy of Management Review, 37, 441-470.
Certo, S. T., Lester, R. H., Dalton, C. M. and Dalton, D. R. (2006). Top Management Teams,
Strategy and Financial Performance: A Meta-Analytic Examination. Journal of
Management Studies, 43, 813-839.
Chang, S.-J., Van Witteloostuijn, A. and Eden, L. (2010). From the editors: Common method
variance in international business research. Journal of International Business Studies,
41, 178-184.
Cheung, G. W. and Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing
measurement invariance. Structural equation modeling, 9, 233-255.
Chrisman, J. J. and Patel, P. C. (2012). Variations in R&D investments of family and
nonfamily firms: Behavioral agency and myopic loss aversion perspectives. Academy
of Management Journal, 55, 976-997.
Chua, J. H., Chrisman, J. J. and De Massis, A. (2015). A Closer Look at Socioemotional
Wealth: Its Flows, Stocks, and Prospects for Moving Forward. Entrepreneurship
Theory and Practice, 39, 173-182.
Chua, J. H., Chrisman, J. J. and Sharma, P. (1999). Defining the family business by behavior.
Entrepreneurship: Theory & Practice, 23, 19-39.
Colbert, A. E., Kristof-Brown, A. L., Bradley, B. H. and Barrick, M. R. (2008). CEO
transformational leadership: The role of goal importance congruence in top
management teams. Academy of Management Journal, 51, 81-96.
Cruz, C. and Arredondo, H. (2016). Going back to the roots of socioemotional wealth.
Management Research: Journal of the Iberoamerican Academy of Management, 14,
234-243.
35
Cruz, C. C., Gomez-Mejia, L. R. and Becerra, M. (2010). Perceptions of benevolence and the
design of agency contracts: CEO-TMT relationships in family firms. Academy of
Management Journal, 53, 69-89.
Cyert, R. M. and March, J. G. (1963). A Behavioral Theory of the Firm. Englewood Cliffs,
NJ: Prentice Hall.
Dahlin, K. B., Weingart, L. R. and Hinds, P. J. (2005). Team Diversity and Information Use.
Academy of Management Journal, 48, 1107-1123.
Daily, C. M. and Dollinger, M. J. (1993). Alternative methodologies for identifying familyversus nonfamily-managed businesses. Journal of Small Business Management, 31,
79.
Debicki, B. J., Kellermanns, F. W., Chrisman, J. J., Pearson, A. W. and Spencer, B. A.
(2016). Development of a socioemotional wealth importance (SEWi) scale for family
firm research. Journal of Family Business Strategy, 7, 47-57.
Delaney, J. T. and Huselid, M. A. (1996). The impact of human resource management
practices on perceptions of organizational performance. Academy of Management
Journal, 39, 949-969.
Dess, G. G. (1987). Consensus on strategy formulation and organizational performance:
Competitors in a fragmented industry. Strategic Management Journal, 8, 259-277.
Dooley, R. S. and Fryxell, G. E. (1999). Attaining Decision Quality and Commitment From
Dissent: The Moderating Effects of Loyalty and Competence in Strategic Decision
making Teams Academy of Management Journal, 42, 389-402.
Edmondson, A. (1999). Psychological Safety and Learning Behavior in Work Teams.
Administrative Science Quarterly, 44, 350-383.
Edmondson, A. C. and Lei, Z. (2014). Psychological safety: The history, renaissance, and
future of an interpersonal construct. Annu. Rev. Organ. Psychol. Organ. Behav., 1, 2343.
Eisenhardt, K. M. and Bourgeois, L. J. (1988). Politics of Strategic Decision Making in HighVelocity Environments: Toward a Midrange Theory. Academy of Management
Journal, 31, 737-770.
Ellis, A. P. J., Mai, K. M. and Christian, J. S. (2013). Examining the Asymmetrical Effects of
Goal Faultlines in Groups: A Categorization-Elaboration Approach. Journal of
Applied Psychology, 98, 948-961.
Farrington, S. M., Venter, E. and Boshoff, C. (2012). The Role of Selected Team Design
Elements in Successful Sibling Teams. Family Business Review, 25, 191-205.
