and Human Decision Processes 167 (2021) 28–41

Contents lists available at ScienceDirect

Organizational Behavior and Human Decision Processes

journal homepage: www.elsevier.com/locate/obhdp

A relational account of low power: The role of the attachment system in reduced

Jieun Pai a,*, Jennifer Whitson b, Junha Kim c, Sujin Lee d a Frank Batten School of Leadership and Public Policy, University of Virginia, 235 McCormick Road, Charlottesville, VA 22904, United States b The UCLA Anderson School of Management, 110 Westwood Plaza, Los Angeles, CA 90095, United States c College of Business, The Ohio State University, 2100 Neil Ave, Columbus, OH 43210, United States d College of Business, School of Business and Technology Management, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea

ARTICLE INFO ABSTRACT

Keywords: Although the literature provides widespread evidence for the effect of power on action, a systematic under­ Attachment theory standing of why low power individuals are less prone to action is still lacking. We focus on proactive behavior as Proactivity a particular form of action and propose a relational underpinning of the link between low power and reduced Power proactive behavior through the framework of attachment theory. We predict that the experience of low power will increase attachment anxiety, and that this increase in attachment anxiety will reduce proactive behavior. We test the proposed theory in a series of four pre-registered experiments and one quasi-experimental fieldstudy. In Study 1, 2, and 3, we test the mediating role of attachment anxiety on the relationship between low power and proactivity. In Study 4 and 5, we test theoretically driven attachment security interventions to mitigate the link between low power and reduced proactivity. We demonstrate that attachment security interventions help low power individuals to be more proactive in an organizational field quasi-experiment (Study 4) and in an exper­ iment (Study 5). Overall, we findsupport for attachment anxiety as an important factor in the experience of low power individuals which inhibits proactive behavior; we further offer a powerful intervention grounded in attachment theory to ameliorate this effect.

1. Introduction Exploratory actions function as a method via which individuals can proactively explore, learn about, and influencetheir environment (Elliot are rife with differences in power (i.e., the capacity to & Reis, 2003; Feeney & Thrush, 2010; Grant & Ashford, 2008; Hazan & influenceor control resources, rewards, punishments, and other people; Shaver, 1990; Noe, 1986). The need of modern organizations to Magee & Galinsky, 2008). Power is the basis of hierarchical differenti­ decentralize and innovate requires proactive behaviors from their em­ ation and is inherently relational. That is, power determines and is ployees (i.e., self-directed engagement with the environment; Campbell, determined by bonds between people who distribute resources (whether 2000; Frese & Fay, 2001; Frohman, 1997; Parker, 2000; Wu, 2019). tangible or intangible) and those who need them (De Waal & Waal, When employees do so, it can lead to improved sales performance 2007; Magee & Galinsky, 2008). Given the pyramid-shaped nature of (Crant, 1995), entrepreneurship (Becherer & Maurer, 1999), and inno­ hierarchies, the majority of organizational members will report to vation at the individual and firmlevel (Kickul & Gundry, 2002; Seibert, someone who has power over them, thus putting them in the position to Kraimer, & Crant, 2001). High power individuals’ capacity to control think, feel, and behave as low power individuals. For example, even valued resources reduces their dependence on others and affords more departmental directors will feel low power when interacting with C- proactive behaviors (Keltner, Gruenfeld, & Anderson, 2003), such as suite executives. The experience of low power can create challenges for initiating the first offer in negotiations (Magee, Galinsky, & Gruenfeld, the by inhibiting the very behaviors that are useful in 2007), changing their surroundings to fit their needs (Galinsky, driving organizational success – i.e., the self-initiated exploratory ac­ Gruenfeld, & Magee, 2003), and displaying risk- and pressure-tolerance tions that constitute proactive behavior. which allows them to persistently pursue beneficial outcomes even

* Corresponding author. E-mail addresses: [email protected] (J. Pai), [email protected] (J. Whitson), [email protected] (J. Kim), [email protected] (S. Lee). https://doi.org/10.1016/j.obhdp.2021.06.003 Received 25 July 2019; Received in revised form 20 May 2021; Accepted 14 June 2021 0749-5978/© 2021 Elsevier Inc. All rights reserved. J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 under less than ideal conditions (Anderson & Galinsky, 2006). In In this paper, we specifically focus on proactive behavior, as its contrast, low power individuals’ limited resources increase their exploratory nature allows us to utilize the lens of attachment theory to dependence on those who do have power and inhibit them from dis­ understand power dynamics (Wu, 2019; Wu, Parker, Wu, & Lee, 2018). playing proactive behaviors (Anderson & Galinsky, 2006; Galinsky We propose and demonstrate that the attachment system does indeed et al., 2003). function as an underlying mechanism of the relationship between low For organizations to thrive, low power members often need to power and reduced proactive action, and that interventions based on behave proactively, by pointing out organizational problems (Fang, attachment theory mitigate this effect, insulating individuals from the Kim, & Milliken, 2014; Kim & Kim, 2020; Milliken, Morrison, & Hewlin, action-discouraging effects of low power. 2003; Vuori & Huy, 2016), taking initiative in pursuing goals (Frese & We seek to make several contributions to theories of power. First, our Fay, 2001; Roberson, 1990), and actively adapting to new environments study builds on recent explorations of individuals at the bottom of the (Ashford & Black, 1996). However, low power members are less likely to hierarchy (e.g., Hays & Bendersky, 2015; Schaerer et al., 2018). As propose new ideas (Kim & Kim, 2020), to share negative, yet accurate demonstrated by Schaerer et al. (2018), extant theories, study designs, information with top managers (Detert & Burris, 2007; Fang et al., 2014; and attributions of results in social power research predominantly focus Milliken et al., 2003; Vuori & Huy, 2016), to be innovative (Zhang & on high power individuals, leaving us an insufficient understanding of Bartol, 2010), and to learn new skills, all of which risk leading to major low power individuals. Our work extends this line of research on the organizational failure (Vuori & Huy, 2016). These findingshighlight the unique system of responses of low power individuals and introduces the importance of understanding the underlying mechanisms of why low attachment system as a key explanatory factor. Second, organizational power employees are less prone to proactive behaviors when needed and research has long been interested in the proactive behaviors of organi­ what interventions can encourage them to behave proactively. zation members (see Grant & Ashford, 2008 for a review). Our research Nevertheless, much research has largely focused on the experience of carries implications for the effect of low power on key organization- high power (Anderson & Galinsky, 2006; Anderson & Thompson, 2004; relevant measures of proactivity. Specifically, many of the behaviors Galinsky et al., 2003; Galinsky, Ku, & Wang, 2005; Galinsky, Magee, that are often seen as desirable in employees may be muted during the Inesi, & Gruenfeld, 2006; Magee et al., 2007). This focus on high power experience of low power, for example, spontaneously pointing out er­ implies, by the lack of attention to low power, that low power might be rors, sharing innovative ideas, and employing voice. Third, we provide thought of as simply the absence of power. However, Schaerer, du interventions based on our theoretical framework that attachment se­ Plessis, Yap, and Thau (2018) have demonstrated that while low power curity can disinhibit low power employees and increase their ability to individuals sometimes behave in opposite ways than high power in­ be proactive. dividuals (e.g., they exhibit less exploratory proactive behaviors), at other times, they behave similarly (e.g., they objectify others to a similar 2. Theory and hypotheses extent). Congruently, Smith and Hofmann (2016) found that low power and high power often had different, and not just opposite effects. This 2.1. Low power and the behavioral inhibition system suggests that unique phenomena come into play only during the expe­ rience of low power and that it is inappropriate to infer the psycholog­ A key paradigm of power reasearch is based on the notion that the ical processes of low power individuals from those of high power experience of low power motivates individuals to avoid negative out­ individuals. The current research investigates the experience of low comes (Keltner et al., 2003), and that this motivation is responsible for power individuals, providing a relational lens through which to inter­ the behavioral inhibition which reduces action (Keltner et al., 2003). pret and predict their proactive behaviors (and lack thereof), as well as However, recent papers have suggested that low power individuals offering potential interventions to increase low power individuals’ exhibit behaviors for which the BIS is an insufficientexplanation (Case, proactive behaviors within organizations. Because power exists within Conlon, & Maner, 2015; Rucker & Galinsky, 2008, 2009; Schaerer et al., social relationships, we use attachment theory to elucidate a relational 2018; Waytz, Chou, Magee, & Galinsky, 2015). For example, the power underpinning for the link between low power and reduced proactive framework based on the BIS (Keltner et al., 2003) suggests low power behaviors. should increase inhibition and thus reduce objectification of others In this research, we put forth that the inherently relational nature of (Gruenfeld, Inesi, Magee, & Galinsky, 2008). However, a recent exam­ power suggests another system, composed of learned patterns of re­ ination demonstrated that both high and low power individuals show sponses built on individuals’ past relationships with others (Bretherton greater objectification of others, compared to those in a baseline con­ & Munholland, 2008). This attachment system is a regulatory system dition (Schaerer et al., 2018). There is also evidence that at times in­ which evolved to achieve the goal of felt-security (Bowlby, 1969, 1982; dividuals with low power are actually more likely to act than those with Bretherton, 1985). We suggest the attachment system provides impor­ high power, but only for certain behaviors. Interestingly, these actions tant mechanisms that illuminate the experiences of low power, and are all related to gaining social approval and acceptance. For example, integrate theories of power and attachment theory to offer a novel and low power individuals were more interested in friendship-facilitating cogent relational underpinning for the link between low power and gatherings, and positioned their chair physically closer to their reduced proactive behavior. We propose and discuss below that the assigned partner’s than high power individuals (Case et al., 2015). vulnerability of being in a low power position invokes the attachment Furthermore, low power consumers were more willing to pay for status- system as a coping strategy, specifically, attachment anxiety. Conse­ signaling products and preferred conspicuous logos on high-end clothing quently, the increase in attachment anxiety distracts from engaging in than high power consumers (Rucker & Galinsky, 2008, 2009; Rucker, exploratory actions (e.g., proactive workplace behaviors). Thus, Dubois, & Galinsky, 2010), all actions intended to increase social appeal attachment anxiety helps explain the link between low power and to others. reduced action (i.e., reduced proactivity). We further explore whether the effect of low power on proactive behavior is mitigated when 2.2. Low power and the attachment system attachment anxiety is reduced. It is worth noting that our approach is not intended to supplant those We argue that a crucial window into the psychology of low power is grounded in the approach/inhibition theory of power (Keltner et al., to consider it in terms of its social aspects. We suggest that low power is a 2003), but rather to complement them. Keltner et al. (2003) have threatening state in a social relationship, which invokes a learned sys­ theorized that low power individuals are less active due to the behav­ tem of responses derived from past relationships with others. This ioral inhibition system (BIS; Carver & White, 1994). Yet, it is still an learned system of responses, called the attachment system, focuses on open question whether the BIS accounts for all the effects of low power. how to achieve a greater felt-security (Pietromonaco & Barrett, 2000)

