The Perceived Influence Model of Trust: Toward a Multi-Trustee Theory

Abstract

Prior research investigating situations involving one trustor and multiple trustees often examines how a trustor’s trust in one party affects their amount of trust in another party. This paper fills a gap by predicting the effects of trust. The Perceived Influence (PI) Model of Trust is an individual-level model focused on the perceptions of a trustor. It builds upon the Mayer et al. (1995) model by integrating insights from literature on task interdependence and expanding to two trustees. The PI model describes and explains three possibilities for how a trustor’s trust in two trustees may combine to form a sense of aggregate multi-trustee trust via: (1) additive effects, such that the trustor’s trust in each of the trustees has independent effects on the aggregate; (2) compulsory effects, such that increasing the amount of trust in one trustee increases the effect of trust in the other trustee; and (3) compensatory effects, such that increasing trust in one trustee decreases the effect of trust in the other trustee. We propose that the theoretical mechanism explaining which of these three possibilities takes place is the trustor’s perceived influence of the trustees, which is tightly linked to perceptions of task requirements necessary to attenuate the trustor’s risk. The PI model begins to fill an important gap in the literature pertaining to pervasive, but rarely considered, multi-trustee situations, and proposes the importance of trustor perceptions of trustee influence and task requirements for future models of trust.

Keywords

trust trustee additive effects interaction effects influence control coordination cooperation collaboration competition

Imagine for a moment that you are a relatively new business owner who is just beginning to build your reputation. You are very passionate about your business—it creates novel and fantastical widgets, “the likes of which no one has seen before!” you say with a flourish. Your fledgling business is seeing some success, and you have hired two independent contractors to help full-time. Well, one helps full-time. In fact, “A” is industrious, reliable, and often works overtime. “B,” however, may have a drinking problem, is often late to work, and usually leaves early. “But, B is a creative genius,” you tell yourself as you turn to other matters. The matter at hand today is whether to take a new contract from a high-profile customer. You are brimming with cautious excitement as you study the proposal. You know that successful completion of the contract would provide a huge boost to your company’s reputation. However, given the stature and visibility of this customer, failure would probably destroy your business. Should you accept the proposal and sign the contract? Much depends upon whether you can trust A and B to fulfill it successfully. The question is: How do you decide?

Trust—a person’s willingness to rely upon or otherwise be vulnerable to a trustee based on positive expectations of the trustee in a situation involving risk (Currall & Inkpen, 2002; Mayer et al., 1995; PytlikZillig & Kimbrough, 2016; Rousseau et al., 1998)—has been extensively investigated at the dyadic level, resulting in a vast literature on its psychological mechanisms (e.g., Thielmann & Hilbig, 2015). However, the focus on one trustor and one trustee often fails to represent reality. In many cases, a trustor must evaluate multiple distinct trustees before determining whether to take a “trusting action” in a situation involving risk. Individual employees may need to simultaneously consider trust in multiple colleagues, or trust in both a supervisor and an administrator, when considering decisions pertaining to job satisfaction or organizational citizenship. Similarly, a consumer may work with multiple specialists, such as when a cancer patient works with both an oncologist and a primary provider.

As we review later, prior research does not entirely ignore such situations; however, research investigating one trustor’s simultaneous trust in multiple distinct trustees (i.e., trustees that do not comprise a team per se) is scant and not guided by any clear theoretical model. Almost all quantitative studies incorporating trust in multiple trustees as predictors assume trust combines additively (composition emergence, Kozlowski & Klein, 2000), with trust in each trustee independently and incrementally contributing to some outcome. Because of the lack of theoretical attention given thus far, the current literature often overlooks the strong possibility that other forms of emergence (e.g., compilation) are possible.

Returning to the opening example, consider the different ways a new business Owner¹ might weigh their trust in two contractor-employees, depending on the nature of the proposed contract. If the contract requires production of a certain number of widgets quickly and independently by each employee, Owner’s additive or average trust across one hardworking and one lazy employee might be a good indicator of aggregate trust in the employees to meet the goal. However, if a contract requires a new widget design, Owner may take the contract if they believe employee B’s genius suffices or compensates for low or uncertain trust in A’s creativity. In yet a third situation, if the production of the widgets requires unique qualifications and contributions from each employee, then Owner may need a high amount of trust in both employees before being willing to accept the contract. In this last case, trust in both employees is compulsory and Owner prioritizes trust in the least trusted employee when making their decision.

Despite the likely prevalence of compensatory and compulsory emergence suggested by the last two scenarios, empirical examinations rarely consider them. Current theoretical models overlook these scenarios and therefore do not explain why nor predict when they might occur. Teams research and research on multilevel trust have begun to acknowledge that multiple forms of emergence of trust across levels (e.g., from individual to group) are possible (Costa et al., 2018; De Jong & Dirks, 2012), but this research is still in its infancy. To our knowledge, no existing treatment of the emergence of trust across levels considers trust × trust interactions: How and when trust in one entity might change the effects of trust in another (e.g., Owner’s high trust in employee A making trust in employee B irrelevant or unnecessary). Our conceptual model, the Perceived Influence (PI) Model of Trust, begins to fill this gap by addressing the question: When a trustor’s risk in a situation is affected by multiple distinct trustees, what determines how the trustor formulates their aggregate trust in, and willingness to rely on, those multiple trustees?

The contributions of the PI model include both theoretical advancement and opportunities for practical application. In the realm of practical applications, a greater understanding of situations in which a trustor need not trust every trustee before taking positive action provides a novel and promising avenue for dealing with situations when low trust, or even active distrust in at least one trustee, seems impossible to avoid (e.g., workplace conflicts). More broadly, the PI model extends opportunities to explain more real-world situations with greater precision. In an era of increasing role specialization, trustors often must assess their trust in multiple distinct trustees. A model predicting and explaining the real-world differential weighing of trust in such situations enhances understanding and increases opportunities for positive impacts.

Pertaining to theoretical advancement, the PI model addresses calls to expand the investigation of trust beyond dyadic trust to situations involving more than two actors (e.g., Dirks & de Jong, 2022). The model contributes to a nascent, but burgeoning, interdisciplinary literature moving beyond the single trustor-trustee dyad and exploring the more complicated dynamics involved when one trustor is simultaneously relying upon multiple trustees (e.g., Birmingham et al., 2020; Cho et al., 2014; Lu et al., 2010). The PI model does this, first, by focusing on the emergence of multi-trustee trust, aggregate trust in multiple distinct individual trustees. Second, by identifying contextual concepts (perceived trustee influence and perceived task requirements) which can be integrated into existing models as explanatory and moderating variables, it also addresses calls for greater attention to and explication of “context” in trust research (Schoorman et al., 2007; Spreitzer & Mishra, 1999).

Third, the PI model addresses a lacuna in the research on trust in multiple trustees, specifically, the dearth of research and theory pertaining to trust × trust interactions. The PI model details three possible effect scenarios, two of which (compulsory and compensatory) involve predictions that trust in one trustee will impact (modify) the importance (effects) of trust in the other trustee. While it is common for models to predict that trust in one trustee may affect the amount of trust in another trustee (e.g., Fulmer & Ostroff, 2017), how trust in one trustee might change the impact (i.e., importance or strength of effect) of trust is rarely considered in current trust studies.

In this paper, we first describe the PI model, its foundation in the Mayer, Davis, and Schoorman (MDS) model of organizational trust (Mayer et al., 1995; Schoorman et al., 2007), and its extensions drawn from conceptualizations of task interdependence (Rizk et al., 2023; Steiner, 1972). After arguing for the importance of perceived trustee influence and task requirements as moderators of the trust-outcome relationship, we then demonstrate how to apply these concepts to one- and two-trustee models. In this analysis, we describe possibilities for “how” multi-trustee trust might emerge from trust in two trustees (i.e., additively vs. interactively via compensatory or compulsory weighting), “when” such weighting might occur (i.e., under different task requirements), and “why” (i.e., due to perceptions of joint influence). Following the description of our model, we compare the PI model with other models and literatures that consider multiple trustees. Finally, we conclude by suggesting future research to advance our model and reduce the research gap concerning multiple trustees.