Fiegener, M. K., Brown, B. M., Prince, R. A. and File, K. M. (1996). Passing on strategic
vision: Favored modes of successor preparation by CEOs of family and nonfamily
firms. Journal of Small Business Management, 34, 15.
Finkelstein, S., Hambrick, D. C. and Cannella, A. A. (2009). Strategic leadership: theory and
research on executives, top management teams and boards. Minneapolis Oxford
University Press.
Fornell, C. and Larcker, D. F. (1981). Structural equation models with unobservable variables
and measurement error: Algebra and statistics. Journal of marketing research, 382388.
Gibson, C. and Vermeulen, F. (2003). A Healthy Divide: Subgroups as a Stimulus for Team
Learning Behavior. Administrative Science Quarterly, 48, 202-239.
Gibson, C. B. and Gibbs, J. L. (2006). Unpacking the Concept of Virtuality: The Effects of
Geographic Dispersion, Electronic Dependence, Dynamic Structure, and National
Diversity on Team Innovation. Administrative Science Quarterly, 51, 451-495.
36
Gomez-Mejia, L. R., Cruz, C., Berrone, P. and De Castro, J. O. (2011). The Bind that Ties:
Socioemotional Wealth Preservation in Family Firms. The Academy of Management
Annals, 5, 653-707.
Gomez-Mejia, L. R., Haynes, K. T., Núñez-Nickel, M., Jacobson, K. J. L. and MoyanoFuentes, J. (2007). Socioemotional wealth and business risks in family-controlled
firms: evidence from spanish olive oil mills. Administrative Science Quarterly, 52,
106-137.
Gomez‐Mejia, L. R., Campbell, J. T., Martin, G., Hoskisson, R. E., Makri, M. and Sirmon,
D. G. (2014). Socioemotional wealth as a mixed gamble: Revisiting family firm R&D
investments with the behavioral agency model. Entrepreneurship Theory and
Practice, 38, 1351-1374.
González-Romá, V., Fortes-Ferreira, L. and Peiró, J. M. (2009). Team climate, climate
strength and team performance. A longitudinal study. Journal of Occupational &
Organizational Psychology, 82, 511-536.
Gujarati, D. N. (1995). Basic Econometrics 3 edition. New York: McGraw-Hill.
Guth, W. D. and MacMillan, I. C. (1986). Strategy implementation versus middle
management self‐interest. Strategic Management Journal, 7, 313-327.
Hambrick, D. C. (1994). Top Management Groups: A Conceptual Integration and
Reconsideration of the "Team" Label. Research in Organizational Behavior, 16, 171.
Hambrick, D. C. (1995). Fragmentation and the Other Problems CEOs Have with Their Top
Management Teams. California Management Review, 37, 110-127.
Hambrick, D. C. and Brandon, G. L. (1988). Executive values. Elsevier Science/JAI Press.
Hambrick, D. C. and Mason, P. A. (1984). Upper Echelons: The Organization as a Reflection
of Its Top Managers. Academy of Management Review, 9, 193-206.
Harrison, D. A. and Klein, K. J. (2007). What's The Difference? Diversity Constructs as
Separation, Variety, or Disparity in Organizations Academy of Management Review,
32, 1199-1228.
Harrison, D. A., Price, K. H., Gavin, J. H. and Florey, A. T. (2002). Time, Teams, And Task
Performance: Changing Effects Of Surface - And Deep-Level Diversity On Group
Functioning. Academy of Management Journal, 45, 1029-1045.
Hauck, J., Suess-Reyes, J., Beck, S., Prügl, R. and Frank, H. (2016). Measuring
socioemotional wealth in family-owned and-managed firms: A validation and short
form of the FIBER Scale. Journal of Family Business Strategy, 7, 133-148.
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. Guilford Press.
Hogue, A. and Steinberg, L. (1995). Homophily of internalized distress in adolescent peer
groups. Developmental psychology, 31, 897.
Homberg, F. and Bui, H. T. M. (2013). Top Management Team Diversity: A Systematic
Review. Group & Organization Management, 38, 455-479.