29 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 and provides an important relational mechanism to unpack the experi­ 2.3. Attachment anxiety as mechanism ence of low power. Low power has been portrayed as an aversive, threatening state, We predict that the threatening state of being low power will in­ compared to high power (Keltner et al., 2003; Magee & Galinsky, 2008). crease attachment anxiety. The attachment system has two orthogonal People with low power face a world of threats and uncertainty (Keltner continuous dimensions—attachment anxiety and attachment avoi­ et al., 2003), depend largely on the powerful to obtain rewards and dance—with low scores on both dimensions indicating attachment se­ avoid punishments (Emerson, 1962; Galinsky, Magee, Gruenfeld, curity (Ainsworth et al., 1978; Brennan, Clark, & Shaver, 1998; Whitson, & Liljenquist, 2008), and are often under threat of losing the Simpson, Rholes, & Phillips, 1996). When individuals are threatened, favor of powerful others (Chance, 1967; Keltner et al., 2003). Also, qualities of the situation that are similar to past experiences invoke a people in low power tend to perceive threats in ambiguous social in­ particular attachment representation to shape a response to the threat teractions (Schwartz, Dodge, & Coie, 1993). Even physiological evi­ (Hazan & Shaver, 1987; Hazan & Zeifman, 1999). We propose that dence suggests that experiencing low power is an aversive, threatening experiencing low power will increase attachment anxiety, but not state (e.g., lower levels of serotonin; Munafo et al., 2003). Because low attachment avoidance. Attachment anxiety results from an attachment power only exists in the context of a relationship with someone who has figurewho is inconsistent or not prompt in responding to one’s bids for the power to provide or withhold resources, low power is inherently approval, proximity, protection, and support (Ainsworth et al., 1978; relational. Brennan et al., 1998; Simpson et al., 1996). When state attachment As experiencing low power is an aversive and threatening state in a anxiety increases, individuals become more preoccupied with uncer­ relational context, we propose it will involve the attachment system tainty about whether their partner will accept versus reject them, worry (Bowlby, 1969, 1982). Attachment theory (Bowlby, 1969, 1982) con­ about losing their current partner, and remain vigilant to signs that their tends that humans are born with a system that, when they feel under partner could pull away (Mikulincer & Shaver, 2007). threat or in danger, drives individuals to seek proximity, comfort, and We argue that the experience of low power is threatening, and assistance from caring and supportive others (i.e., attachment figures). qualities of that situation share similarities with experiences of anxious This attachment system is a psychological regulatory system that governs attachment, therefore increasing attachment anxiety as individuals seek the employment of behavioral sequences aimed at attaining the goal of to cope with that threat. For example, low power individuals are often felt-security, a state that occurs when individuals are confidentthat their unaware of the exact reason they are being rewarded or punished attachment figureswill respond to their needs (Bretherton, 1985; Hazan (Butterfield,Trevino, & Ball, 1996), meaning they are more likely to feel & Shaver, 1987; Hazan & Zeifman, 1999). However, in an organization, that responses to their needs will be equally inconsistent. In fact, low it is rarely the case that attachment figures (e.g., peers, leaders, or power individuals generally experience higher uncertainty, lower con­ managers; Davidovitz, Mikulincer, Shaver, Izsak, & Popper, 2007; fidence,and greater dependence (Lammers, Dubois, Rucker, & Galinsky, Mayseless, 2010; Mayseless & Popper, 2007; Popper & Mayseless, 2007) 2013; Smith, Wigboldus, & Dijksterhuis, 2008), exaggerating for them are always available to provide felt-security. When low power in­ both the need for approval, protection, and support, as well as the dividuals cannot be sure the attachment figure will be available or concern that they will not receive it. Low power individuals also tend to predictably responsive to address experienced threats, they will expe­ be more uncertain about whether they would receive social support rience a sense of anxiety, which arouses the attachment system. when needed (Brion & Anderson, 2013). We therefore predict that the Despite its root in infant-caregiver relationships, the attachment experience of low power will increase attachment anxiety, as it will system is continuously updated to reflect individuals’ experiences with drive uncertainty with regard to whether approval, proximity, protec­ attachment figures (Mikulincer & Shaver, 2003). Specifically, in­ tion, and support will be promptly and consistently attained. dividuals accumulate vivid episodes and semantic information about In contrast, we expect the avoidance dimension of the attachment attachment figures throughout life. These figures include caregivers, system to be less relevant to power relationships. Attachment avoidance teachers, friends, romantic partners, coworkers, and managers, and are results from an attachment figure who is consistently unavailable, re­ used to create partner-specificmental representations which guide how jects, and does not provide support, security, protection, and approval. individuals think, feel, and behave in interpersonal contexts to attain the When state attachment avoidance increases, individuals orient away goal of felt-security (Ainsworth, Blehar, Waters, & Wall, 1978). from their partner and instead try to be self-reliant and independent to It is important to note that although individuals develop a general handle insecurity via social distancing (Mikulincer & Shaver, 2007). attachment representation, differing contexts and relationships make Low power individuals’ proximity-seeking efforts are unlikely to context-specificand partner-specificforms of attachment representation engender consistent rejection—the spur most relevant to the attachment temporarily and contextually more accessible (Baldwin, Keelan, Fehr, system’s avoidance dimension. Rather, low power individuals seeking Enns, & Koh-Rangarajoo, 1996; La Guardia, Ryan, Couchman, & Deci, reassurance are likely to feel they receive or are denied it at unpre­ 2000). In other words, interacting with a particular person or finding dictable intervals. As classical conditioning shows, random reinforce­ oneself in a particular situation can invoke specificforms of attachment ment creates some of the most persistent behaviors in pursuit of rewards representations. For instance, when individuals experience a situation (Papini & Bitterman, 1990). Furthermore, a drive for independence is where they need to protect themselves from threats and alleviate more relevant to attachment avoidance, and low power increases distress, this will prompt memories of similar past experiences and dependence on others as compared to baseline, rather than decreasing it motivate attempts at proximity seeking (Green & Campbell, 2000; (Rucker et al., 2010). The dependence of low power individuals on high Mikulincer & Shaver, 2007). Even experimentally manipulated cues can power individuals likely also makes distancing attempts appear even prompt a specific attachment dimension and influence in turn partici­ riskier. Thus, we expect power relationships not to activate the attach­ pants’ negotiation outcomes (Lee & Thompson, 2011) and ethical be­ ment system’s avoidance dimension. haviors (Chugh, Kern, Zhu, & Lee, 2014). These findingssuggest that the Hypothesis 1. Low power individuals exhibit a higher level of effect of the attachment system relevant to one relationship can extend attachment anxiety than baseline or high power individuals. beyond that relationship into other relationships and situations. In sum, the attachment system is relevant to and updated via varying social We further predict that low power individuals will show reduced relationships throughout life, and various context and relationship cues proactive behaviors because of increased attachment anxiety. Attach­ can temporarily invoke the attachment system. ment theory involves both attachment and exploration systems, which act in complementary balance with one another (Ainsworth et al., 1978; Bowlby, 1969, 1982; Elliot & Reis, 2003). Individuals with a more secure sense of attachment use attachment figuresas launching pads to

30 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 explore and learn about their environment as part of the exploration sense of felt-security, reduce attachment anxiety, and thus allow the system (Ainsworth et al., 1978; Bowlby, 1969, 1982; Hazan & Shaver, exploration system to move to the forefront. While this sense of felt- 1990). As threat increases, so too does the influence of the relevant security can be provided by the attachment figure who has power over attachment system, which suppresses the exploration system as in­ the low power individual, past research has also demonstrated that dividuals shift their focus to regaining felt-security. In other words, the people were more open to exploratory behaviors when attachment anxious attachment system distracts the individual from exploration as anxiety was reduced by thinking about secure relationships in another they become preoccupied with uncertainty with regard to the avail­ domain (Green & Campbell, 2000). Similarly, adults who felt securely ability of an attachment figure, which saps their confidence about attached to their spouses demonstrated higher persistence and better venturing out to engage with their environment (Fraley, Garner, & performance even in a relationship-irrelevant laboratory exploration Shaver, 2000; Hazan & Shaver, 1994). activity (Feeney & Thrush, 2010). This sense of support strengthens A widespread form of adulthood exploratory behavior involves work individuals’ perception of their own competence and motivation to and the workplace (Hazan & Shaver, 1990), particularly proactive be­ interact with the environment and thus helps them with exploration haviors (Wu & Parker, 2012, 2017; Wu et al., 2018). Proactive behaviors behaviors in other domains (Grossmann, Grossmann, Kindler, & Zim­ are often defined as “action that employees take to impact themselves mermann, 2008). Thus, reducing attachment anxiety by providing and/or their environments” (Grant & Ashford, 2008, p. 8), motivated by attachment security should allow low power individuals to exhibit the desire to actively control their environment (Bateman & Crant, increased levels of proactive behaviors. 1993). Congruently, Elliot and Reis (2003) tie exploration behaviors in We also suggest that this sense of security can come from a number of adulthood specifically to the, “desire for effective, competent in­ sources, including leaders and managers. Indeed, empirical work shows teractions with the environment,” which drives individuals to, “inves­ that leaders can serve as a secure base for their followers (Davidovitz tigate, manipulate, and master the environment” (p. 318). When et al., 2007; Mayseless & Popper, 2007; Mayseless, 2010; Popper & employees engage in proactive behaviors such as seeking feedback Mayseless, 2007). Specifically, certain types of leaders seem to foster (Ashford, Blatt, & VandeWalle, 2003; Ashford & Cummings, 1985) they attachment security among their employees (Molero, Moriano, & are investigating their environment by actively pursuing more infor­ Shaver, 2013), and good leader-member exchange relations are posi­ mation. Employees seeking to manipulate and master their environment tively related to attachment security (Richards & Hackett, 2012); em­ proactively take initiative in pursuing personal and organizational goals ployees who reported attachment security at work displayed less (Frese & Fay, 2001; Roberson, 1990), use their voice (LePine & Van burnout (Simmons et al., 2009) and more organizational citizenship Dyne, 1998, 2001), sell issues to colleagues (Dutton & Ashford, 1993), behaviors (Richards & Schat, 2011). Accordingly, we suggest that secure revise their tasks (Staw & Boettger, 1990), craft jobs (Wrzesniewski & support from a manager or supervisor will reduce state attachment Dutton, 2001), and manage their social networks to further their careers anxiety for low power individuals, increasing their proactive behaviors: (Morrison, 2002; Ostroff & Kozlowski, 1992). In other words, because Hypothesis 3. Low power individuals will be more proactive when they involve employees seeking out information and influencing their attachment security is made salient as opposed to not. environment, proactive behaviors are a form of adult exploratory behavior occurring in professional and organizational contexts. 3. Overview of studies Because proactive behaviors depend on the engagement of the exploration system, which is suppressed when attachment anxiety in­ In a series of four pre-registered experiments and one quasi- creases, they should also be reduced when attachment anxiety increases. experimental field study, we examine our hypothesized relationships. Indeed, past research has shown that individuals with high attachment First, to provide evidence that low power individuals are less proactive, anxiety tend to be distracted in the workplace, where they frequently and that this relationship is mediated by attachment anxiety, we worry about rejection for poor performance and are mostly motivated by manipulated each chain of the proposed mediation (Spencer, Zanna, & social acceptance and approval, congruent with their focus on attaining Fong, 2005). Specifically,in Study 1, we experimentally test the effects a sense of secure attachment (Hazan & Shaver, 1990; Richards & Schat, of low power on attachment anxiety, and in Study 2, we experimentally 2011; Simmons, Gooty, Nelson, & Little, 2009). Further, people with test the effects of attachment anxiety on proactivity. Next, in Study 3, we high attachment anxiety construed tasks as more of a negative threat integrate Study 1 and 2 and manipulate low power to examine its effects and less of a positive challenge, which predicted reduced levels of task on proactivity as mediated by attachment anxiety. Study 4 and 5 then motivation and performance (Elliot & Reis, 2003). Since this process of explore possible interventions to weaken the low power-reduced pro­ seeking security preoccupies attention and energy and hinders the activity link through state attachment security. In Study 4, we conduct a exploration system (Mikulincer & Shaver, 2003), we suggest that low quasi-experiment in a company fieldsetting with low power employees power individuals, compared to high power individuals, will be less (i.e., those with managers but without subordinates) and examine the likely to engage in proactive behaviors (e.g., using their voice, offering effects of a secure attachment intervention on proactive behavior. Study feedback, etc.). Thus, we hypothesize the following: 5 employs an interaction paradigm in which all participants are low Hypothesis 2. Attachment anxiety mediates the link between low power and interact with a high power counterpart who behaves in a power and reduced proactive behaviors. fashion congruent with either attachment-security- or anxiety-inducing figures. To lend support for the robustness of the effects, we used mul­ tiple methods to manipulate power (i.e., recalling experiences of power 2.4. Reducing attachment anxiety via attachment security and simulating power via a subordinate role in a task), multiple mea­ sures of proactive behaviors (i.e., providing feedback on task design, Based on our theory that low power individuals are less likely to voicing opinions to high power counterparts, and sharing innovative engage in proactive behaviors because of attachment anxiety, we further ideas), and explore two forms of intervention (i.e., making a secure propose that reducing attachment anxiety by providing attachment se­ relationship with a manager salient, and interacting with an attachment- curity should mitigate the effect of low power on proactive behaviors. As security-inducing figure). The study materials and data without identi­ discussed earlier, a central theme in attachment theory is that the fiable information for Study 1, 2, 3, and 5 are posted as a supplement availability of supportive attachment figures provides felt-security, online at the Open Science Framework (https://osf.io/abzq7/?vie which promotes an individual’s exploration (Ainsworth et al., 1978; w_only=8be0d1bbecd64e95a984facdcff3e836). Elliot & Reis, 2003), enabling them to learn about and become compe­ tent in interacting with novel environments (Bowlby, 2008). Making salient a secure relationship with an attachment figureshould provide a