The Perceived Influence Model of Trust

To explain the psychological processes an individual trustor engages in when formulating multi-trustee trust, the PI model draws heavily from social cognition approaches (Smith & Semin, 2007; Tory Higgins, 2000) and the MDS model of trust (Mayer et al., 1995). The MDS model was proposed to clarify and explain interpersonal trust in organizational contexts and has since been applied across various disciplines and contexts (Schoorman et al., 2007). The PI model also draws from theories of emergence (Kozlowski & Klein, 2000) and conceptualizations of task interdependence (Courtright et al., 2015; Puranam, 2018). Before delving into the details of the PI model, we provide an overview and background on its foundational definitions, assumptions, and boundary conditions. Then, we describe the central conceptual contributions of the PI model. Next, we apply those concepts to the case of a one trustee model, followed by an expansion to a multiple (two) trustee model.

Model Overview

The heart of the PI model is the idea that how a trustor weighs their trust in each trustee depends on two highly related factors: (1) what is needed or required to attenuate the trustor’s risk (i.e., the perceived task requirements) and (2) the trustor’s perception of the trustees’ joint potential for influencing the trustor’s desired end (i.e., their perceived joint influence). The effect of individual amounts of trust in each trustee on the trustor’s aggregate trust and trust-relevant decisions and behaviors will vary depending on these factors.

As shown in Figure 1, the PI model proposes that, when a trustor is faced with a situation involving multiple trustees, the trustor will assess three main aspects of the situation. First, the trustor is aware of their vulnerability and risk. Trust is not relevant unless there is risk (PytlikZillig & Kimbrough, 2016). This risk is tied to the trustor’s preference for a particular end or goal, which is the primary source of trustor vulnerability, because, depending on the situation and various threats, their desired end may not be met.

Figure 1.

Conceptual overview of the Perceived Influence (PI) model of multi-trustee trust.

Second, the trustor is aware of the trustees who have influence over the trustor’s risk in the situation, as well as the trustor’s amount of trust in each trustee. A trustee’s influence is conceptually subsumed under the requirement of “interdependence” between the trustor and trustee (PytlikZillig & Kimbrough, 2016). However, most conceptualizations of trust gloss over the fact that the nature and extent of trustor-trustee interdependence and trustee influence may vary across specific trustor-trustee dyads. The PI model explicitly recognizes this variance and proposes it is an important, but often overlooked, aspect of trust situations.

The third aspect of the situation considered by the trustor is not typically referenced in prior treatments of trust and comprises a major contribution of the PI model. As the trustor attends to “what is needed” to achieve the trustor’s desired end, they think about the task requirements. When multiple trustees are involved, perceptions of such task requirements include perceptions of task interdependence (Steiner, 1972). These perceptions drive additional trustor perceptions of the amount and type of shared or joint influence which the trustees have over the risk. Said differently, the trustor looks at the actions or tasks that will impact the trustor’s risk, and considers whether, how, and how much each trustee, individually and jointly, influences their desired end through those tasks.

Next, the PI model proposes it is the trustor’s perception of trustee joint influence, particularly the extent of perceived cooperative influence, which drives how the trustor differentially weighs their trust in each trustee (i.e., the trust × trust interaction) to formulate their aggregate trust, resulting in three potential effect scenarios: additive, compensatory, and compulsory, which are detailed under the PI model description.

Foundations and Definitions

Consistent with its foundations in the MDS model, the PI model focuses on trust as the trustor’s psychological state of willingness to rely upon and/or be vulnerable to some entity—or, our model argues, multiple entities—termed trustees. In other words, trust is the “willingness to take risk” (Mayer et al., 1995, p. 712; Schoorman et al., 2007, p. 346) or, more specifically, willingness to take “risk in relationship” with the trustee(s) upon whom the trustor depends. Within an individual psychological framework, trust is experienced by the trustor as an affective and motivational state (e.g., see analysis by PytlikZillig & Kimbrough, 2016) which varies over time and across situations (Audrey Korsgaard et al., 2018). In any given moment, a trustor’s trust is directed toward a specific target/trustee, exists in the context of a situationally specific goal, and exists somewhere on a continuum from high to low. Consistent with the social-cognition perspective, a trustor can perceive multiple social actors (trustees) at once. The trustor’s states of trust toward such trustees are dependent upon characteristics of the trustor (e.g., propensity to trust) and the trustor’s subjective perceptions and interpretations of the “situation,” which includes characteristics of the individual trustee(s) and other specific aspects of the context.

The PI model identifies and explains ways in which multi-trustee trust is derived from trust in individual trustees. A trustee is defined as an entity the trustor perceives as having influence over the trustor’s risk (where risk refers to the potential for a negative outcome; Currall & Inkpen, 2002) in the situation. Meanwhile, multi-trustee trust is the trustor’s sense of aggregate trust in multiple, distinct, simultaneously considered trustees. It is important to emphasize that the model distinguishes itself from other models which have multiple trustees by focusing on trust × trust interactions : how trust in one trustee impacts the effects of (rather than the amounts of) trust in another trustee—that is, how strongly trust in one trustee affects a trust-relevant or trusting outcome. In contrast to our focus on trust × trust interactions, the MDS model also includes consideration of the moderation of trust effects, primarily by perceived risk. As we elaborate later, MDS explicitly define perceived risk as risk from the situation which is not attributable to the (one) trustee. The MDS model suggests that perceived risk may moderate the extent to which a given amount of trust leads to the trusting outcome of risk-taking in relationship, because the trust must be sufficient to offset the other situational risks.

By trusting outcome we generically refer to those outcomes theorized in the literature as directly or indirectly caused or affected by a psychological state of trust. This includes trusting actions like MDS’s construct of risk taking in relationship (e.g., Owner taking the contract and depending on their employees), and possibly many other proposed outcomes (Dirks & de Jong, 2022). However, the primary outcome we focus upon in this paper is multi-trustee trust, which the PI model proposes mediates the impacts of trust in individual trustees upon other outcomes.

Caveats and Boundary Conditions

Before detailing the PI model, it is important to mention a few caveats and boundary conditions. Consistent with a social cognitive approach, and to use the distinctions offered by Fulmer and Gelfand (2012), the model occurs entirely “at” the level of the individual trustor—i.e., the level of analysis is the individual trustor who is a single person. Thus, the model does not cover other levels such as trust “at” the level of the team. At the same time, our model explicitly examines how the trustor’s trust in individual trustees might transform into multi-trustee trust, making it inherently applicable to discussions of multilevel trust (Costa et al., 2018).

The PI model focuses on how an aggregate multi-trustee trust emerges from consideration of trust in each individual trustee. Thus, the model only applies when a trustor can separately ascertain their trust in multiple distinct trustees. Also, when moving beyond more than one trustee, we could discuss any number of trustees, from two to infinity. For the sake of parsimony, however, in this article the PI model focuses on one trustor and two trustees. Later, in the discussion, we consider the implications of this model for more than two trustees.

Finally, in this initial articulation of the PI model, we focus on the case of interpersonal trust (trustors and trustees who are individual persons). By focusing on two trustees and interpersonal trust, we hope to clearly articulate the important underlying psychological concepts that drive the model propositions. However, in the discussion we consider how the model applies to distinct trustees that can vary in type (e.g., organizations, technologies).

Perceived Influence

As indicated by the model’s name, drawing attention to perceived influence comprises a primary and central contribution of the PI conceptual model. We define perceived influence as the trustor’s perceptions of the extent and nature by which the trustee(s) actions could increase or decrease the trustor’s risk in the situation. For example, Owner may perceive their employees as having more or less influence over successfully meeting the contract. Employees may have high influence over the speed of production of the widgets but low influence over availability of widget source materials. If the customer asked for something Owner feels is impossible (a widget to control the weather!), then Owner may perceive employees as having zero influence over contract success.

Although “influence” is not a word used in the original MDS definition of trust, its importance is implied by the MDS model references to trustee actions and the concept of “risk taking in relationship” which separates risk from the trustee and risk from other sources. The importance of perceived trustee influence is also clear in research operationalizing trust as the expressed willingness to allow a trustee to have influence or control over issues of importance to the trustor and willingness to forgo monitoring the trustee (e.g., Mayer & Davis, 1999). Also, the MDS model implies trustee influence over risk when it describes the trust as taking “risks at the hands of the trustee” (Schoorman et al., 2007, p. 347).

In current models of trust, the influence of the trustee is first assumed and then ignored. However, encompassing influence within the definition of trust or as a boundary condition fails to acknowledge and account for the fact that trustee influence may vary from low to high, which we will subsequently argue has implications for predicting trusting outcomes from trust.