Homburg, C., Krohmer, H. and Workman Jr, J. P. (1999). Strategic consensus and
performance: The role of strategy type and market-related dynamism. Strategic
Management Journal, 339-357.
Hough, J. R. and White, M. A. (2003). Environmental Dynamism And Strategic Decision
making Rationality: An Examination At The Decision Level. Strategic Management
Journal, 24, 481.
Huybrechts, J., Voordeckers, W. and Lybaert, N. (2013). Entrepreneurial Risk Taking of
Private Family Firms: The Influence of a Nonfamily CEO and the Moderating Effect
of CEO Tenure. Family Business Review, 26, 161-179.
37
Iaquinto, A. L. and Fredrickson, J. W. (1997). Research Notes and Communications Top
Management Team Agreement about the Strategic Decision Process: A Test of Some
of Its Determinants and Consequences. Strategic Management Journal, 18, 63-75.
Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure
and access in an advertising firm. Administrative Science Quarterly, 422-447.
James, L. R., Demaree, R. G. and Wolf, G. (1993). An Assessment of Within-Group
Interrater Agreement. Journal of Applied Psychology, 78, 306-309.
Janis, I. L. (1972). Victims of groupthink: a psychological study of foreign-policy decisions
and fiascoes.
Jehn, K. A. and Mannix, E. A. (2001). The Dynamic Nature of Conflict: A Longitudinal
Study of Intragroup Conflict and Group Performance. Academy of Management
Journal, 44, 238-251.
Joshi, A. and Roh, H. (2009). The Role of Context in Work Team Diversity Research: A
Meta-Analytic Review. Academy of Management Journal, 52, 599-627.
Kahn, W. A. (1990). Psychological Conditions of Personal Engagement and Disengagement
At Work. Academy of Management Journal, 33, 692-724.
Kam, C. D. and Franzese, R. J. (2007). Modeling and interpreting interactive hypotheses in
regression analysis. Ann Arbor: University of Michigan Press.
Kearney, E. and Gebert, D. (2009). Managing Diversity and Enhancing Team Outcomes: The
Promise of Transformational Leadership. Journal of Applied Psychology, 94, 77-89.
Kearney, E., Gebert, D. and Voelpel, S. C. (2009). When And How Diversity Benefits
Teams: The Importance Of Team Members' Need For Cognition Academy of
Management Journal, 52, 581-598.
Kellermanns, F. W., Walter, J., Lechner, C. and Floyd, S. W. (2005). The lack of consensus
about strategic consensus: Advancing theory and research. Journal of Management,
31, 719-737.
Klimoski, R. J. and Koles, K. L. K. (2001). The chief executive officer and top management
team interface. The nature of organizational leadership: Understand the performance
imperatives confronting today’s leaders, 219-269.
Korsgaard, M. A., Schweiger, D. M. and Sapienza, H. J. (1995). Building Commitment,
Attachment, and Trust in Strategic Decision making Teams: The Role of Procedural
Justice Academy of Management Journal, 38, 60-84.
Lau, D. C. and Murnighan, J. K. (2005). Interactions Within Groups and Subgroups: The
Effects Of Demographic Faultlines. Academy of Management Journal, 48, 645-659.
Lawrence, B. S. (1997). The Black Box of Organizational Demography. Organization
Science, 8, 1-22.
Lee, R. M. (1993). Doing research on sensitive topics. Sage.
Liang, T.-P., Wu, J. C.-H., Jiang, J. J. and Klein, G. (2012). The impact of value diversity on
information system development projects. International Journal of Project
Management, 30, 731-739.
Lindell, M. K. and Whitney, D. J. (2001). Accounting for common method variance in crosssectional research designs. Journal of Applied Psychology, 86, 114.
Ling, Y. and Kellermanns, F. W. (2010). The Effects of Family Firm Specific Sources of
TMT Diversity: The Moderating Role of Information Exchange Frequency. Journal of
Management Studies, 47, 322-344.
Ling, Y., Simsek, Z., Lubatkin, M. H. and Veiga, J. F. (2008). Transformational Leadership's
Role in Promoting Corporate Entrepreneurship: Examining The CEO-TMT Interface
Academy of Management Journal, 51, 557-576.