31 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41

4. Study 1 experience greater attachment anxiety than both high power and base­ line individuals. Furthermore, low power was only associated with Study 1 set out to experimentally test Hypothesis 1, that individuals increased state attachment anxiety, not attachment avoidance, sug­ with low power will exhibit greater state attachment anxiety than those gesting the latter is less relevant to the experience of power. with high power. To rule out the possibility that high power is driving the results, we conducted a pre-registered study (https://aspredicted. 5. Study 2 org/4zi57.pdf), in which we compared low power to both high power and a baseline condition. To do so, we employed a power recall task Study 1 demonstrated a causal relationship between low power and (Galinsky et al., 2003) and expected that low power participants, as state attachment anxiety. Next, we conducted a pre-registered study opposed to high power and baseline participants, would experience (https://aspredicted.org/wf6ti.pdf) to test whether heightened state greater state attachment anxiety. We also measured state attachment anxiety would reduce proactivity. To do so, we manipulated state avoidance to test whether it was indeed less relevant to the experience of attachment anxiety using a priming task (Chugh et al., 2014) and power. measured participants’ proactivity (Grant & Rothbard, 2013), predict­ ing that participants primed with attachment anxiety would be less 4.1. Participants and design proactive compared to a baseline condition. Study 2 also intentionally primed attachment anxiety in the context of a relationship with someone We recruited 905 participants from ProlificAcademic. As in the pre- who had power over the participants to help confirmthe organizational registration, we excluded participants who incorrectly answered our relevance of the findings. attention check, leaving 787 participants (396 men, 13 non-binary 1 people, 4 prefer not to say; Mage = 31.38, SDage = 11.18). 5.1. Participants and design Participants were randomly assigned to a high power, low power, or baseline condition. We recruited 595 participants from Amazon Mechanical Turk (243 men, 10 non-binary people, 3 prefer not to say; Mage = 38.19, SDage = 4.2. Procedure 12.64) randomly assigning them to the attachment anxiety or baseline condition. Power manipulation. We instructed participants in the high power condition to recall and write about an experience in which they had 5.2. Procedure power over others, while those in the low power condition recalled and wrote about an experience in which someone else had power over them State attachment manipulation. Participants read a paragraph (e.g., Anderson & Galinsky, 2006; Galinsky et al., 2003). In the baseline priming attachment anxiety (Chugh et al., 2014), which asked them to condition, we asked participants to recall and write about the most recall a situation in which their relationship with their manager or su­ recent bird they saw. pervisor reflected attachment anxiety at the workplace. They read: State attachment measure. After completing the recall task, par­ In the workplace, people frequently interact with their boss (i.e., ticipants reported their state attachment anxiety and state attachment their manager or supervisor). Try to recall a particular situation in avoidance using the modified Experiences in Close Relationship Scale- which you were anxious because you found your boss was reluctant Short Form (ECR-S; Wei, Russell, Mallinckrodt, & Vogel, 2007) with to get as close as you would like and to depend on you. You wanted to regard to how they felt “at the moment.” Example items for state know them more, but you somehow felt that they didn’t want it. You attachment anxiety included “At this moment, I worry that other people were unsure whether they liked you. You often worried that they do not really care for me.” (3 items; α = 0.89). Example items for state didn’t really accept you. attachment avoidance included “At this moment, I would talk things over with people.” (reversed; 6 items; α = 0.79). All items used a 7-point The baseline condition, as in Study 1, asked participants to recall and response scale ranging from 1 (strongly disagree) to 7 (strongly agree). write about the last time they saw a bird. Proactive behavior measure. Participants were told we were 4.3. Results and discussion planning to conduct another study in the future to understand employee proactivity. They were asked to write up to five statements defining As pre-registered, a one-way between-subjects ANOVA was con­ proactivity. They were then told they would be shown seven statements, ducted to compare the effect of power on state attachment anxiety in the ostensibly written by previous participants who had completed the same low power, high power, and baseline conditions. There was a significant task, that the experimenters were planning to use in future studies. effect of power on attachment anxiety, F(2, 784) = 5.10, p = 0.006. Post Following Grant and Rothbard (2013), each of the seven statements had hoc comparisons using the Tukey HSD test indicated that attachment definitional and grammatical errors (e.g., “waiting for things to unfold anxiety for participants in the low power condition (M = 3.96, SD = before responding” is not a definition of proactivity; similarly, “Proac­ 1.66) was significantly higher than the baseline (M = 3.59, SD = 1.72; tivity is very important!” is an irrelevant response to a request for a 95% CI = [0.098,0.808], p = .008) or the high power condition (M = definition). Participants had the option—though they were not requir­ 3.64, SD = 1.75; 95% CI = [0.021, 0.751], p = .035). State attachment ed—to comment on each sentence. Following the pre-registration, we anxiety for participants in the high power condition was not signifi­ operationalized proactivity as the number of statements participants cantly different from those in the baseline condition (95% CI = [ 0.288, commented on, as well as the length of those comments (given that 0.422], p = .898). Power had no significant effect on attachment lengthier comments indicate more effort made in voicing corrections; & avoidance, F(2,784) = 1.40, p = .247 (Mlow power = 3.74, SD = 1.11; Detert Burris, 2007). Both of these factors constitute change-oriented Mhigh power = 3.64, SD = 1.13; Mbaseline = 3.59, SD = 1.18). action directed toward improving the task, consistent with research on These results support Hypothesis 1, in that low power individuals proactive behaviors such as task revision (Grant & Rothbard, 2013; Staw & Boettger, 1990) and suggestion voicing (Grant, Gino, & Hofmann, 2011). 1 We also ran an analysis including participants who failed the attention check and the effect of power on attachment anxiety remained significant, F (2,902) = 4.20, p = .015, and its effects on attachment avoidance were again not significant, F(2,902) = 1.42, p =.242.