Perceived Task Requirements

“Throughout the history of organizational research on groups, one sentiment has been pervasive: task type matters” (Beal et al., 2003, p. 992). Consistent with this sentiment, the PI model proposes that it is the perceived task requirements that either implicitly or explicitly determines the trustor’s perception of the trustee(s) influence. Broadly, the perceived task requirements are the features of the situation-context that determine what is needed to eliminate risk and achieve the trustor’s desired ends. The phrase “task requirements” implies a necessary action or inaction which would attenuate the trustor’s risk. The PI model’s emphasis on both task requirements and trustee influence underscores that trustors are not generically evaluating a trustee’s trustworthiness characteristics. Instead, they are evaluating specific trustworthiness characteristics relevant to trustee influence in the context of the situation’s task requirements.

Using the Owner example, to successfully fulfill a contract and protect Owner’s reputation, the task requirements involve actions the employees might engage in to create widgets (e.g., creativity, competence, speed, and/or reliability). However, in other contexts the task requirements could involve actions such as keeping confidential information to themselves (requiring integrity), protecting the relationship of a colleague (requiring benevolence), etc. Task requirements and trustee influence are thus closely linked. However, they are distinguished by the questions of “what is needed?” (task requirements) and “how much might a trustee affect that which is needed?” (influence). As evident in our examples, not all requirements are things which trustees will be expected to be able to control, even if they are very trustworthy. In this definition, we also explicitly focus on the perception of the trustor. If perception and reality do not align, it is the perception—not objective reality—that is influential.

Applying the PI Model to One Trustee

Having introduced the concepts of perceived task requirements and perceived trustee influence, we next apply them to a simplified model with only one trustee. We adopt this approach, initially, to highlight how these constructs can also enhance trust models with one trustee.

Perceived Influence of One Trustee

Figure 2 includes relevant constructs from the original MDS model in white boxes, and PI model constructs in shaded boxes. Within the larger MDS one-trustee model, the PI model expands consideration of the relationship between trust in the trustee and the trusting outcome by adding perceived influence as a moderator of that relationship. In the case of a single trustee, the PI model proposes trustors not only perceive whether trustee influence is present, they also perceive “how much” influence the trustee wields. If a given entity has no influence over risk in the situation, then the entity is not a trustee: i.e., there is no need to trust or depend upon the entity in that specific situation. Thus, the trustor is unlikely to engage in the evaluations necessary to support a sense of trust/distrust relative to that entity in the specific situation. However, like similar concepts in social psychology such as self-efficacy (Bandura, 1997) and perceptions of one’s own “perceived behavioral control” (Ajzen, 1991), perceptions of trustee influence are not all-or-nothing; they likely vary continuously. Further, as perceived influence increases, the trustee’s “trustworthiness” characteristics (e.g., extent of creativity, competence, speed) are also perceived as positively or negatively affecting the trustor’s desired end. Therefore, the effect size of trust also increases, leading to our models’ first proposition.

Figure 2.

Applying the concepts of perceived task requirements and perceived trustee influence to the MDS one trustee model. Note: Because we are focused on expanding a subsection of the MDS model, we have left out of Figure 2 a number of the elements of the full MDS model, which we assume would operate similar to how they operate in the original model.

Proposition 1

Perceived influence of a trustee moderates the effect of trust in that trustee on trusting outcomes: The greater the perceived influence of a trustee over risk in the situation, the greater the effect of trust in that trustee on trusting outcomes.

Importantly, the extent of influence a trustee wields is separate from a trustor’s evaluation of the trustee’s trustworthiness. Consider the Owner example, this time with only one employee. The PI model proposes Owner first perceives the task requirements and trustee influence (e.g., “create widgets quickly” may be a task requirement the Owner expects an employee to have influence over). Next, Owner gauges their trust in the specific employee to meet the task requirements. Because Owner perceives task success as dependent on the employee (i.e., high influence), low trust in the employee has a strong effect on Owner’s decision to take the contract. One might imagine a different situation in which the Owner felt the client requesting the widgets was asking for an impossible-to-create product. Owner may perceive the employee as having significantly less influence over the desired end of contract success. Then, even if Owner has high trust in the employee, Owner’s trust may only weakly impact the decision to take the contract.

Task Requirements of One Trustee

As illustrated in the examples above, the PI model proposes that trustor perceptions of task requirements and trustee influence are linked. If key features of the situation-context driving the task requirements change, then the perceptions of trustee influence may also change. The task of creating widgets requires both obtaining materials and manufacturing processes. Potential supply chain issues may reduce trustee influence over the task of obtaining materials. Likewise, the perceived influence of the employee over widget production may change if there is a groundbreaking new widget-producing technology available, thereby potentially increasing or decreasing the employee’s influence over widget production.

Proposition 2

Trustor perceptions of the perceived influence of a trustee are a function of the perceived task requirements: Tasks will vary in what a trustee must or can do to eliminate the trustee-relevant risk, thereby also changing perceived trustee influence.

Again, the PI model conceives of perceived task requirements as separate from trustor trust. The PI model assumes, but does not focus on, the trustworthiness aspect of the MDS model (where trustworthiness is defined as characteristics of the trustee contributing to trustor trust). Thus, our simplified figures do not show the trustworthiness factors. Nonetheless, consistent with findings from studies of disclosure-based and reliance-based trust (Gillespie, 2003; Tomlinson et al., 2020), it is likely that task requirements affect which specific trustworthiness characteristics of a trustee are considered most important for assessing one’s trust in that trustee, without driving trust per se. For example, consider if Owner needs a creative solution to enhance widget capacities versus quick production of widgets. In these cases, Owner may use the different trustworthiness criteria of creativity versus speed to ascertain their trust in the employee. For other ends desired by a trustor, such as organizational citizenship, trustworthiness characteristics of benevolence or integrity may be more important. This is consistent with findings from studies of disclosure-based and reliance-based trust which seem to acknowledge that trustees’ influence may be specific to instrumental versus emotional ends (Gillespie, 2003; Tomlinson et al., 2020).

Applying the PI Model to Two Trustees

Having considered the PI model concepts of perceived task requirements and perceived trustee influence as applied to one trustee, we now turn our attention to how a trustor determines whether to be vulnerable to multiple (two) distinct trustees.

In the MDS model, only one trustee is specified. Any additional trustees (and their influence) are implicitly assumed to fall under the MDS perceived risk construct (“perceived risk (other than trustee),” in Figure 2). To expand to two trustees, it is necessary to account for the impact of both trustees by explicitly considering trust in trustee A and trust in trustee B. In so doing, the PI model recognizes the potential for an emergent trust cognition: multi-trustee trust. We detail the PI model with two trustees (see Figure 3), beginning with a description of the new concept of multi-trustee trust. This is followed by a discussion of how perceived task requirements and perceived influence are modified by the inclusion of an additional trustee. From there, we discuss how perceived joint influence impacts the emergence of multi-trustee trust and revisit the three possible scenarios for how multi-trustee trust emerges.

Figure 3.

Expanding the perceived influence model to two trustees. Note: To simplify model depiction, we did not draw in P1 and P2 effects from Figure 2. We assume these effects are subsumed within the two-trustee model.

Multi-Trustee Trust

As shown in Figure 3, we propose the construct of multi-trustee trust as both an outcome of trust in individual trustees and a mediator between trust in individual trustees and some trusting outcome.

Proposition 3

(mediation): Multi-trustee trust mediates the effects of trust in each of the multiple trustees upon the trusting outcome (e.g., risk-taking in relationship).

We define multi-trustee trust as a trustor’s aggregate assessment of their willingness to be vulnerable to the multiple, distinct trustees simultaneously. The concept of multi-trustee trust reflects a holistic trust assessment arising from individual-level trust in individual trustees. Thus, our conceptualization of multi-trustee trust has parallels with the conceptualization of a sense of general trust as arising from specific trust beliefs (Chen et al., 2014).

The PI model further proposes multiple ways multi-trustee trust emerges from trust in individual trustees (Kozlowski & Chao, 2012; Kozlowski & Klein, 2000). The three scenarios pertaining to additive, compensatory, and compulsory trust effects reflect different patterns of emergence: additive effects reflect composition emergence, whereas compensatory and compulsory effects reflect compilation emergence. At the same time, the additive effects and two qualitatively different interaction effects also describe different points on a continuum. The scenarios are a function of whether and how a trust × trust interaction occurs between trust in trustee A and trust in trustee B. The P4 arrow in Figure 3 indicates this interaction effect.

Proposition 4

(trust × trust interaction): Trust in some trustees in the situation can interact with trust in other trustees in their effect upon multi-trustee trust. This interactive effect can be positive, zero, or negative.