38
Locke, K. D. and Horowitz, L. M. (1990). Satisfaction in interpersonal interactions as a
function of similarity in level of dysphoria. Journal of personality and social
psychology, 58, 823.
Mansfield, E. R. and Helms, B. P. (1982). Detecting multicollinearity. American Statistician,
36, 158.
Marks, M. A., Mathieu, J. E. and Zaccaro, S. J. (2001). A Temporally Based Framework and
Taxonomy of Team Processes Academy of Management Review, 26, 356-376.
Miller, D. and Le Breton-Miller, I. (2014). Deconstructing Socioemotional Wealth.
Entrepreneurship: Theory & Practice, pp. 713-720.
Minichilli, A., Corbetta, G. and MacMillan, I. C. (2010). Top management teams in familycontrolled companies: ‘familiness’, ‘faultlines’, and their impact on financial
performance. Journal of Management Studies, 47, 205-222.
Mustakallio, M., Autio, E. and Zahra, S. (2002). Relational and Contractual Governance in
Family Firms: Effects on Strategic Decision Making. Family Business Review, 15,
205-222.
Nielsen, S. (2010). Top Management Team Diversity: A Review of Theories and
Methodologies. International Journal of Management Reviews, 12, 301-316.
O'Reilly, C. A. and Roberts, K. H. (1976). Relationships Among Components of Credibility
and Communication Behaviors in Work Units. Journal of Applied Psychology, 61, 99102.
Olson, B. J., Parayitam, S. and Bao, Y. (2007). Strategic decision making: The effects of
cognitive diversity, conflict, and trust on decision outcomes. Journal of Management,
33, 196-222.
Papadakis, V. M., Lioukas, S. and Chambers, D. (1998). Strategic decision making
processes: The role of management and context. Strategic Management Journal, 19,
115.
Patel, P. C. and Cooper, D. (2014). Structural Power Equality Between Family and NonFamily TMT Members and The Performance of Family Firms. Academy of
Management Journal, 57, 1624-1649.
Peterson, R. S., Martorana, P. V., Smith, D. B. and Owens, P. D. (2003). The Impact of Chief
Executive Officer Personality on Top Management Team Dynamics: One Mechanism
by Which Leadership Affects Organizational Performance. Journal of Applied
Psychology, 88, 795.
Pitcher, P. and Smith, A. D. (2001). Top Management Team Heterogeneity: Personality,
Power, and Proxies. Organization Science, 12, 1-18.
Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y. and Podsakoff, N. P. (2003). Common
method biases in behavioral research: a critical review of the literature and
recommended remedies. Journal of Applied Psychology, 88, 879.
Podsakoff, P. M., MacKenzie, S. B. and Podsakoff, N. P. (2012). Sources of method bias in
social science research and recommendations on how to control it. Annual Review of
Psychology, 63, 539-569.
Priem, R. L., Lyon, D. W. and Dess, G. G. (1999). Inherent limitations of demographic
proxies in top management team heterogeneity research. Journal of Management, 25,
935-953.
Rapert, M. I., Velliquette, A. and Garretson, J. A. (2002). The strategic implementation
process: evoking strategic consensus through communication. Journal of Business
Research, 55, 301-310.
Saunders, M., Lewis, P. and Thornhill, A. (2007). Research methods for business students.
Pearson Education UK.
39
Schulze, W. S. and Kellermanns, F. W. (2015). Reifying socioemotional wealth.
Entrepreneurship Theory and Practice, 39, 447-459.
Schulze, W. S., Lubatkin, M. H. and Dino, R. N. (2003). Exploring the agency consequences
of ownership dispersion among the directors of private family firms Academy of
Management Journal, 46, 179-194.
Schweiger, D. and Sandberg, W. R. (1991). A team approach to top management's strategic
decisions. Handbook of business strategy, 6, 6-20.
Shepherd, N. G. and Rudd, J. M. (2014). The Influence of Context on the Strategic Decision
making Process: A Review of the Literature. International Journal of Management
Reviews, 16, 340-364.
Siemsen, E., Roth, A. and Oliveira, P. (2010). Common method bias in regression models
with linear, quadratic, and interaction effects. Organizational Research Methods, 13,
456-476.