32 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41

5.3. Results and discussion corrections. For the number of errors corrected, a Poisson regression found a significant effect for power (Mlow power = 3.60, SD = 3.03; We examined the number of errors corrected using Poisson regres­ Mbaseline = 4.41, SD = 2.86), b = 0.202, SE = 0.043, z = 4.88, 95% CI sion (as the number of errors is a count) with conditions (attachment = [ 0.283, 0.121], p < .001. For the length of written corrections, a anxiety coded 1, baseline coded 0) as a predictor. As expected, we found regression also found a significanteffect for power (Mlow power = 128.59, that participants in the attachment anxiety condition (M = 3.74, SD = SD = 144.90; Mbaseline = 161.72, SD = 150.29), b = 33.131, SE = 2.94) corrected fewer errors compared to those in the baseline condition 12.126, t = 2.73, 95% CI = [ 56.95, 9.32], p = .006. Low power (M = 4.32, SD = 2.91), b = 0.144, SE = 0.041, z = 3.48, 95% CI = reduced proactivity as compared to baseline. [ 0.225, 0.063], p = .001. Likewise, regressing the length of written We then examined the effect of power and attachment anxiety on corrections onto condition, as predicted, showed that participants in the proactivity. For the number of errors corrected, a Poisson regression attachment anxiety condition (M = 137.00, SD = 158.41) made signif­ found a significant negative association by condition (low power icantly shorter corrections than those in the baseline condition (M = compared to baseline), b = 0.194, SE = 0.042, z = 4.67, 95% CI = 165.62, SD = 186.25), b = 28.61, SE = 14.43, t = 1.98, 95% CI = [ 0.275, 0.113], p = <0.001, and a significant negative association [ 56.962, 0.267], p = .048. with attachment anxiety, b = 0.029, SE = 0.012, z = 2.44, 95% CI = Study 2 found that attachment anxiety reduced proactivity in the [ 0.053, 0.006], p = .015. Bootstrapped 95% bias-corrected confi­ form of task revision, for both number of errors participants corrected dence intervals with 1,000 iterations included zero [ 0.029, 0.00004], and the length of their written corrections. but at 90% confidence intervals excluded zero [ 0.025, 0.0004], indicating a marginally significantindirect effect. We also examined the 6. Study 3 length of written corrections as a function of condition and attachment anxiety. As expected, a regression on length of written corrections found Study 1 demonstrated a causal relationship between low power and a significant negative association by condition (low power condition state attachment anxiety, with low power increasing state attachment compared to baseline), b = 31.378, SE = 12.11, t(597) = 2.58, 95% anxiety, and Study 2 demonstrated a causal relationship between state CI = [ 51.385, 11.372], p = .010, and a marginally significant attachment anxiety and proactivity, with heightened attachment anxi­ negative association with attachment anxiety, b = 6.275, SE = 3.52, t ety reducing proactivity. In Study 3, which we pre-registered (https:// = 1.78, 90% CI = [ 12.066, 0.483], p = .075. Bootstrapped 95% aspredicted.org/4mt24.pdf), we integrate the full model and test both bias-corrected intervals with 1000 iterations excluded zero [ 6.396, Hypothesis 1 (that low power increases state attachment anxiety), and 0.027], indicating that state attachment anxiety mediated the rela­ Hypothesis 2 (that state attachment anxiety mediates the relationship tionship between low power and low proactivity, in support of Hy­ between low power and reduced proactivity). To explore the proposed pothesis 2. relationship, we manipulated power and measured attachment anxiety We ran the same analyses with attachment avoidance as a mediator as in Study 1 and measured proactive behavior as in Study 2. for the effects of power on both the number of errors corrected [ 0.021, 0.003] as well as the length of written corrections [ 4.637, 0.339] and 6.1. Participants and design both included zero, showing that attachment anxiety, but not attach­ ment avoidance, is related low power’s effects on proactive behaviors. We recruited 600 participants (278 men, 7 non-binary people, 4 In Study 3, we integrated results from Study 1 and 2 and demon­ prefer not to say; Mage = 37.87, SDage = 11.82) from Amazon Mechanical strated that heightened attachment anxiety indeed mediated the rela­ Turk, randomly assigning them to a low power or baseline condition. tionship between low power and reduced proactivity. While the results of Study 1, 2, and 3 support the proposed theory, with a significant in­ 6.2. Procedure direct effect when proactivity was measured as a continuous variable (i. e., the length of written corrections), there was only a marginal indirect Power manipulation. We manipulated low power by asking par­ effect when measured as a count variable (i.e., the number of errors ticipants to recall a time someone had power over them at the work­ corrected), despite the study being well-powered.2 To explore how place. In the baseline condition, we asked participants to recall and write generalizable the effects were, different operationalizations of proac­ about the last bird they saw. tivity were used in Study 4 and 5. State attachment measure. We used the same attachment measure as in Study 1 (α = 0.91 for attachment anxiety, α = 0.86 for attachment 7. Study 4 avoidance). Proactive behavior measure. We used the same proactivity task as Study 1 through 3 provided support for both Hypotheses 1 and 2 in Study 2. using experimental manipulations of power and state attachment anxi­ ety, finding a causal relationship between power, state attachment 6.3. Results and discussion anxiety, and reduced proactivity. Study 4 and 5 focus on the moderating effects of attachment security, with the expectation that low power First, we examined the effect of power (low power coded 1, baseline participants will be more proactive when attachment security is salient coded 0) on state attachment. A one-way between-subjects ANOVA as opposed to not. found a significant effect of power on attachment anxiety, F(1, 598) = Study 4 had two goals: (1) To manipulate state attachment security 3.93, p = .048. Participants in the low power condition reported to further establish causality; (2) To show that state attachment security significantlymore attachment anxiety (M = 3.13, SD = 1.75) than those in the baseline condition (M = 2.85, SD = 1.69; 95% CI = [0.003,0.556], 2 p = .048, ηp = 0.007, supporting Hypothesis 1. As expected, we did not a 2 We asked coders to examine the participant responses and conservatively find significant effect of power on attachment avoidance, replicating count whether the responses were truly proactive in nature. With the coded Study 1, F(1,198) = 2.49, p = .115 (Mlow power = 3.51, SD = 1.32; = = data, we finda significantindirect effect at the 95% level for the count measure Mbaseline 3.34, SD 1.35). [ 0.073, 0.001] as well as the length measure [ 3.738, 0.072]. Given that ′ Next, to replicate previous findings in power research (e.g., Karre­ Study 2 used the same paradigm, we also submitted Study 2 s participant re­ mans & Smith, 2010; Keltner et al., 2003; Lammers et al., 2013), we sponses to the same process, and the results remain significant. The coding examined the effect of power on proactivity. As in Study 2, we examined schema and additional analyses have been uploaded to the OSF repository both the number of errors corrected as well as the length of written associated with this project.

33 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 gained from sources within a workplace context would be effective. As [company name] better by visiting [website] or by scanning this QR discussed above, leaders and managers can provide attachment security code!” (Davidovitz et al., 2007; Mayseless & Popper, 2007; Mayseless, 2010; In the baseline condition, the stories did not involve any elements of Popper & Mayseless, 2007). Following this logic, we ran a field experi­ attachment security, instead emphasizing each employee’s uniqueness ment and manipulated state attachment security in the workplace (e.g., “My knowledge of military methods and procedures helped,” “My among low power employees. We predicted that state attachment se­ familiarity with and passion for graphic arts are an important part of my curity would moderate the link between low power and reduced pro­ idea”). The poster concluded with, “Research shows that effective activity, such that the link will be weaker when attachment security is thinkers often have a unique perspective – think about what makes you salient compared to when it is not. In this study we focused on inno­ different. Then share your ideas to make [company name] better by vation (Crant, 2000) as a form of proactivity in the workplace which visiting [website] or by scanning this QR code!” For both posters the QR contributes to the identification of a problem, the generation of novel code and website address both led to the idea gathering tool. ideas, and their implementation (Seibert et al., 2001). Innovative Assignment to condition. The company grouped its branches into behavior (e.g., coming up with new ideas), is considered an extra-role five different regions, consisting of four different geographical regions behavior, as it involves the development of new ideas, processes, and (Canada, Western US, Northern US, and Southern US) and warehouses routines at organizations (Welbourne, Johnson, & Erez, 1998). Inno­ and corporate offices (Others). For each region there were different vative behaviors can also drive bottom-up change via the voicing of numbers of branches, which themselves differed in size and amount of constructive ideas (Grant et al., 2011; Van Dyne & LePine, 1998). We technicians participating in the first six months of the idea generation examined innovative behaviors among low power employees (definedas tool’s existence. To control for these three differences (region, branch those who do not have supervisory responsibilities and therefore do not size, past participation in the tool), we employed non-bipartite matching have control over any subordinates’ outcomes). (Gerber & Green, 2012) with region as block, and branch as cluster with a matched pair design. Within each of the four regions, we sorted 7.1. Context and design branches by a) total number of technicians, and b) number of techni­ cians participating in the idea generation tool thus far. Then (using R A Fortune 500 manufacturing company implemented a tool aimed at package ‘nbpMatching’), the branches were compared, matched based gathering ideas from technicians. The technician role in this company on total number of technicians and number of technicians participating involved frontline employees without managerial or supervisional re­ in the idea generation tool thus far, then randomized such that one sponsibilities but with supervisors they reported to (i.e., the technicians member of each pair was assigned to each of the two conditions (see are low power in this context). The total number of technicians in the Appendix B). Because this method of assignment to condition balances organization numbered 3823 (3489 men; 20–39 years old, 49.5%; 40 out the relative number of technicians and number of past participants, years or above, 50.5%), and this population constituted the number of our dependent measure of innovation behavior would consist of the potential idea contributors. In the firstsix months following the launch number of technicians submitting ideas by branch to the idea generation of the tool, only 55 technicians submitted ideas. tool after the posters were put up in the break rooms of each branch. The The fieldstudy occurred in the context of a company relaunch of the poster was distributed on a Monday to be hung in the breakroom the idea gathering tool. The company planned to select two ideas submitted following day. The number of technicians submitting ideas to the idea in the past six months, recognize the employees who submitted the generation tool were tracked from Tuesday through Friday on the week ideas, and make salient the existence of the idea gathering tool via of the poster’s distribution.3 posters hung in break rooms across all branches of the organization The expectation was that the number of technicians submitting ideas encouraging technicians to submit their ideas via the idea gathering from branches exposed to the attachment security poster would be tool. greater than the number of technicians submitting ideas from branches In the planned intervention, one of two different posters announcing exposed to baseline poster. the relaunch of the idea gathering tool was to be distributed to each of Attrition. Staff at each branch were asked to confirm that they had the different branches in the company, one attachment security poster received the poster and hung it on the breakroom wall. Of 198 posters making salient attachment security with one’s managers, and one delivered, 97 confirmations were received. No differences in confirma­ baseline poster which did not make salient attachment security with tions emerged between the attachment security and baseline conditions one’s managers. The number of technicians submitting ideas to the idea in the regions of Canada (12 each), Others (13 attachment security gathering tool from each branch would be counted, with the expectation posters and 12 baseline posters), and the Northern US region (30 each). that more technicians would submit ideas from the branches in which However, confirmationswere distributed unevenly across conditions in attachment security had been made salient as opposed to those in which the Southern US region (7 attachment security posters and 13 baseline it had not. posters) and the Western US region (5 attachment security posters and 10 baseline posters). Furthermore, in both the Southern US region and in 7.2. Procedure the Western US region branches were closed due to COVID-19 in the attachment security condition (2 in each region) but not in the baseline Intervention design. The company selected two successful ideas condition. Due to these problems with distribution in the Southern US 4 submitted during the first six months of the idea generation tool’s ex­ and Western US regions, both were excluded from the analysis. istence with the intent of recognizing the ideas and their submitters throughout the company. Interviews were conducted with both technicians for the poster 3 content. Two posters were designed, each featuring both technicians Data from the following week was not collected, as that week constituted with a quote from the interviews (see Appendix A for anonymized ver­ the US Thanksgiving holiday, with differing patterns of which branches were open across the company regions. sions of these posters). In the attachment security condition, the quotes 4 At the time of distribution, the administrative staff was advised to work emphasized having a secure attachment with one’s supervisor (e.g., “It from home in these regions (e.g., Texas, California), both of which reported makes you feel like you can depend on them and have them depend on high numbers of COVID-19 cases, and this may have been a factor in the uneven ” “ ” you, you feel that you are being valued as much as you value them ). distribution of the posters. Including these regions in the analysis decreased the The poster concluded with, “Research shows that effective thinkers often difference between conditions (number of technicians submitting ideas in the have strong bonds with their managers – think about times you and your attachment security condition, 90, number submitting ideas in the baseline manager had each other’s backs. Then share your ideas to make condition, 71, χ2 (1) = 2.248, p = .134).