Interaction effects occur when two predictor variables non-additively combine to produce an effect on an outcome variable. The P4 arrow could have been drawn from Trust in A to the effect of Trust in B with equivalent meaning and effect. Statistically, the interaction effect is more than the sum of their parts. It is a unique effect due to the combination of the two predictor variables. These effects can be numerically positive, negative, or zero, resulting in qualitatively different effects, which the PI model classifies as compensatory, compulsory, and additive effects, respectively. Consistent with the one-trustee model, in the two-trustee model we propose the trust × trust interaction effect is moderated by perceptions of the influence of two trustees (which are linked to perceived task requirements of two trustees).

Perceived Influence of Multiple (Two) Trustees

When considering two trustees, there is the potential for varied types of shared or joint influence. Thus, perceived joint influence —how and how much the trustees, together, are perceived to influence (increase or decrease) the trustor’s risk—is likely to be more complex than individual influence. Conceptually speaking, perceptions of the influence of each individual trustee are nested within and affect perceptions of joint influence.

In many cases, joint influence is not simply “the sum of” individual influences. For example, there may be a unique effect attributable to perception of how the trustees’ influence is inherently interconnected. Collaboration costs refer to the extra hassles of working with other parties (e.g., extra efforts to solve conflict) and their results (e.g. delays, budget overruns) (Hansen, 2009). If collaboration costs are high, the benefits of working together may be less than the sum of the benefits of working separately. Meanwhile, synergistic effects, or effects from partnering synergies, reflect situations in which benefits to cooperating entities are achieved above and beyond individual benefits (Venkatesh & Bala, 2012).

While collaboration costs and partnering synergies may seem to imply that two trustees are interacting directly, this is not always the case. The PI model applies even in cases of low or no interaction between trustees. For example, if Owner’s employees, two independent contractors, separately provide different widget components, the contractors may share joint influence over contract success even without interacting.

These examples lead us to suggest one specific type of joint influence which is likely to act as an important moderator: Across situations, multiple trustees may share perceptibly different degrees of cooperative influence, which we define as the extent to which the trustees’ joint or combined influence is greater than or less than the sum of their individual influences. Cooperative influence is closely related to task interdependence, discussed in the next section (Rizk et al., 2023, citing Puranam, 2018). To assess cooperative influence, one would first measure the benefits of each trustee behaving in a trustworthy manner alone, without the other trustee being trustworthy. Then one would sum these independent contributions (= summed individual influences). Next, one would assess the benefits gained from both trustees behaving in a trustworthy manner (= total joint influence). Finally, the difference between joint influence and the summed individual influence (joint – summed) is the extent of cooperative influence. Cooperative influence thus is continuous and can be positive, negative, or zero. In the case of collaboration costs, cooperative influence would be negative as joint influence is less than the summed individual influences (the PI model denotes these as compensatory effect scenarios). In the case of partnering synergies, cooperative influence would be positive as the joint influence would be greater than the summed individual influences (the PI model denotes these as compulsory effect scenarios).

The PI model proposes that perceptions of the cooperative influence of multiple trustees can affect whether the trust in trustee A by trust in trustee B (trust × trust) interaction effect is additive, compensatory, or compulsory. As such, the nature of the trust × trust interaction effect depends on perceptions of cooperative influence. In this instance, moderation represents the change in the trust × trust interaction effect on multi-trustee trust that is due to variation in a specific type of perceived cooperative influence, which is one type of joint influence. Statistically, this would be represented by a three-way interaction between the two independent variables, trust in trustee A and trust in trustee B, and cooperative influence. This is indicated by the P5 arrow in Figure 3 and can be stated as the following proposition.

Proposition 5

(three-way interaction effect): The interactive effect of trust in trustee A and trust in trustee B on multi-trustee trust depends upon the perceived joint influence of the trustees (e.g., perceptions of cooperative influence).

Perceived Task Requirements of Multiple (Two) Trustees

As a final element of the PI model, we propose perceived joint influence is a contextual variable affected by the nature of the perceived task requirements needed to reduce trustor risk. That is, perceptions of cooperative influence—one type or means of achieving “perceived joint influence”—are affected by trustor’s perceptions of the task requirements . This is shown by the P6 arrow in Figure 3 and summarized by the following proposition.

Proposition 6

Trustor perceptions of the task requirements (e.g., task interdependence) will impact perceptions of trustee joint influence (e.g., cooperative influence).

As previously noted, “task requirements” is a broad concept and the literature has yet to fully elucidate what trustors are attending to in order to ascertain perceptions of joint influence. Nonetheless, there are some clues in prior literatures. For example, one construct likely to inform task situation-context perceptions is task interdependence. We define task interdependence as the extent to which unique contributions and influences from each of the trustees are necessary to achieve overall task success (c.f., Courtright et al., 2015; De Jong et al., 2016; Rizk et al., 2023). Within the PI model, task interdependence occurs even when the tasks leading to the trustor’s desired outcomes involve, or are affected by, individuals who may not interact at all. It is the contributions and influences of each trustee that “work together” toward or against the trustor’s desired end, not necessarily the individual trustees.

The concept of task interdependence has its roots in the literature on the social science of teams (e.g., Steiner, 1972), which identifies specific tasks of interdependence (conjunctive, disjunctive, and additive). Drawing from a recent proposal that task interdependence reflects a difference in benefits gained from performing one task, depending upon performance of another task (Puranam, 2018; Rizk et al., 2023), we suggest disjunctive, additive, and conjunctive tasks actually lie upon a continuum in which a task provides greater, lesser, or equal benefits, dependent upon another task being performed. In addition, task requirements have direct implications for how trust in differentially trusted trustees will be weighed.

Conjunctive tasks are high in interdependence. These are tasks in which members must contribute their unique resources to achieve success. They are defined by greater benefits accruing from one task if another is also performed. In the Owner assembly line contract scenario, the benefit of adding part1 to the widget is much greater if part2 is added (and vice versa). Since maximum success depends upon the weakest contributor, trust in the least trusted trustee should be given greatest weight.

In contrast, disjunctive tasks can be completed successfully and independently by a single actor. These tasks are structured such that there is less benefit of a task if another has been performed. There is great benefit of finding a new widget design for Owner if no one else has done so. However, once the optimal solution is found, there is much less benefit from another person also finding it. Since maximum success is dependent upon the strongest contributor, trust in the most trusted trustee should be given greatest weight.

Purely conjunctive and disjunctive tasks have either/or success states: Success or no success. Solutions to these tasks are “optimized” rather than “maximized” (Steiner, 1972). Additive task success, however, varies more continuously and improves with each added but independent contribution. Additive tasks result in equal benefits regardless of whether another task was performed: If maximum supply is the goal, creating one widget beneficially moves another step further toward the Owner’s desired end, even if another widget was already created. This implies trust in each trustee would be additively considered. Note, the placement of tasks on a continuum of interdependence is congruent with, and tightly linked to, the concept of cooperative influence discussed in the last section.

Bringing it All Together: Three Different Trust Effect Scenarios

Table 1 summarizes the alignment of the three effect scenarios and major PI model constructs. As shown, the PI model proposes that multi-trustee trust emerges due to trust × trust interaction effects which vary on a continuum from negative (compensatory) to zero (additive) to positive (compulsory). As previously noted, the trustor’s vulnerability is the same in the scenarios. What varies is the perceived task requirements (what is needed from the trustees) and the perceived influence of the trustees over the desired end. The following subsections detail these effect scenarios, providing more detail on the nature of these effects.

Table 1.

Alignment of Model Constructs with Each of the Three Scenarios.

Scenario/Effects	Task Requirements	Perceived Joint Influence	Trustor Conclusions	Trust × Trust Interaction	Trust Weighting
Compulsory trust effects	Conjunctive: Achieved end is as good as efforts of least trusted trustee	High cooperative: Trustee A success MORE valued if trustee B succeeds	“I need to trust all trustees or I will not get my desired end”	Positive: Trust in B increases effect of trust in A	Trust in least trusted trustee is weighed heavier than trust in most trusted
Additive trust effects	Additive: Each trustee's efforts incrementally influence end achieved	Medium cooperative: Trustee A success valued independent of trustee B success	“Each trustee independently increases my desired end”	Zero: Trust in B has no impact on effects of trust in A	Trust in each trustee is additively combined, unaffected by trust in others
Compensatory trust effects	Disjunctive: Achieved end is as good as most trusted trustee	Low cooperative: Trustee A success valued LESS if trustee B succeeds	“I need only to trust one trustee to achieve my desired end”	Negative: Trust in B decreases the effect of trust in A	Trust in most trusted trustee is weighed heavier than trust in least trusted

Note. Corresponding to Figure 3, it is the trust × trust interaction (P4) which varies due to perceived joint influence (P5) which results from perceptions of task requirements (P6).