Simons, T., Pelled, L. H. and Smith, K. A. (1999). Making Use of Difference: Diversity,
Debate, and Decision Comprehensiveness In Top Management Teams Academy of
Management Journal, 42, 662-673.
Simsek, Z., Veiga, J. F., Lubatkin, M. H. and Dino, R. N. (2005). Modeling The Multilevel
Determinants Of Top Management Team Behavioral Integration Academy of
Management Journal, 48, 69-84.
Smith, K. G., Smith, K. A., Sims Jr, H. P., O'Bannon, D. P., Scully, J. A. and Olian, J. D.
(1994). Top Management Team Demography and Process: The Role of Social
Integration and Communication. Administrative Science Quarterly, 39, 412-438.
Tagiuri, R. and Davis, J. (1996). Bivalent attributes of the family firm. Family Business
Review, 9, 199-208.
Van der Merwe, S. P. (2007). An exploratory study of some of the determinants of
harmonious family relationships in small and medium-sized family businesses.
Management Dynamics: Journal of the Southern African Institute for Management
Scientists, 16, 24-35.
van Knippenberg, D., De Dreu, C. K. W. and Homan, A. C. (2004). Work Group Diversity
and Group Performance: An Integrative Model and Research Agenda. Journal of
Applied Psychology, 89, 1008-1022.
van Knippenberg, D. and Schippers, M. C. (2007). Work Group Diversity. Annual Review of
Psychology, 58, 515-541.
Venter, E., Boshoff, C. and Maas, G. (2003). The influence of relational factors on successful
succession in family businesses: A comparative study of owner-managers and
successors. South African Journal of Business Management, 34, 1-13.
Wei, L.-Q. and Wu, L. (2013). What a Diverse Top Management Team Means: Testing an
Integrated Model. Journal of Management Studies, 50, 389-412.
West, M. A. and Anderson, N. R. (1996). Innovation in top management teams. Journal of
Applied Psychology, 81, 680.
Wiersema, M. F. and Bantel, K. A. (1992). Top Management Team Demography and
Corporate Strategic Change Academy of Management Journal, 35, 91-121.
Williams, K. Y. and O'Reilly, C. A. (1998). Demography and Diversity in Organizations: A
Review of 40 Years of Research Research in Organizational Behavior, 20, 77.
Woehr, D., Arciniega, L. and Poling, T. (2013). Exploring the Effects of Value Diversity on
Team Effectiveness. Journal of Business & Psychology, 28, 107-121.
Wooldridge, B. and Floyd, S. W. (1990). The strategy process, middle management
involvement, and organizational performance. Strategic Management Journal, 11,
231-241.
40
Zellweger, T. M., Nason, R. S., Nordqvist, M. and Brush, C. G. (2013). Why do family firms
strive for nonfinancial goals? An organizational identity perspective. Entrepreneurship
Theory and Practice, 37, 229-248.
Zenger, T. R. and Lawrence, B. S. (1989). Organizational demography: The differential
effects of age and tenure distributions on technical communication. Academy of
Management Journal, 32, 353-376.
41
Psychological
safety
SEW
separation
Behavioral
integration
Figure 1. Research model
42
TMT
decision making
quality
Figure 2. Conditional indirect effect of SEW separation on decision making quality through
behavioral integration
43
Table I. Descriptive statistics and pairwise correlations
Mean
SD
1
2
3
4
5
6
7
1 Decision making
quality
4.08
0.33
1
2 SEW separation
0.507
0.22
-0.297*
1
3 Psychological safety
3.95
0.38
0.726**
-0.275*
1
4 Behavioral integration 3.80
0.47
0.764**
-0.297*
0.718**
1
5 Firm size¥
374.16
699.62
0.229†
-0.055
0.138
0.209
1
6 Team size¥
5.45
1.87
-0.086
0.103
-0.135
-0.227†
0.466**
1
7 Age diversity
0.162
0.075
-0.058
0.011
-0.117
-0.175
0.110
0.163
1
8 Gender diversity
0.226
0.20
0.130
0.134
0.082
0.104
-0.095
0.147
0.101
N = 55 teams.