34 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41

7.3. Results and discussion We predicted that interacting with the attachment-security-inducing figure would increase proactivity as compared to interacting with an Because the posters were displayed in the context of an overall attachment-anxiety-inducing figure. relaunch of the idea gathering tool, overall participation increased significantly across all regions as expected: for branches receiving the attachment security poster, compared to 18 technicians participating in 8.1. Participants and design the first 6 months, 71 technicians submitted ideas during the data collection period, χ2(1) = 33.76, p < .001; for branches receiving the We recruited 602 participants from Amazon Mechanical Turk (293 baseline poster, compared to 17 technicians participating in the first 6 men, 3 non-binary people; Mage = 40.63, SDage = 13.43) randomly months, 45 technicians submitted ideas during the data collection assigning them to the attachment anxiety condition or the attachment period, χ2(1) = 13.11, p < .001. Central to our hypothesis, when security condition. comparing the number of technicians submitting ideas in each condition during the data collection period, a greater number of technicians sub­ 8.1.1. Procedure mitted ideas in the attachment security condition as compared to in the Interactive paradigm. The study was advertised as an interactive baseline condition, χ2(1) = 5.83, p = .016, supporting Hypothesis 3. study which specifiedthat participants would be assigned to a team with The results support Hypothesis 3 that low power individuals behave another individual participating in the study at the same time. To in­ more proactively when attachment security is made salient as opposed crease believability, the advertisement also noted that participants who to not, and further, that attachment security moderates the link between were ill-mannered, rude, and/or impolite would be subject to rejection. low power and low proactivity, such that the link will be weaker when Upon arrival, participants were told that the study focused on how attachment security is high compared to low. As participants engaged in people interact and make decisions in online environments because non- real workplace behaviors without knowing that they were participating in-person communications are increasingly common. Participants were in a research study and that an experimental manipulation had occurred, told that direct communication would be limited but that they and their the results provide external validity. Yet it is worth noting that the study counterpart would be given opportunities to share messages with one design did not allow for the measurement of subjective feelings of another. We asked participants’ initials and how long they had been power, and there is evidence that regardless of the positional power employed throughout their life. Participants waited for about 30 s while individuals hold (e.g., if they are subordinates), their feelings of power the survey told them they were being assigned to a team with another fluctuatethroughout the day (Smith & Hofmann, 2016), suggesting the participant. They were then told they had been assigned to a team with study results should be interpreted with some caution. Nonetheless, another participant whose initials were JMH and that they would be individuals’ positional power is strongly related to their subjective working on a series of tasks. feelings of power (Smith & Hofmann, 2016), and it is likely that situa­ Low power induction. Most importantly, participants were told tional factors (such as the poster which itself constituted the study their team would be working on an idea generation task. Participants intervention, and other physical and social aspects of employees’ were shown a table with their and JMH’s initials, work experience, and workplace context) acted as cues which brought to mind their low power assigned role for the idea generation task. In all cases, JMH always had position (Tost, 2015). five more years of work experience than participants did and was This study illustrates in a multi-regional company context that when assigned the role of idea evaluator. All participants were assigned to the attachment security is made salient, the link between low power and role of idea worker. reduced proactive behavior in the form of innovative idea sharing is Participants were told that the idea worker would generate ideas weakened. Study 5 goes on to test an attachment security intervention in while the idea evaluator answered a series of questions independently. a more controlled experimental context. They were told they would then switch tasks, with the idea evaluator evaluating the ideas the participant had just generated, while the idea 8. Study 5 worker went on to answer a series of questions independently. Participants were told that a bonus of 50 cents would be divided Study 4 demonstrated that low power individuals behave more between them and their idea evaluator, and that the idea evaluator’s proactively when attachment security is made salient as opposed to not, scoring of their performance would determine how the bonus money supporting Hypothesis 3. Study 5 also tests Hypothesis 3, this time in a would be divided. Thus the idea evaluators had power over the evalu­ more controlled experiment. ations the participants received, and their outcomes. Study 1 through 3 utilized a recall task, asking participants to recall a Secure attachment intervention. Participants were given a chance time when someone else had power over them or when they experienced to write a message to the idea evaluator to introduce themselves and attachment anxiety. Although vividly recalling past experiences is a received a message from JMH in turn. As noted in our theorizing, low powerful way to induce a sense of power (Galinsky et al., 2003) or sense power increases attachment anxiety because low power individuals’ of attachment anxiety (Chugh et al., 2014), it is important for general­ dependence on high power individuals leads them to desire a feeling of izability to induce power in multiple ways. To do so, we ran a pre- security and acceptance, while also increasing their perception that the registered study (https://aspredicted.org/m798e.pdf) in which all par­ relationship is more variable and less reliable; therefore, in the attach­ ticipants were put in a low power position with an ostensible interaction ment anxiety condition the high power individual behaved in a way partner whose task it was to evaluate the participants’ performance. To which did not provide reassurance that the relationship would be secure ensure a standardized series of communications, the interaction partner – showing variability in responses, and not explicitly affirmingsupport. was not a real participant, but a pre-programmed set of responses within In the attachment anxiety condition, participants received the following the survey that participants believed to be another individual; these message: ostensible interaction partners behaved in a fashion congruent with Hello I’m Jake and I guess I’ll be evaluating you. Let’s see how you do at attachment security- or attachment-anxiety-inducing figures. Partici­ this. Let’s stay professional so I can evaluate you objectively on this task. pants were then given an opportunity for proactivity. To capture pro­ activity in this study we used voice (Burris, 2012), a proactive workplace In contrast, attachment security is invoked when an individual feels behavior (LePine & Van Dyne, 2001; Van Dyne & LePine, 1998) which accepted and supported by an attachment figure,or in other words feels involves speaking up on important issues and problems in organizations. comfortable depending on and being depended on by them (Baldwin Voice is thus a measure of employee action (Grant & Ashford, 2008) that et al., 1996; Chugh et al., 2014). Thus, in the attachment security condi­ is highly relevant to organizational contexts. tion, the ostensible counterpart explicitly affirmed the dependability of

35 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 the relationship, in that participants received the following message: to not. This study illustrates in an interactive setting that when attach­ ment security is made salient, the link between low power and reduced Hello, I’m Jake and i guess I’ll be evaluating you. You’ll be great at this. proactive behavior in the form of voicing opinions is weakened. Let’s support each other on this task so we can have each other’s backs.

Participants then engaged in a task where they were asked to 8.3. General discussion generate as many original and creative ideas they could for how to use a brick (Christensen, Guilford, & Wilson, 1957). After completing the task, Across fivestudies we explored the relationship between low power participants in the attachment security condition received another and reduced proactivity through the lens of attachment theory. Specif­ message from JMH: ically, we found that low power individuals show greater attachment anxiety (Study 1 and 3) and that attachment anxiety decreases proac­ ’ ’ I m about to see your answers - I m sure you did great. tivity (Study 2 and 3). We illustrated that attachment anxiety was In the attachment anxiety condition, they did not receive another increased in low power (as opposed to decreased in high power) via the message from JMH. use of baseline and high power conditions (Study 1 and 3), and that Voice measure. Participants were told that while they had been power did not consistently predict attachment avoidance (Study 1 and completing the idea generation task, the idea evaluator was given a 3). Interventions which induced attachment security weakened the word puzzle, and asked to rewrite the puzzle in their own words. Par­ relationship between low power and reduced proactivity in both a quasi- ticipants were then presented with the word puzzle ostensibly written by experiment (Study 4) and interactive experimental design (Study 5). In JMH, which had a number of grammatical errors and typos, and asked to all, we replicated the effect using multiple methods to manipulate solve it. After solving the word puzzle, participants were given an option power, operationalizing proactivity in a variety of ways, and exploring to send a message to JMH, in which they could tell them about any errors interventions in multiple contexts; all speaking to the robustness of the in what they wrote or how to improve their writing in the future. By effects. Overall, we consistently find support for attachment anxiety as sending a message, participants could provide information to improve an important factor in the experience of low power individuals which JMH’s work but which might also displease JMH, who had the authority inhibits them from being proactive. We also offer a powerful interven­ to control their outcomes; we considered this to constitute voice (Detert tion that facilitates the proactivity of those with low power. & Burris, 2007). Choosing to send a message to JMH was coded as voice, a 1. Not sending a message was coded as 0. 8.4. Theoretical contributions

The present research expands the power-related literature by illu­ 8.2. Results and discussion minating a new consequence of low power: attachment anxiety. Power is a relational construct existing only within social relationships (Emerson, First, to verify that our manipulation indeed induced low power and 1962; Fiske, 1993; Salancik & Pfeffer, 1974; Thibaut & Kelley, 1959). attachment security, we conducted a pre-test of the paradigm with 118 However, in spite of its grounding in interpersonal connections, there MTurkers (60 men; Mage = 37.96, SD = 11.96). We administered the has been a lack of systematic examination of how power, especially low same process as above; however after receiving their role description power, is interlaced with mental representations of relationships. We and before they introduce themselves to the interaction partner, we argue that, via attachment theory, we are able to introduce the construct asked participants to indicate to what extent they felt powerful using of attachment anxiety as a key link between the experience of low power three items (powerful, dominant, and in control; α = 0.96; 1 = not at all, and reduced proactivity. Previously, Smith, Jostmann, Galinsky, and 7 = a great deal). As expected, participants experienced low power (M = van Dijk (2008) demonstrated that low power diminishes people’s ex­ 3.24, SD = 1.59), which is significantly below the midpoint, t(117) = ecutive functions, as they have difficulty maintaining focus on their 5.20, 95% CI = [ 1.053, 0.472], p < .001. In addition, after their goal. Our findings have implications for this phenomenon, as attach­ interaction, instead of measuring voice, we measured participants’ state ment anxiety has been reported to distract people from focusing on tasks attachment security using an attachment security measure drawn from because they place increased attention on seeking reassurance of sup­ the work of Gillath, Hart, Noftle, and Stockdale (2009; on a scale from 1 port and acceptance (Lee & Ling, 2007; Shaver, Schachner, & Miku­ = not at all to 7 = very much; 7 items, α = 0.96). As expected, partici­ lincer, 2005). Our studies suggest that the experience of low power pants interacting with the attachment-security-inducing figure (M = creates relational concerns that reduce action-taking, and this increased 5.24, SD = 1.14) reported significantly greater levels of attachment se­ need for reassurance, support, and acceptance is also consistent with the curity than participants interacting with attachment-anxiety-inducing increased motivation of low power individuals to seek social affiliation figure (M = 4.70, SD = 1.72), t(116) = 2.01, 95% CI = [ 1.075, (Case et al., 2015). 0.007], p = .047. Our research also complements work on the impact of power on Following the pre-registration, we examined voice (in terms of hierarchy (Anderson & Galinsky, 2006; Galinsky et al., 2003, 2006; sending a message to JMH) using logistic regression by condition Magee et al., 2007), particularly findings that people lower in a power (attachment security coded 1, attachment anxiety coded 0) as a pre­ hierarchy exhibit less competitive behavior (Hays & Bendersky, 2015). dictor. As expected, we found that participants interacting with the Our findings not only offer a mechanism for the reduced action associ­ attachment-security-inducing figure had 2.06 times the odds of using ated with low power but also suggest a reason why those with low power voice compared to participants interacting with the attachment-anxiety- may perceive more challenges related to moving up a hierarchy. The inducing figure, b = 0.723, SE = 0.170, z = 4.25, 95% CI = [0.389, attachment literature suggests that individuals with high attachment 1.057], p < .001. In addition, because people differ in how they interact anxiety have a more negative view of the self and are vigilant for in online studies, we also ran the same logistic regression controlling for rejection cues (Mikulincer, 1998); they will thus be less likely to engage whether participants’ sent an initial message to JMH (coded 1 for yes, in security-risking activities such as attempts to ascend within a hier­ 0 for no). In this analysis participants interacting with the attachment- archy. Thus, both low power individuals’ reduced likelihood of action as security-inducing figure had 2.22 times the odds of using voice well as increased fear of the cost of bids to move up a hierarchy likely compared to participants interacting with the attachment-anxiety- contribute to power hierarchies’ stability. inducing figure, over and above their initial messaging pattern, b = Lastly, we provide a unique intervention grounded in an under­ 0.799, SE = 0.177, z = 4.51, 95% CI = [0.451, 1.146], p < .001. standing of how attachment theory helps to explain the relationship The results support Hypothesis 3 that low power individuals behave between low power and reduced action. Previous research has theorized more proactively when attachment security is made salient as opposed and tested that low power (compared to high power) individuals take