Additive Effects

The arrow from Trust in B to the effect of Trust in A on multi-trustee trust indicates that trust in B and trust in A may interact—i.e., one may increase or decrease the effect of the other on multi-trustee trust. If this interactive effect is zero, all that is left are the additive effects of trust in A and trust in B on multi-trustee trust. To return to the Owner example, if the potential contract requires that each employee create a certain number of widgets by a given deadline, then task interdependence is low but not zero. Each employee can independently and incrementally contribute to the number of widgets needed. Owner may have different amounts of trust in each employee relating to this task because one employee is fast and the other slower at making the widgets. If Owners’ trust summed across both employees suffices, they will accept the contract.

Compensatory Effects

If, when rating trust in the trustees, higher values indicate more trust, and the sign of the interaction effect is negative instead of zero, it means that the more the trustor trusts in trustee B, the less it matters whether they trust in trustor A (and vice versa). That is, trust in one trustee can partially or completely compensate for trust in the other. At the extreme, the trustor’s trust in one trustee could make trust in the other trustee irrelevant and completely unnecessary. This would occur if what Owner needs from the employees (e.g., come up with a novel widget design) could be achieved mostly or entirely by only one employee. In this case, task interdependence is even lower than in the additive scenario. If Owner perceives their “genius” employee B as particularly creative and well-suited for this task, Owner may have high multi-trustee trust and accept the contract regardless of their amount of trust in employee A. Trust in A is nice, but not necessary. Owner weighs more heavily their trust in the most trusted (genius) employee, while giving less weight to their trust in the least trusted employee.

Compulsory Effects

Interaction effects can also be numerically positive. If there is a positive interaction effect, it means that trust in one of the trustees increases rather than decreases the effect of the other trustee. At the extreme, this would mean trust in both trustees would be required (compulsory) to achieve some trust-relevant outcome. In the compulsory scenario, Owner’s widgets may require an assembly line approach to production whereby employees A and B each create a different widget part. In this case, task interdependence is very high. Neither employee can even create one widget without the help of the other employee. Thus, trust in each employee is compulsory to successfully meet the needs of the contract and trust in the least trusted employee is weighed more heavily than one’s trust in the most trusted employee when deciding to take the contract and the risks implied to Owner’s reputation.

Discussion

The PI model of trust and multi-trustee trust proposed here especially contributes to prior research pertaining to situations involving multiple trustees. Here we discuss in greater detail how the PI model contributes to areas of prior research by comparing and contrasting work in each area with the PI model and its predictions.

First, however, it is important to note that the work we next review does cross levels of analysis. A desire for clarity and consistency with the original MDS model led us to focus our explanation of the PI model upon interpersonal trust situations (c.f., Dirks & de Jong, 2022; Fulmer & Gelfand, 2012). While the MDS model intentionally focused on interpersonal trust, the authors originally chose model constructs that would apply across levels (e.g., interpersonal and organizational levels; Schoorman et al., 2007). Thus, researchers have often successfully applied the MDS model of trust across different trustees representing different types and levels (Schoorman et al., 2007). The PI model is expected to also apply beyond interpersonal trust and help explain how individuals form trust “in” multiple entities of different types (e.g., persons, organizations, institutions, technologies).

Relation to Prior Research Considering Multiple Trustees

Prior research that has considered one trustor and multiple trustees has included research on trust transfer, third party trust (trust brokering), social networks, and studies which included multiple trustees. The PI model treats multi-trustee trust somewhat differently than each of these other bodies of work as we discuss in the following subsections.

Trust Transfer

Trust transfer involves a trustor transferring trust from one more well-known trustee to a lesser-known trustee, which implies a two-step process. Trust transfer (or trust spillover) models (e.g., Fulmer & Ostroff, 2017; Stewart, 2003) highlight that one’s trust in a known trustee A can affect trust in a second trustee B with whom a trustor may be less familiar. The trustor sees trustee A as an implied or explicit source of information regarding trustee B. Theories and studies of trust transfer typically examine statistical mediation models in which a trustor’s trust in one trustee increases trust in another trustee, thereby allowing the trustor to feel more comfortable engaging in behavior that may make them vulnerable to that lesser-known second trustee (e.g., Belanche et al., 2014; Esmaeilzadeh, 2019). For example, employee trust in a firm or upper management may transfer to a lower-level manager, or vice versa (Fulmer & Ostroff, 2017). Trust transfer models, however, do not examine whether trust × trust interactions may occur. The PI model suggests this consideration is important for fully understanding situations with multiple trustees.

Social Network Models

Other models of trust involving multiple parties (e.g., Burt & Knez, 1995; Ferrin et al., 2006; Pavlou & Gefen, 2004) often focus on how social structure surrounding a trustor and multiple trustees provides opportunities or constraints affecting the trustor’s behaviors. Social network theory provides a solid framework for understanding how amounts of interpersonal trust are affected by surrounding social structure. Social network scholars have long investigated the relationship between interpersonal ties, networks, and interpersonal trust. Some studies (Ermisch et al., 2023) primarily focused on the characteristics of a dyad, including presence (or absence) of tie, strength of ties (e.g. strong vs. weak ties), reciprocal (e.g. symmetric or asymmetric) ties, etc. Other studies (Burt & Knez, 1995; Coleman, 1988; Ferrin et al., 2006) extended a dyadic relationship to triadic relationships (e.g. mutual third party) and even whole networks (e.g. dense or cohesive networks and sparse networks).

For example, as one type of triad, some studies examined the role of a third party (who is often viewed as a trustee) in brokering, building, or otherwise encouraging trust between the trustor and another trustee. Trust brokering occurs when a third party serves as a connection for two parties who initially do not trust each other (Kilduff et al., 2017). In some cases, third-party trust effects resemble trust transfer models. For example, trustee A may be trusted to effectively (i.e., in a trustworthy manner) constrain untrustworthy behavior or encourage trustworthy behavior by trustee B (Pavlou & Gefen, 2004), thereby resulting in trust in A predicting trust in B. However, in contrast to the trust transfer model, some third-party models hypothesize indirect effects, and the third parties need not necessarily be “trustees.” For example, third parties may be onlookers whose presence is a threat to the focal trustee’s reputation if that trustee behaves in an untrustworthy manner (Burt & Knez, 1995). In such cases, the onlookers are not trustees and it is their presence that is important, not their trustworthiness. Alternatively, third parties’ common ties with the trustor and trustee may encourage good organizational citizenship behaviors and trustworthiness by the trustee (Ferrin et al., 2006).

In contrast to social network models, the PI model primarily focuses on how the amount of one trustor’s trust in one trustee (or a trustor’s perception of one trustee) affects the effects of trustor’s trust in another trustee. Meanwhile, social network theories tend to focus on how ‘relations’ (e.g. dyad, triad, and/or whole networks) provide opportunities or constraints affecting an actor’s amounts of trust in others.

Studies of Additive Trust Effects

Trust transfer and social network models provide valuable insights into how the amount of trust in a referent might be affected or fostered by a different entity. In these models, trustee A’s impact on an ultimate outcome of interest (e.g., behavior, performance, turnover) is generally indirect and occurs through its impact on the trustor’s trust in trustee B. In contrast, some studies examine whether a trustor’s trust in multiple trustees has simultaneous independent, additive, direct effects on various outcomes (e.g., Chen et al., 2009; Lu et al., 2010).

In the additive case, increases in trust in each party are hypothesized to independently contribute to some outcome. For example, Lu et al. (2010) examined how members of a virtual community on a web-based platform developed trust in both other community members and the platform. In addition to finding that trust in other community members facilitated trust in the platform (consistent with trust transfer models), they found trust in each trustee (i.e., other community members and the platform) independently and positively contributed to purchase intentions on the platform, a trust-relevant behavior. However, trust in the platform had a stronger effect than trust in the other community members. Lu et al. (2010) also examined the trust-relevant outcome of “intention to get information” from the platform but found only trust in the platform (not trust in community members) independently predicted that outcome.