†, *, ** Correlation is significant at the 0.1 level, 0.05 level, 0.01 level (2-tailed).
¥ Natural logarithm used in regression model.
44
8
1
TABLE II. OLS regression results for the effect of SEW separation on decision making
quality
Model
b coeff
SE
t
Constant
4.190
0.242 17.303**
SEW separation
-0.440
0.198
-2.221*
Firm size
.094
0.038
2.486*
Team size
-0.239
0.141
-1.689†
Age diversity
-0.348
0.566
-0.614
Gender diversity
0.418
0.221
1.891†
R² = 0,226, F =2,854 , p = 0,024
N= 55 teams. Unstandardized regression coefficients are reported.
† p < .10. * p < .05. ** p < .01, two-tailed.
45
Table III. Regression results for simple mediation model of SEW separation on decision
making quality through behavioral integration
Model
b coeff
SE
t
Mediator variable model (DV = Behavioral integration)
4.246
0.318
13.351**
Constant
SEW separation
-0.572
0.260
-2.198*
Firm size
0.156
0.0497
3.146**
Team size
-0.548
0.186
-2.953**
Age diversity
-1.110
0.744
-1.492
Gender diversity
0.614
0.290
2.113*
R² = 0,319, F = 4,590, p = 0,0016
Dependent variable model (DV= Decision making quality)
1.927
0.376
5.121**
Constant
Behavioral integration
0.533
0.0785
6.793**
SEW separation
-0.135
0.150
-0.902
Firm size
0.0107
0.0299
0.359
Team size
0.0534
0.111
0.483
Age diversity
0.244
0.418
0.583
Gender diversity
0.0910
0.167
0.546
R² = 0,6052, F = 12,261, p = 0,000
Total, direct and indirect effects
Total effect of SEW on dmq
Effect
SE
t
LLCI
ULCI
-0.440
0.198
-2.221*
-0.838
-0.0419
Direct effect of SEW on dmq
Effect
SE
t
LLCI
ULCI
-0,135
0.150
-0.902
-0.437
0.166
Indirect effect of SEW on dmq
Effect
Boot SE
z
BootLLCI BootULCI
-0.305
0.158
-2.071*
-0.625
-0.0099
N= 55 teams. Unstandardized regression coefficients are reported. Bootstrap sample size = 10000.
LL = lower limit, UL = upper limit, CI = confidence interval.
† p < .10. * p < .05. ** p < .01, two-tailed.
46
Table IV. Regression results for moderated mediation model of SEW separation on decision
making quality through behavioral integration with psychological safety as moderator
Model
b coeff
SE
t
Mediator variable model (DV = Behavioral integration)
4.0231
0.202
19.878**
Constant
SEW separation
-0.351
0.184
-1.907†
Psychological safety
0.830
0.111
7.499**
SEW separation x
Psychological safety
2.126
0.543
3.912**
Firm size
0.0775
0.0350
2.214*
Team size
-0.277
0.130
-2.135*
Age diversity
-1.101
0.518
-2.126*
Gender diversity
0.299
0.200
1.496
R² = 0,704, F = 15,936, p = 0,000
Dependent variable model (DV= Decision making quality)
1.927
0.376
5.121**
Constant
Behavioral integration
0.533
0.0785
6.793**
SEW separation
-0.135
0.150
-0.902
Firm size
0.0107
0.0299
0.359
Team size
0.0534
0.111
0.483
Age diversity
0.244
0.418
0.583
Gender diversity
0.0910
0.167
0.546
R² = 0,6052, F = 12,261, p = 0,000
Conditional indirect effects of psychological safety
Psychological safety
Bootstrap indirect Bootstrap BootLLCI BootULCI
effect
SE
3,5678
-0.617
0.163
-0.951
-0.307
3,9471
-0.187
0.104
-0.401
0.0118
4,3264
0.243
0.146
-0.0351
0.555
N= 55 teams. Mean centered regression coefficients are reported. Bootstrap sample size = 10000.
LL = lower limit, UL = upper limit, CI = confidence interval.
† p < .10. * p < .05. ** p < .01, two-tailed
47