36 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 less action (Anderson & Berdahl, 2002; Galinsky et al., 2003). Compared For instance, it has been theorized that powerlessness activates the to the high power to action link, which can be explained by multiple behavioral inhibition system (BIS; Keltner et al., 2003), but it is unclear factors (e.g., via freedom from social interference, the behavioral how much or under what context the BIS and the attachment system approach system, and reduced deliberation; Croizet & Claire, 1998; contribute to reduced action. In fact, research has shown that attach­ Guinote, 2007; Keltner et al., 2003; Whitson et al., 2013), research on ment dimensions, although weakly to moderately related to the BIS power has paid little attention to the question of how to weaken or break (Jiang & Tiliopoulos, 2014), predicted individuals’ response to rela­ the low power-inaction link. Our work provides one mechanism and a tionship threats independently from the BIS (Meyer, Olivier, & Roth, related intervention to illuminate—and ameliorate—the link between 2005). It may be that differing contexts call up the BIS and the attach­ low power and reduced proactivity. ment system to different extents, and future research should look into these factors in the context of power dynamics. 8.5. Limitations and future research Future research should also investigate (and potentially attempt to systematize) findingsregarding the boundary conditions around power While these findings illustrate that attachment anxiety is one and action. For example, there is already some evidence that low power mechanism driving the effects of low power on reduced action (and individuals may more actively pursue affiliative goals (which would therefore proactivity), it is important to acknowledge some limitations satisfy their needs for felt-security) than high power individuals. As of our findings. Specifically, while Study 5 focused on how low power mentioned earlier, Case et al. (2015) showed increased action in low individuals would be influenced by interactions with a secure versus a power individuals when action was operationalized as seeking social baseline leader, there are some reasons to interpret the results with affiliation.Conversely, Durso, Brinol,˜ and Petty (2016) have found that caution. Attachment security depends on instilling positive associations under certain conditions, high power individuals can be less active with interdependence and trust (Arriaga & Kumashiro, 2019). Because compared to low power individuals. An organized exploration on the attachment security is defined as low on both attachment anxiety and circumstances under which low power leads to greater action than high avoidance, we incorporated language on mutual dependence in our power or vice versa would be a promising direction for future research. attachment-security-inducing figure manipulation. However, it is Relatedly, there may be benefitsfrom future research which explores possible that the wording used to induce attachment security may have the boundary between proactive and exploration behaviors. In this also prompted participant perceptions of the leader as more friendly or paper, we examined proactive behaviors, which we consider to be of their power relationship as more equal (given the exploratory behavior occurring in professional and organizational con­ inter-dependencies), which may have affected their behaviors. The texts. Yet, it is worth examining exploratory behaviors which are not perception of the leader as more friendly may have made participants generally considered proactive in nature. For instance, reduced explor­ more willing to engage in voicing behavior, and it would be beneficial atory behavior has been shown empirically in non-organizational con­ for future work to test how leader friendliness influences or interacts texts, such as when participants show reduced interaction with a puzzle with their ability to induce a sense of security. Similarly, while a pre-test box (Main, 1983) or reduced exploration of relationships with peers (e. confirmed the effects on participants’ sense of power and attachment g., less time spent interacting with playmates and more time playing security, it is possible that the statements made by the secure leader may with toys). Studies in attachment theory have also investigated explo­ have resulted in a sense of a more equal power relationship, also making ration in other domains such as curiosity (i.e., information search) and participants more willing to engage in voicing behavior. While Study 4, cognitive closure (i.e., openness to new information; Hazan & Shaver, in using an intervention around attachment security in a quasi 1990; Johnston, 1999; Mikulincer, 1997). While this paper focused on field-experimental context, lends confidence to effects found, future proactive behaviors due to their relevance for organizations, the po­ research would still benefit from continuing to disentangle these tential exists to use attachment-related accounts to explain a variety of phenomena. power-related phenomena in domains external to professional or orga­ Future work should also continue to explore mechanisms relevant to nizational contexts. low rather than high power states. Theorizing by Tost (2015) is partic­ ularly relevant here, wherein a delineation is made between the effects 8.6. Practical implications of structural power (i.e., actual control over valuable resources) as opposed to psychological sense of power (i.e., feeling powerful). Indeed, Power hierarchies emerge spontaneously in all types of groups we examined the effects of the experience of low power via both (Bales, Strodtbeck, Mills, & Roseborough, 1951). Hierarchies are sup­ structural (Study 4 and 5) and psychological (Study 1 and 3) approaches. posed to facilitate efficient decision-making and motivate group mem­ We argue that the effects were the same across all of these inductions of bers lower in the hierarchy to perform well to move up it via rewards power because they all relied on low power individuals’ awareness of and promotions (McClelland, 1987). However, our research findingscall being in a relationship with someone who had power over them. Tost into question the efficiencyof having too many low powered individuals does argue that differing effects in the literature, particularly with re­ in a given hierarchy; when individuals are overly concerned about gard to sense of responsibility, can be traced back to whether partici­ others’ acceptance, their proactivity maybe hampered. pants experienced structural power or a psychological sense of power. It is worth noting that the messages we used in the attachment- Indeed, further work by Tost and Johnson (2019) find different effects anxiety condition in Study 5 are on their face objective and fair state­ on prosocial orientation as a result of structural versus psychological ments—indeed, it seems reasonable for managers to communicate that power inductions. However, the theorizing here is focused on the state of they are interested in seeing how their subordinates perform on a task. high power, in particular how structural power over others may also While past research has shown that negative feedback can reduce em­ prompt greater awareness of their dependence on oneself, and thus one’s ployees’ proactive behavior (e.g., creativity; Kim & Kim, 2020), our responsibility for their well-being. It may be that this congruence of results suggest that when followers received even seemingly objective effects for low power goes beyond the activation of the attachment and fair communications they still did not feel secure and did not in­ system, and that the structural experience of low power and the psy­ crease their proactive behaviors. This, in combination with the fact that chological experience of low power are more aligned, such that low many managers feel more effective as leaders when criticizing and power states have similar effects regardless of whether they come via providing negative feedback (Zenger & Folkman, 2017), likely means structural or relational power inductions, and future research would that managers who wish to encourage proactive behaviors of their benefit from exploring this possibility more thoroughly. subordinates will need to do more than reduce the negativity of their More broadly, further investigation is needed in order to understand statements to provide feelings of security. the broader framework within which power’s effects on action operate. Another way to mitigate this problem is to empower those in the

37 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41 lower part of the hierarchy. A manager may empower each member in (Bunderson, Van Der Vegt, Cantimur, & Rink, 2016) and task-distracting the lower domain of the hierarchy using multiple sources of power (for conflicts (Bendersky & Hays, 2012). Similarly, given that hierarchies instance, expertise, referent; French & Raven, 1959, social versus per­ may remain stable in part because those lower in the hierarchy exhibit sonal power; Greer, 2014; Lammers, Stoker, & Stapel, 2009), giving less competitive behavior (Hays & Bendersky, 2015), widespread pro­ each individual a sense of reduced dependence on their superior, thus activity may contribute to the destabilization of existing power hierar­ protecting them from increases in anxious attachment. In a group with a chies, making organizational coordination more difficult. more balanced power distribution—where one member excels in expertise power and another in referent power, for example—the 9. Conclusion amelioration of low power may facilitate each member’s action; their hierarchical power remains largely constant. That is, even if each The dynamic nature of contemporary organizations requires em­ member holds the same hierarchical rank, empowering each member ployees to be more proactive in the workplace. At the same time, with relevant, different sources of power may prevent the activation of organizational hierarchy and the resultant power relationships can be a attachment anxiety and further inhibition of action. deterrent for employees to behave proactively. Because of the centrality In addition, organizational members might consider how to imple­ of power relationships in organizational contexts, understanding how ment the logic proposed by the interventions in this paper. An effective the nature of those relationships refracts the experience of power illu­ manager can encourage proactive behavior from their subordinates by minates how and why its downstream effects, such as reduced proac­ building a secure environment. The activation of attachment anxiety tivity, come about. The experience of lacking power changes not only stems primarily from the threat low power poses for security. At work, one’s affect, cognition, and behavior, but also one’s attachment repre­ anxiously attached individuals often spend too much of their attention sentation. It is thus important to understand the relational un­ on worrying about failure and trying to please others (Hazan & Shaver, derpinnings of low power, which can account for the inhibition of action 1990). A manager, should they observe a subordinate spending their for those in less powerful positions. attention in this way, might respond by consistently signaling that the subordinate is accepted and valued, which would contribute to the CRediT authorship contribution statement activation of the exploration system, and thus increase the likelihood the subordinate would engage more confidently with their environment. Jieun Pai: Conceptualization, Methodology, Formal analysis, Congruently, employees for whom proactive behaviors are an important Writing - original draft. Jennifer Whitson: Conceptualization, Meth­ part of the job should seek out individuals with whom they can establish odology, Resources, Writing - review & editing. Junha Kim: Concep­ a secure relationship. If their own managers cannot be a source of this tualization, Writing - original draft. Sujin Lee: Conceptualization, security, they might interact with other potential attachment figures in Writing - review & editing. the organization (e.g., peers, leaders, mentors), thus building a network of support that will make it easier for them to behave proactively. However, an important caveat is relevant here—an organization may Declaration of Competing Interest not, ceteris paribus, desire that individuals lower in the hierarchy engage unexaminedly in widespread proactive behaviors. For instance, some The authors declare that they have no known competing financial proactive behaviors, such as unexpectedly taking charge, may disrupt interests or personal relationships that could have appeared to influence effective group functioning by introducing coordination problems the work reported in this paper.