Note the PI model proposes that this differential weighting stems from perceptions of the separate influence each trustee has over the trustor’s desired end. Thus, it may be that purchasers perceived the platform (trustee A) as exerting greater control over successful and safe purchases, as well as quality of information, than the community members (trustee B). On the other hand, because Lu et al. (2010) did not check to see if there was a significant trust × trust interaction between trust in the platform and trust in the community members, the researchers may have underestimated the effects of trust in these trustees on their trusting outcomes. For example, it is possible that trust in the platform compensated for (and reduced the effect of) low trust in community members, and vice versa, hiding the joint effects within an interaction term.

When multiple trustees are considered as having influence over a single outcome, additive effects are the most common effects investigated. Other researchers have found independent additive effects of employee trust in upper management and immediate supervisors or co-workers predicting job satisfaction and organizational commitment (Cho & Park, 2011) or perceived organizational performance (Javed et al., 2018). Additive effects of trust in multiple trustees also predict outcomes such as support for social welfare policies (Daniele & Geys, 2015) and willingness to share health care information (Esmaeilzadeh, 2019). Additive models can compare the independent and direct effects of trust in different trustees to identify which trustees are most important to trustors under different conditions. However, they do not tell the full story of the joint effects of multiple trustees if hidden trust × trust interaction effects exist.

Studies of Trust × Trust Interactions

Current studies of trust involving more than one trustee rarely test for or find interaction effects. However, a few do. Cho et al. (2014) examined viral electronic advertising and how trust in distributors (senders) of an email message and trust in producers of the message (advertisers) influenced message recipient behaviors, perceptions, and attitudes. The authors found several negative interaction effects such that trust in the sender reduced the effect of trust in the advertiser or vice versa. For example, when predicting (1) whether the recipient opened the message and (2) positive attitudes after opening the message, trust in the advertiser was only important if trust in the sender was low. The authors concluded that trust in a sender could compensate for negative effects of a less trusted advertiser.

A different study by Agahari et al. (2022) examined how trust in “multi-party computation” processes reduced the need for organizations to trust one another when sharing data. They found that trust in the computation processes can compensate for low/uncertain trust between organizations to facilitate data sharing. Another rare example of a study of trust × trust interaction effects is a study by Brock et al. (2011) ² examining how trust in an e-retail platform (Facebook) and trust in the seller (e-retailer) affected trust in shopping and purchase intentions. They found a positive interaction between perceptions of the e-retailer and platform trustworthiness: If participants perceived either the platform or retailer as not trustworthy, then the trustworthiness of the other trustee mattered less for predicting trust in shopping. The authors also found a positive interaction between perceptions of e-retailer competence and the platform’s benevolence/integrity for predicting purchase intentions. Thus, trust in both trustees was necessary (compulsory) to support purchase intentions.

While these studies tested for and found trust × trust interaction effects, they do not explain why or predict when these interactions occur. The PI model suggests answers are found in the trustor’s perceptions of the task requirements that would lead to the trustor’s desired end, and their perceptions of the trustee’s joint influence over those requirements. In the Brock et al. (2011) study, the PI model would predict that trustors/buyers are especially focused on some tasks requiring both seller competence (operationalized as perceptions of ability to do internet business) and platform benevolence/integrity (operationalized as protecting buyer’s interests) to achieve a successful purchase.

Model Contributions and Strengths

We began this article by posing the question: When a trustor’s risk in a situation is affected by multiple distinct trustees, what determines how the trustor formulates their aggregate trust in and willingness to rely on multiple trustees? In the conceptual PI model presented, we address an important, but heretofore neglected, aspect of trust research: the differing ways trust in multiple distinct trustees may affect one another and combine to create multi-trustee trust. While the MDS model and other models provide insight into dyadic trusting relationships, an expansion of the model into an examination of multi-trustee trusting relationships is needed to reflect the types of situations that a trustor may encounter. We introduce perceived trustee influence, perceived task requirements, and multi-trustee trust as a foundation for discussing the varied (additive, compensatory, and compulsory) effects of trust in multiple trustees.

This conceptual paper makes several contributions to the literature on trust. First, it identifies a significant gap pertaining to multiple-trustee trust effects. Most real-life situations involving “trust-relevant outcomes” such as trusting behaviors also involve perceived vulnerabilities to multiple distinct trustees. However, there exists very little research and theory to guide thinking about how varied amounts of trust in multiple trustees might combine to predict such outcomes. This comprises a large gap. Technological advancements and role specialization has increased the number of situations in which a trustor must rely on multiple distinct trustees. Oncology as a field came about due to rapid advances in cancer research. This is to a patient’s benefit, but the fact remains that having two clinicians (primary care physician and oncologist) to trust rather than just one is more complex and needs to be better understood.

Second, it begins to address this gap in research on multi-trustee trust effects by identifying three ways in which trust in two trustees might combine to create emergent multi-trustee trust. Specifically, we observe the possibility that trust in one trustee may at times moderate the effects of trust in another, and two continuous variables (e.g., high to low trust in two trustees) can combine additively or interactively to affect outcomes. Further, when contingency (interaction) effects occur, they can be positive or negative. As “interaction effects” exist on a continuum ranging from very negative, to zero, to very positive, we offer the descriptive concepts of additive, compulsory, and compensatory effects to link statistical concepts with continuously varying but qualitatively distinctive patterns of effects observable in human contexts.

Third, the paper offers a conceptual extension of definitions of trust as “willingness to be vulnerable” by focusing greater attention on perceptions of trustee influence. We argue the reason trustors consider their vulnerability to trustees is because of trustee influence over trustor risks in specific situations and that trustee influence can vary in amount and type. Thus, the PI model proposes the concept of perceived trustee influence as a potential testable mechanism explaining the extent to which trust in one trustee may have a stronger or weaker effect upon a trusting outcome, and when and why trust in two trustees might combine additively versus interactively.

Fourth, the PI model introduces the ideas of trustor perceptions of task requirements as tightly related to perceptions of trustee influence, because trustees likely will have varied influence over the tasks or conditions supporting the trustor’s desired end. Fifth, while the concepts of perceived task requirements and trustee influence are not yet fully articulated, we identify task interdependence as one task requirement trustors may perceive which will drive their perceptions of what we have termed the cooperative influence of trustees. We further offer an operational definition of perceived cooperative influence which mirrors recent suggestions made for detecting task interdependence (i.e., change in benefits of a task depending on whether the other task is executed). In so doing, we have offered ideas for initial tests of the proposed conceptual model.

A major strength of the PI model is that it offers a set of propositions which can form the basis for testable hypotheses. A second strength is that we expect that it will apply to trust “in” trustees of multiple types. As previously noted, the MDS model has been successfully applied across various disciplines, contexts, and levels (Schoorman et al., 2007). Thus, we envision that the individual distinct trustees to which the PI model could refer may be any trustee high in entitativity (c.f., Schilke & Lumineau, in press), including persons (interpersonal), groups (e.g., teams or organizations), or other entities like technologies, institutions, and policies. This is important because rarely do risks and vulnerabilities arising from a specific trust-relevant behavior involve reliance upon only one trustee, or one type of trustee. In the workplace, outcomes such as job satisfaction may be affected by vulnerability to one’s organization, one’s managers, and one’s co-workers, as well as the technology one needs to do one’s job (Li et al., 2008). Furthermore, it is common to employ different types of trustees to compensate for low trust in other trustees, such as when various technologies are employed to monitor and assist humans (e.g., cars alert drivers to dangers in their blind spots). Thus, our contribution extends other research examining trust in multiple trustees whereby one is a person and the other is a type of technology such as a robot (Birmingham et al., 2020).

Limitations and Future Directions

Although we expect the PI model to apply to trust “in” different types of trustees, a limitation of the PI model is that its focus on individual-level psychological mechanisms may restrict its applicability to situations when the trust is “at” a different level such as the group or the organization (Fulmer & Gelfand, 2012). The PI model’s focus on trustors’ perceptions and cognitions also suggests the most pertinent outcomes will be those proximal to the trustor (trustor attitudes, decisions, and behaviors or behavioral intentions). It is unknown whether the model would apply to longer-term or multilevel outcomes beyond the individual, such as organizational turnover.

A second limitation is the need for further development of the concepts of task requirements and joint influence. Future research is needed to identify what trustors attend to within the task situation and the task requirements. This represents a significant area of potential future research, especially when more than one trustee is involved. Type and amount of influence also have not been extensively considered in trust theory and research. Most research assumes the trustor is giving up some control, thereby allowing one or more trustees to have some influence, without detailing the control/influence characteristics (see Agahari et al., 2022, for an exception). Beyond concepts such as collaboration, coordination, and competition, prior work has conceptualized variations in control in terms of formal and informal mechanisms focused on inputs, outputs, and processes (e.g., Wiener et al., 2016). Greater theoretical and empirical attention to the types of influence and control is needed to fruitfully advance understanding of multi-trustee trust effects.

A third limitation is that the PI model, as described in this paper, focuses on only two trustees. Trustors often must deal with three or more trustees, leading to increased complexity and challenges in predicting the emergence of configurations of trust and their effects on trustor behavior and outcomes. There could be other ways in which multi-trustee trust emerges, not yet identified in this paper. This is especially likely as the number of trustees increases. Suppose a trustor is dealing with three trustees, A, B, and C, while making decisions. In such a scenario, the trustor may perceive trustees A and B as having compensatory effects on the decision, trustees B and C having additive effects, and trustees A and C having compulsory effects. An investigation into the complexity of competing trust configurations and how they might impact trust-relevant outcomes would provide additional insight into multi-trustee trust as well as advance the conceptual model presented here. Thus, as the number of trustees grows and their patterns of influence over a trustor’s vulnerability become increasingly complex, we propose that it will be increasingly important to attend to, theorize, and test for emergent properties and processes. Different forms of emergence (Kozlowski & Klein, 2000) have increasingly been considered in the context of considering trust across levels, individuals, and groups (Costa et al., 2018; De Jong & Dirks, 2012). Advancing understanding of such configurations is important and reflective of real-world phenomena.

The PI model also does not currently elaborate on the connections between various trustworthiness constructs and perceived influence. Prior research implicitly supports the importance of trustee influence by focusing on those trustworthiness features most likely to influence risk in the situation (Kappmeier et al., 2021; Mayer et al., 1995; McAllister, 1995). This suggests different types of influence may matter for different types of threats (Gillespie, 2003; Tomlinson et al., 2020). However, additional research is needed to advance understanding of how best to link trustee trustworthiness and influence.

Because the PI model primarily focuses on how a trustor’s distinct amounts of trust in two trustees might combine to influence their behaviors, the model is limited to understanding how trusting (or distrusting) relationships between multiple trustees shape a trustor’s trusting behaviors. In this regard, integrating relevant social network research will offer valuable insight into incorporating a trustor’s perception of the relations between multiple trustees (Krackhardt & Hanson, 1993). This implies that the perceived presence (or absence), strength, and/or direction (e.g., symmetric or asymmetric) of trusting relationships between trustees is likely to influence a trustor’s willingness to trust multiple trustees. For example, trusting relationships are often bidirectional or asymmetric (Stewart, 2003), which means that trustee A may trust in trustee B while trustee B does not necessarily trust in trustee A. Future theory and research might expand our model by incorporating the strength of trusting relationships between trustees and the direction of such relationships.

Finally, the PI model does not consider the impact of individual differences on a trustor’s assessment of perceived task situation requirements and trustee influence. Results from a meta-analysis suggest that risk propensity is independent of the Five Factor Model (FFM) personality traits (Highhouse et al., 2022), suggesting the relationship between personality traits and perceived trustee influence could be reliably examined. Also, research has shown that intra-individual perceptions of risk may vary such that people may be willing to accept more risk in domains such as work or sports, but less risk in domains such as childcare or retirement savings (Blais & Weber, 2006; Weber et al., 2002). An examination of variability in the trustees’ FFM personality traits or in other factors such as age and gender (e.g., Figner & Weber, 2011; Kanthak & Woon, 2015) may shed light on whether additive, compensatory, or compulsory trust effects are more likely to occur in a particular situation or be associated with certain individual differences.

Conclusion

Returning to our initial request: Envision yourself again as the business owner, deciding whether you trust your employees enough to take the new contract. Initially, you assess your amount of trust in each employee based on their specific trustworthiness characteristics. However, this alone is not sufficient for your decision. Guided by the new PI model, you recognize the importance of considering the contract task requirements and if and how your employees can influence task success. The contract proposal lists several requirements: (1) the use of a certain unavailable material—which is outside your employees’ control; (2) the design of a new widget, which matches employee B’s skills and enthusiasm; and (3) a production plan which would require reliable effort from both employee A and B.

Having analyzed the task requirements and employee influence, it is clear that trust in your employees is irrelevant if the unavailable material is necessary. Thus, you suggest to the client a different (more available) material to address requirement #1. Meanwhile, trust in employee B for their ingenuity strongly influences your confidence you can meet disjunctive requirement #2, regardless of your trust in A. Finally, B’s unreliability makes you hesitate about requirement #3. However, you think you might hire a third employee, C, to help A and compensate for B, or shift the production plan from a conjunctive to an additive task to allow reliable employee A more independent influence.

In sum, the PI model provides greater clarity and detail regarding how and when trust influences outcomes such as multi-trustee trust and trusting decisions, as well as suggesting new solutions to problems caused by low trust. Researchers exploring multi-trustee effects can incorporate PI model concepts in their analyses and empirical studies to contribute to further refinement of the model.

Footnotes

Acknowledgments

The authors are grateful to Claire Jumper and Aiden Quinn of the University of Nebraska Public Policy Center (NUPPC) for their assistance with literature review activities, to Taylor St. Cin of NUPPC for assistance with preparation of the figures, and to the editors of the special issue and anonymous reviewers for their assistance in developing and clarifying our ideas.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a collaboration grant from the University of Nebraska (NU Collaboration Initiative, 2021-2022, #32182, awarded to Lisa PytlikZillig, Jooho Lee, Ashley Votruba, and Changsoo Song, and including Michelle Fleig-Palmer and Mariska Kappmeier as external partners).

ORCID iDs

Lisa M. PytlikZillig

Ashley M. Votruba

Michelle M. Fleig-Palmer

Jooho Lee

Mariska Kappmeier

Notes

Author Biographies

Lisa M. PytlikZillig, Ph.D., is an Associate Research Professor at the University of Nebraska Public Policy Center, has a Courtesy Appointment at the University of Nebraska-Lincoln Department of Psychology, and serves as one of two Faculty Ombuds. Dr. PytlikZillig's research focuses on public engagement, trust in institutions, and conflict and conflict resolution.

Ashley M. Votruba, J.D., Ph.D., is an Assistant Professor in the Law-Psychology Program and Social Psychology Program at the University of Nebraska – Lincoln and has a Courtesy Appointment at the Nebraska College of Law. Dr. Votruba is the director of the Culture, Conflict, and Law Lab. She maintains an active research program examining how social perception and dispute system design influence conflict management. Dr. Votruba’s areas of expertise include cultural psychology, conflict theory, restorative justice, civil justice, and alternative dispute resolution.

Michelle M. Fleig-Palmer, Ph.D., is an Associate Professor at South Dakota State University. Her research in trust, mentoring, electronic health records, and crisis leadership has won awards from the Academy of Management Health Care Division (best theory to practice paper), Midwest Academy of Management (best paper), and Medical Care Research and Review (2022 article of the year). Dr. Fleig-Palmer’s research interests include building trust in healthcare settings and emergent leadership in diverse populations such as MENA women.

Jooho Lee, Ph.D., is a professor at School of Public Administration and an associate director of Digital Governance and Analytics Lab at University of Nebraska, Omaha. Areas of his research interest include public management with an emphasis on information and communication technology use in public organizations and the application of social network theories in public management issues such as interagency and interorganziational collaboration.

Mariska Kappmeier, Ph.D., is a researcher in the Psychology Department at the University of Otago, New Zealand. She received her Ph.D. in Psychology from the University of Hamburg, Germany and completed her post-doctoral tenure in the Psychology Department of Harvard University, USA. Her research focuses on intergroup trust and trust building, collective memories, and trust between institutions and diverse communities. She developed the Intergroup Trust Model, which conceptualizes intergroup trust as a multidimensional construct and provides a solid basis for understanding how trust between groups can be built or eroded.

References

Agahari

Ofe

de Reuver

(2022). It is not (only) about privacy: How multi-party computation redefines control, trust, and risk in data sharing. Electronic Markets, 32(3), 1577–1602. https://doi.org/10.1007/s12525-022-00572-w

Ajzen

(1991). The theory of planned behavior. Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-t

Audrey Korsgaard

Kautz

Bliese

Samson

Kostyszyn

(2018). Conceptualising time as a level of analysis: New directions in the analysis of trust dynamics. Journal of Trust Research, 8(2), 142–165. https://doi.org/10.1080/21515581.2018.1516557

Bandura

(1997). Self-efficacy: The exercise of control. W. H. Freeman and Company.