Appendix A

a. Attachment Security Poster

38 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41

b. Baseline Poster

Appendix B

Attachment Security Poster Baseline Poster

Number of Participants in previous 6 Number of Number of Participants in previous 6 Number of branches months technicians branches months technicians

Canada 12 4 291 12 3 260 Others 13 0 58 12 1 53 Northern 30 14 708 30 13 727 US Southern 24 3 481 24 7 470 US Western US 20 6 364 21 4 411 Total 99 27 1902 99 28 1921

References Baldwin, M. W., Keelan, J. P. R., Fehr, B., Enns, V., & Koh-Rangarajoo, E. (1996). Social- cognitive conceptualization of attachment working models: Availability and accessibility effects. Journal of Personality and Social Psychology, 71, 94–109. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A Bales, R. F., Strodtbeck, F. L., Mills, T. M., & Roseborough, M. E. (1951). Channels of psychological study of the strange situation. Hillsdale, NJ: Erlbaum. communication in small groups. American Sociological Review, 16(4), 461–468. Anderson, C., & Berdahl, J. L. (2002). The experience of power: Examining the effects of Bateman, T. S., & Crant, J. M. (1993). The proactive component of organizational power on approach and inhibition tendencies. Journal of Personality and Social behavior: A measure and correlates. Journal of Organizational Behavior, 14(2), Psychology, 83(6), 1362–1377. 103–118. Anderson, C., & Galinsky, A. D. (2006). Power, optimism and risk-taking. European Becherer, R. C., & Maurer, J. G. (1999). The proactive personality disposition and Journal of Social Psychology, 36, 511–536. entrepreneurial behavior among small company presidents. Journal of small business Anderson, C., & Thompson, L. L. (2004). Affect from the top down: How powerful management, 37(1), 28–36. individuals’ positive affect shapes negotiations. Organizational Behavior and Human Bendersky, C., & Hays, N. A. (2012). Status conflictin groups. Organization Science, 23(2), Decision Processes, 95(2), 125–139. 323–340. Arriaga, X. B., & Kumashiro, M. (2019). Walking a security tightrope: Relationship- Bowlby, J. (1969/1982). Attachment and loss (Vol. 1, attachment). New York: Basic induced changes in attachment security. Current Opinion in Psychology, 25, 121–126. Books. Ashford, S. J., & Black, J. S. (1996). Proactivity during organizational entry: The role of Bowlby, J. (2008). A secure base: Parent-child attachment and healthy human development. desire for control. Journal of Applied Psychology, 81(2), 199. Basic books. Ashford, S. J., Blatt, R., & VandeWalle, D. (2003). Reflections on the looking glass: A Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult review of research on feedback-seeking behavior in organizations. Journal of romantic attachment: An integrative overview. In J. A. Simpson, & W. S. Rholes Management, 29(6), 773–799. (Eds.), Attachment theory and close relationships (pp. 46–76). New York: Guilford Ashford, S. J., & Cummings, L. L. (1985). Proactive feedback seeking: The instrumental Press. use of the information environment. Journal of Occupational Psychology, 58(1), Bretherton, I. (1985). Attachment theory: Retrospect and prospect. Monographs of the 67–79. society for research in child development, 3–35. Bretherton, I., & Munholland, K. A. (2008). Internal working models in attachment relationships: Elaborating a central construct in attachment theory.

39 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41

Brion, S., & Anderson, C. (2013). The loss of power: How illusions of alliance contribute Grant, A. M., Gino, F., & Hofmann, D. A. (2011). Reversing the extraverted leadership to powerholders’ downfall. Organizational Behavior and Human Decision Processes, advantage: The role of employee proactivity. Academy of Management Journal, 54(3), 121(1), 129–139. https://doi.org/10.1016/j.obhdp.2013.01.005. 528–550. Bunderson, J. S., Van Der Vegt, G. S., Cantimur, Y., & Rink, F. (2016). Different views of Grant, A. M., & Rothbard, N. P. (2013). When in doubt, seize the day? Security values, hierarchy and why they matter: Hierarchy as inequality or as cascading influence. prosocial values, and proactivity under ambiguity. Journal of Applied Psychology, 98 Academy of Management Journal, 59(4), 1265–1289. (5), 810. Burris, E. R. (2012). The risks and rewards of speaking up: Managerial responses to Green, J. D., & Campbell, W. K. (2000). Attachment and exploration in adults: Chronic employee voice. Academy of Management Journal, 55(4), 851–875. and contextual accessibility. Personality and Social Psychology Bulletin, 26(4), Butterfield, K. D., Trevino, L. K., & Ball, G. A. (1996). Punishment from the manager’s 452–461. perspective: A grounded investigation and inductive model. Academy of Management Greer, L. L. (2014). Power in teams: Effects of team power structures on team conflictand Journal, 39(6), 1479–1512. team outcomes. In O. B. Ayoko, N. M. Ashkanasy, & K. A. Jehn (Eds.), Handbook of Campbell, D. J. (2000). The proactive employee: Managing workplace initiative. conflict management research (pp. 93–108). Edward Elgar. Academy of Management Perspectives, 14(3), 52–66. Grossmann, K., Grossmann, K. E., Kindler, H., & Zimmermann, P. (2008). A wider view Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and of attachment and exploration: The influence of mothers and fathers on the affective responses to impending reward and punishment: The BIS/BAS scales. development of psychological security from infancy to young adulthood. Journal of Personality and Social Psychology, 67(2), 319. Gruenfeld, D. H., Inesi, M. E., Magee, J. C., & Galinsky, A. D. (2008). Power and the Case, C. R., Conlon, K. E., & Maner, J. K. (2015). Affiliation-seeking among the objectification of social targets. Journal of Personality and Social Psychology, 95(1), powerless: Lacking power increases social affiliativemotivation. European Journal of 111–127. Social Psychology, 45(3), 378–385. Guinote, A. (2007). Behaviour variability and the situated focus theory of power. Chance, M. R. A. (1967). Attention structure as the basis of primate rank orders. Man, 2 European Review of Social Psychology, 18(1), 256–295. (4), 503–518. Hays, N. A., & Bendersky, C. (2015). Not all inequality is created equal: Effects of status Christensen, P. R., Guilford, J. P., & Wilson, R. (1957). Relations of creative responses to versus power hierarchies on competition for upward mobility. Journal of Personality working time and instructions. Journal of Experimental Psychology, 53(2), 82. and Social Psychology, 108(6), 867–882. Chugh, D., Kern, M. C., Zhu, Z., & Lee, S. (2014). Withstanding moral disengagement: Hazan, C., & Shaver, P. R. (1987). Romantic love conceptualized as an attachment Attachment security as an ethical intervention. Journal of Experimental Social process. Journal of Personality and Social Psychology, 52, 511–523. Psychology, 51, 88–93. Hazan, C., & Shaver, P. R. (1990). Love and work: An attachment-theoretical perspective. Crant, J. M. (1995). The Proactive Personality Scale and objective job performance Journal of Personality and Social Psychology, 59(2), 270–280. among real estate agents. Journal of Applied Psychology, 80(4), 532–537. Hazan, C., & Shaver, P. R. (1994). Deeper into attachment theory. Psychological Inquiry, 5 Crant, J. M. (2000). Proactive behavior in organizations. Journal of Management, 26(3), (1), 68–79. 435–462. Hazan, C., & Zeifman, D. (1999). Pair bonds as attachments. In J. Cassidy, & P. R. Shaver Croizet, J.-C., & Claire, T. (1998). Extending the concept of stereotype threat to social (Eds.), Handbook of attachment: Theory, research, and clinical applications (pp. class: The intellectual underperformance of students from low socioeconomic 336–354). New York: Guilford. backgrounds. Personality and Social Psychology Bulletin, 24(6), 588–594. Jiang, Y., & Tiliopoulos, N. (2014). Individual differences in adult attachment and Davidovitz, R., Mikulincer, M., Shaver, P. R., Izsak, R., & Popper, M. (2007). Leaders as reinforcement sensitivity. Personality and Individual Differences, 68, 205–210. attachment figures: Leaders’ attachment orientations predict leadership-related Johnston, M. A. (1999). Influences of adult attachment in exploration. Psychological mental representations and followers’ performance and mental health. Journal of Reports, 84(1), 34-34. Personality and Social Psychology, 93(4), 632–650. Karremans, J. C., & Smith, P. K. (2010). Having the power to forgive: When the De Waal, F., & Waal, F. B. (2007). Chimpanzee politics: Power and sex among apes. JHU experience of power increases interpersonal forgiveness. Personality and Social Press. Psychology Bulletin, 36(8), 1010–1023. Detert, J. R., & Burris, E. R. (2007). Leadership behavior and employee voice: Is the door Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach, and inhibition. really open? Academy of Management Journal, 50(4), 869–884. Psychological review, 110(2), 265–284. Durso, G. R., Brinol,˜ P., & Petty, R. E. (2016). From power to inaction: Ambivalence gives Kickul, J., & Gundry, L. (2002). Prospecting for strategic advantage: The proactive pause to the powerful. Psychological Science, 27(12), 1660–1666. entrepreneurial personality and small firm innovation. Journal of Small Business Dutton, J. E., & Ashford, S. J. (1993). Selling issues to top management. Academy of Management, 40(2), 85–97. management review, 18(3), 397–428. Kim, Y. J., & Kim, J. (2020). Does negative feedback benefit (or harm) recipient Elliot, A. J., & Reis, H. T. (2003). Attachment and exploration in adulthood. Journal of creativity? The role of the direction of feedback flow. Academy of Management Personality and Social Psychology, 85(2), 317–331. Journal, 63(2), 584–612. Emerson, R. M. (1962). Power-dependence relations. American Sociological Review, 27(1), La Guardia, J. G., Ryan, R. M., Couchman, C. E., & Deci, E. L. (2000). Within-person 31–41. https://doi.org/10.2307/2089716. variation in security of attachment: A self-determination theory perspective on Fang, C., Kim, J. H., & Milliken, F. J. (2014). When bad news is sugarcoated: Information attachment, need fulfillment, and well-being. Journal of Personality and Social distortion, organizational search and the behavioral theory of the firm. Strategic Psychology, 79(3), 367–384. Management Journal, 35(8), 1186–1201. Lammers, J., Dubois, D., Rucker, D. D., & Galinsky, A. D. (2013). Power gets the job: Feeney, B. C., & Thrush, R. L. (2010). Relationship influences on exploration in Priming power improves interview outcomes. Journal of Experimental Social adulthood: The characteristics and function of a secure base. Journal of Personality Psychology, 49(4), 776–779. and Social Psychology, 98(1), 57–76. Lammers, J., Stoker, J. I., & Stapel, D. A. (2009). Differentiating social and personal Fiske, S. T. (1993). Controlling other people: The impact of power on stereotyping. power: Opposite effects on stereotyping, but parallel effects on behavioral approach American Psychologist, 48(6), 621–628. tendencies. Psychological Science, 20(12), 1543–1548. Fraley, R. C., Garner, J. P., & Shaver, P. R. (2000). Adult attachment and the defensive Lee, S., & Ling, L. (2007). Understanding affectional ties to groups from the perspective regulation of attention and memory: Examining the role of preemptive and of attachment theory. In M. A. Neale, E. A. Mannix, & C. Anderson (Eds.), Research on postemptive defensive processes. Journal of Personality and Social Psychology, 79(5), managing groups and teams (Vol. 10, pp. 217–248). Oxford, UK: Elseviser Science 816–826. Press. French, J. R. P., & Raven, B. (1959). The basis of social power. In D. Cartwright (Ed.), Lee, S., & Thompson, L. (2011). Do agents negotiate for the best (or worst) interest of Studies in Social Power (pp. 529–569). Ann Arbor, MI: University of Michigan Press. principals? Secure, anxious and avoidant principal–agent attachment. Journal of Frese, M., & Fay, D. (2001). 4. Personal initiative: An active performance concept for Experimental Social Psychology, 47(3), 681–684. https://doi.org/10.1016/j. work in the 21st century. Research in Organizational Behavior, 23, 133–187. jesp.2010.12.023. Frohman, A. L. (1997). Igniting organizational change from below: The power of LePine, J. A., & Van Dyne, L. (2001). Voice and cooperative behavior as contrasting personal initiative. Organizational Dynamics, 25(3), 39–53. forms of contextual performance: Evidence of differential relationships with big five Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of personality characteristics and cognitive ability. Journal of Applied Psychology, 86(2), Personality and Social Psychology, 85(3), 453–466. 326–336. Galinsky, A. D., Ku, G., & Wang, C. S. (2005). Perspective-taking and self-other overlap: LePine, J. A., & Van Dyne, L. (1998). Predicting voice behavior in work groups. Journal of Fostering social bonds and facilitating social coordination. Group Processes & Applied Psychology, 83(6), 853–868. Intergroup Relations, 8(2), 109–124. LePine, J. A., & Van Dyne, L. (2001). Voice and cooperative behavior as contrasting Galinsky, A. D., Magee, J. C., Gruenfeld, D. H., Whitson, J. A., & Liljenquist, K. A. (2008). forms of contextual performance: Evidence of differential relationships with Big Five Power reduces the press of the situation: Implications for creativity, conformity, and personality characteristics and cognitive ability. Journal of Applied Psychology, 86(2), dissonance. Journal of Personality and Social Psychology, 95, 1450–1466. 326–336. Galinsky, A. D., Magee, J. C., Inesi, M. E., & Gruenfeld, D. H. (2006). Power and Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of perspectives not taken. Psychological Science, 17(12), 1068–1074. power and status. The Academy of Management Annals, 2(1), 351–398. Gerber, A. S., & Green, D. P. (2012). Field experiments: Design, analysis, and Magee, J. C., Galinsky, A. D., & Gruenfeld, D. H. (2007). Power, propensity to negotiate, interpretation: WW Norton. and moving first in competitive interactions. Personality and Social Psychology Gillath, O., Hart, J., Noftle, E. E., & Stockdale, G. D. (2009). Development and validation Bulletin, 33(2), 200–212. https://doi.org/10.1177/0146167206294413. of a state adult attachment measure (SAAM). Journal of Research in Personality, 43(3), Main, M. (1983). Exploration, play, and cognitive functioning related to infant-mother 362–373. https://doi.org/10.1016/j.jrp.2008.12.009. attachment. Infant Behavior and Development, 6, 167–174. Grant, A. M., & Ashford, S. J. (2008). The dynamics of proactivity at work. Research in Mayseless, O. (2010). Attachment and the leader—follower relationship. Journal of Social Organizational Behavior, 28, 3–34. and Personal Relationships, 27(2), 271–280. Mayseless, O., & Popper, M. (2007). Reliance on leaders and social institutions: An attachment perspective. Attachment & human development, 9(1), 73–93.

40 J. Pai et al. Organizational Behavior and Human Decision Processes 167 (2021) 28–41

McClelland, D. C. (1987). Human motivation. CUP Archive. Seibert, S. E., Kraimer, M. L., & Crant, J. M. (2001). What do proactive people do? A Meyer, B., Olivier, L., & Roth, D. A. (2005). Please don’t leave me! BIS/BAS, attachment longitudinal model linking proactive personality and career success. Personnel styles, and responses to a relationship threat. Personality and Individual Differences, Psychology, 54(4), 845–874. 38(1), 151–162. Shaver, P. R., Schachner, D. A., & Mikulincer, M. (2005). Attachment style, excessive Mikulincer, M. (1997). Adult attachment style and information processing: Individual reassurance seeking, relationship processes, and depression. Personality and Social differences in curiosity and cognitive closure. Journal of Personality and Social Psychology Bulletin, 31(3), 343–359. https://doi.org/10.1177/0146167204271709. Psychology, 72(5), 1217–1230. Simmons, B. L., Gooty, J., Nelson, D. L., & Little, L. M. (2009). Secure attachment: Mikulincer, M. (1998). Attachment working models and the sense of trust: An Implications for hope, trust, burnout, and performance. Journal of Organizational exploration of interaction goals and affect regulation. Journal of Personality and Behavior: The International Journal of Industrial, Occupational and Organizational Social Psychology, 74(5), 1209–1224. Psychology and Behavior, 30(2), 233–247. Mikulincer, M., & Shaver, P. R. (2003). The attachment behavioral system in adulthood: Simpson, J. A., Rholes, W. S., & Phillips, D. (1996). Conflict in close relationships: An Activation, psychodynamics, and interpersonal processes. In M. P. Zanna (Ed.), attachment perspective. Journal of Personality and Social Psychology, 71(5), 899–914. Advances in experimental social psychology (Vol. 35, pp. 53–152). New York: Smith, P. K., & Hofmann, W. (2016). Power in everyday life. Proceedings of the National Academic Press. Academy of Sciences, 113(36), 10043–10048. Mikulincer, M., & Shaver, P. R. (2007). Attachment patterns in adulthood: Structure, Smith, P. K., Jostmann, N. B., Galinsky, A. D., & van Dijk, W. W. (2008). Lacking power dynamics, and change. New York, NY: Guilford Press. impairs executive functions. Psychological Science, 19(5), 441–447. Milliken, F. J., Morrison, E. W., & Hewlin, P. F. (2003). An exploratory study of employee Smith, P. K., Wigboldus, D. H., & Dijksterhuis, A. (2008). Abstract thinking increases silence: Issues that employees don’t communicate upward and why. Journal of one’s sense of power. Journal of Experimental Social Psychology, 44(2), 378–385. Management Studies, 40(6), 1453–1476. Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why Molero, F., Moriano, J. A., & Shaver, P. R. (2013). The influence of leadership style on experiments are often more effective than mediational analyses in examining subordinates’ attachment to the leader. The Spanish journal of psychology, 16, 1–10. psychological processes. Journal of Personality and Social Psychology, 89(6), 845–851. Morrison, E. W. (2002). Newcomers’ Relationships: The Role of Social Network Ties Staw, B. M., & Boettger, R. D. (1990). Task revision: A neglected form of work During Socialization. Academy of Management Journal, 45(6), 1149–1160. performance. Academy of Management Journal, 33(3), 534–559. Munafo, M. R., Clark, T. G., Moore, L. R., Payne, E., Walton, R., & Flint, J. (2003). Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley. Genetic polymorphisms and personality in healthy adults: A systematic review and Tost, L. P. (2015). When, why, and how do powerholders “feel the power”? Examining meta-analysis. Molecular Psychiatry, 8(5), 471–484. the links between structural and psychological power and reviving the connection Noe, R. A. (1986). Trainees’ attributes and attitudes: Neglected influences on training between power and responsibility. Research in Organizational Behavior, 35, 29–56. effectiveness. Academy of Management Review, 11(4), 736–749. Tost, L. P., & Johnson, H. H. (2019). The prosocial side of power: How structural power Ostroff, C., & Kozlowski, S. W. (1992). Organizational socialization as a learning process: over subordinates can promote social responsibility. Organizational Behavior and The role of information acquisition. Personnel Psychology, 45(4), 849–874. Human Decision Processes, 152, 25–46. Papini, M. R., & Bitterman, M. (1990). The role of contingency in classical conditioning. Van Dyne, L., & LePine, J. A. (1998). Helping and voice extra-role behaviors: Evidence of Psychological Review, 97(3), 396–403. construct and predictive validity. Academy of Management Journal, 41(1), 108–119. Parker, S. (2000). From passive to proactive motivation: The importance of flexible role Vuori, T. O., & Huy, Q. N. (2016). Distributed attention and shared emotions in the orientations and role breadth self-efficacy. Applied Psychology, 49(3), 447–469. innovation process: How Nokia lost the smartphone battle. Administrative Science Pietromonaco, P. R., & Barrett, L. F. (2000). The internal working models concept: What Quarterly, 61(1), 9–51. do we really know about the self in relation to others? Review of General Psychology, 4 Waytz, A., Chou, E. Y., Magee, J. C., & Galinsky, A. D. (2015). Not so lonely at the top: (2), 155–175. The relationship between power and loneliness. Organizational Behavior and Human Popper, M., & Mayseless, O. (2007). The building blocks of leader development: A Decision Processes, 130, 69–78. psychological conceptual framework. Leadership & Journal, Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007). The experiences in Close 28(7), 664–684. Relationship Scale (ECR)-Short Form: Reliability, validity, and factor structure. Richards, D. A., & Hackett, R. D. (2012). Attachment and emotion regulation: Journal of Personality Assessment, 88, 187–204. Compensatory interactions and leader–member exchange. The Leadership Quarterly, Welbourne, T. M., Johnson, D. E., & Erez, A. (1998). The role-based performance scale: 23(4), 686–701. Validity analysis of a theory-based measure. Academy of Management Journal, 41(5), Richards, D. A., & Schat, A. C. (2011). Attachment at (not to) work: Applying attachment 540–555. theory to explain individual behavior in organizations. Journal of Applied Psychology, Whitson, J. A., Liljenquist, K. A., Galinsky, A. D., Magee, J. C., Gruenfeld, D. H., & 96(1), 169–182. Cadena, B. (2013). The blind leading: Power reduces awareness of constraints. Roberson, L. (1990). Prediction of job satisfaction from characteristics of personal work Journal of Experimental Social Psychology, 49(3), 579–582. goals. Journal of Organizational Behavior, 11(1), 29–41. Wrzesniewski, A., & Dutton, J. E. (2001). Crafting a job: Revisioning employees as active Rucker, D. D., Dubois, D., & Galinsky, A. D. (2010). Generous paupers and stingy princes: crafters of their work. Academy of Management Review, 26(2), 179–201. Power drives consumer spending on self versus others. Journal of Consumer Research, Wu, C.-H. (2019). Employee Proactivity in Organizations: An Attachment Perspective. Policy 37(6), 1015–1029. Press. Rucker, D. D., & Galinsky, A. D. (2008). Desire to acquire: Powerlessness and Wu, C.-H., & Parker, S. K. (2012). The role of attachment styles in shaping proactive compensatory consumption. Journal of Consumer Research, 35(2), 257–267. behaviour: An intra-individual analysis. Journal of Occupational and Organizational Rucker, D. D., & Galinsky, A. D. (2009). Conspicuous consumption versus utilitarian psychology, 85(3), 523–530. ideals: How different levels of power shape consumer behavior. Journal of Wu, C.-H., & Parker, S. K. (2017). The role of leader support in facilitating proactive Experimental Social Psychology, 45(3), 549–555. work behavior: A perspective from attachment theory. Journal of Management, 43(4), Salancik, G. R., & Pfeffer, J. (1974). The bases and use of power in organizational 1025–1049. decision making: The case of a university. Administrative Science Quarterly, 19(4), Wu, C.-H., Parker, S. K., Wu, L.-Z., & Lee, C. (2018). When and why people engage in 453–473. https://doi.org/10.2307/2391803. different forms of proactive behavior: Interactive effects of self-construals and work Schaerer, M., du Plessis, C., Yap, A. J., & Thau, S. (2018). Low power individuals in social characteristics. Academy of Management Journal, 61(1), 293–323. power research: A quantitative review, theoretical framework, and empirical test. Zenger, J., & Folkman, J. (2017). Why do so many managers avoid giving praise. Organizational Behavior and Human Decision Processes, 149, 73–96. Harvard Business Review, 2. Retrieved from https://hbr.org/2017/05/whydo-so- Schwartz, D., Dodge, K. A., & Coie, J. D. (1993). The emergence of chronic peer many-managers-avoid-giving-praise. victimization in boys’ play groups. Child development, 64(6), 1755–1772. Zhang, X., & Bartol, K. M. (2010). Linking empowering leadership and employee creativity: The influence of psychological empowerment, intrinsic motivation, and creative process engagement. Academy of Management Journal, 53(1), 107–128.

41