Beal

D. J.

Cohen

R. R.

Burke

M. J.

McLendon

C. L.

(2003). Cohesion and performance in groups: A meta-analytic clarification of construct relations. Journal of Applied Psychology, 88(6), 989–1004. https://doi.org/10.1037/0021-9010.88.6.989

Belanche

Casaló

L. V.

Flavián

Schepers

(2014). Trust transfer in the continued usage of public e-services. Information & Management, 51(6), 627–640. https://doi.org/10.1016/j.im.2014.05.016

Birmingham

Mahajan

Reber

Matarić

M. J.

(2020). Can I trust you? A user study of robot mediation of a support group. In 2020 IEEE International Conference on Robotics and Automation (ICRA), Paris, France, 2020, 8019–8026.

Blais

A.-R.

Weber

E. U.

(2006). A domain-specific risk-taking (DOSPERT) scale for adult populations. Judgment and Decision making, 1(1), 33–47. https://doi.org/10.1017/s1930297500000334

Brock

Blut

Linzmajer

Zimmer

(2011). F-commerce and the crucial role of trust. Proceedings of the thirty second International Conference on Information Systems.

10.

Burt

R. S.

Knez

(1995). Kinds of third-party effects on trust. Rationality and Society, 7(3), 255–292. https://doi.org/10.1177/1043463195007003003

11.

Chen

Lai

Lin

(2014). A trust model for online peer-to-peer lending: A lender’s perspective. Information Technology and Management, 15(4), 239–254. https://doi.org/10.1007/s10799-014-0187-z

12.

Chen

Zhang

(2009). The role of mutual trust in building members' loyalty to a C2C platform provider. International Journal of Electronic Commerce, 14(1), 147–171. https://doi.org/10.2753/jec1086-4415140105

13.

Cho

Huh

Faber

R. J.

(2014). The influence of sender trust and advertiser trust on multistage effects of viral advertising. Journal of Advertising, 43(1), 100–114. https://doi.org/10.1080/00913367.2013.811707

14.

Cho

Y. J.

Park

(2011). Exploring the relationships among trust, employee satisfaction, and organizational commitment. Public Management Review, 13(4), 551–573. https://doi.org/10.1080/14719037.2010.525033

15.

Coleman

J. S.

(1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95–S120. https://doi.org/10.1086/228943

16.

Costa

A. C.

Fulmer

C. A.

Anderson

N. R.

(2018). Trust in work teams: An integrative review, multilevel model, and future directions. Journal of Organizational Behavior, 39(2), 169–184. https://doi.org/10.1002/job.2213

17.

Courtright

S. H.

Thurgood

G. R.

Stewart

G. L.

Pierotti

A. J.

(2015). Structural interdependence in teams: An integrative framework and meta-analysis. Journal of Applied Psychology, 100(6), 1825–1846. https://doi.org/10.1037/apl0000027

18.

Currall

S. C.

Inkpen

A. C.

(2002). A multilevel approach to trust in joint ventures. Journal of International Business Studies, 33(3), 479–495. https://doi.org/10.1057/palgrave.jibs.8491027

19.

Daniele

Geys

(2015). Interpersonal trust and welfare state support. European Journal of Political Economy, 39, 1–12. https://doi.org/10.1016/j.ejpoleco.2015.03.005

20.

De Jong

B. A.

Dirks

K. T.

(2012). Beyond shared perceptions of trust and monitoring in teams: Implications of asymmetry and dissensus. Journal of Applied Psychology, 97(2), 391–406. https://doi.org/10.1037/a0026483

21.

De Jong

B. A.

Dirks

K. T.

Gillespie

(2016). Trust and team performance: A meta-analysis of main effects, moderators, and covariates. Journal of Applied Psychology, 101(8), 1134–1150. https://doi.org/10.1037/apl0000110

22.

Dirks

K. T.

de Jong

(2022). Trust within the workplace: A review of two waves of research and a glimpse of the third. Annual Review of Organizational Psychology and Organizational Behavior, 9(1), 247–276. https://doi.org/10.1146/annurev-orgpsych-012420-083025

23.

Ermisch

Gambetta

Iacono

S. L.

Sonmez

(2023). Trust and strength of Family ties: New Experimental evidence. Social Psychology Quarterly, Advanced Online Publication. https://doi.org/10.1177/01902725231162074

24.

Esmaeilzadeh

(2019). The impacts of the perceived transparency of privacy policies and trust in providers for building trust in health information exchange: Empirical study. JMIR Medical Informatics, 7(4), e14050. https://doi.org/10.2196/14050

25.

Ferrin

D. L.

Dirks

K. T.

Shah

P. P.

(2006). Direct and indirect effects of third-party relationships on interpersonal trust. Journal of Applied Psychology, 91(4), 870–883. https://doi.org/10.1037/0021-9010.91.4.870

26.

Figner

Weber

E. U.

(2011). Who takes risks when and why? Determinants of risk taking. Current Directions in Psychological Science, 20(4), 211–216. https://doi.org/10.1177/0963721411415790

27.

Fulmer

C. A.

Gelfand

M. J.

(2012). At what level (and in whom) we trust trust across multiple organizational levels. Journal of Management, 38(4), 1167–1230. https://doi.org/10.1177/0149206312439327

28.

Fulmer

C. A.

Ostroff

(2017). Trust in direct leaders and top leaders: A trickle-up model. Journal of Applied Psychology, 102(4), 648–657. https://doi.org/10.1037/apl0000189

29.

Gillespie

N. A.

(2003). Measuring trust in working relationships: The behavioral trust inventory. In Annual Meeting of the Academy of Management. Seattle, WA, 2003.

30.

Hansen

(2009). Collaboration: How leaders avoid the traps, build common ground, and reap big results. Harvard Business Press.

31.

Highhouse

Wang

Zhang

D. C.

(2022). Is risk propensity unique from the big five factors of personality? A meta-analytic investigation. Journal of Research in Personality, 98, 104206. https://doi.org/10.1016/j.jrp.2022.104206

32.

Javed

S. A.

Syed

A. M.

Javed

(2018). Perceived organizational performance and trust in project manager and top management in project-based organizations: Comparative analysis using statistical and grey systems methods. Grey Systems: Theory and Application, 8(3), 230–245. https://doi.org/10.1108/gs-01-2018-0009

33.

Kanthak

Woon

(2015). Women don't run? Election aversion and candidate entry. American Journal of Political Science, 59(3), 595–612. https://doi.org/10.1111/ajps.12158

34.

Kappmeier

Guenoun

Fahey

K. H.

(2021). Conceptualizing trust between groups: An empirical validation of the five-dimensional intergroup trust model. Peace and Conflict: Journal of Peace Psychology, 27(1), 90–95. https://doi.org/10.1037/pac0000537

35.

Kilduff

Mehra

Gioia

D. A.

Borgatti

(2017). Brokering trust to enhance leadership: A self-monitoring approach to leadership emergence. Knowledge and networks, 11, 221–240. https://doi.org/10.1007/978-3-319-45023-0_11

36.

Kozlowski

S. W. J.

Chao

G. T.

(2012). The dynamics of emergence: Cognition and cohesion in work teams. Managerial and Decision Economics, 33(5-6), 335–354. https://doi.org/10.1002/mde.2552

37.

Kozlowski

S. W. J.

Klein

K. J.

(2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In Klein

K. J.

Kozlowski

S. W. J.

(Eds.), Multilevel theory, research and methods in organizations: Foundations, extensions, and new directions (pp. 3–90). Jossey-Bass.

38.

Krackhardt

Hanson

J. R.

(1993). Informal networks: The company behind the charts [Article]. Harvard Business Review, 71(4), 104–111.

39.

Hess

T. J.

Valacich

J. S.

(2008). Why do we trust new technology? A study of initial trust formation with organizational information systems. The Journal of Strategic Information Systems, 17(1), 39–71. https://doi.org/10.1016/j.jsis.2008.01.001

40.

Zhao

Wang

(2010). From virtual community members to C2C e-commerce buyers: Trust in virtual communities and its effect on consumers’ purchase intention. Electronic Commerce Research and Applications, 9(4), 346–360. https://doi.org/10.1016/j.elerap.2009.07.003

41.

Mayer

R. C.

Davis

J. H.

(1999). The effect of the performance appraisal system on trust for management: A field quasi-experiment. Journal of Applied Psychology, 84(1), 123–136. https://doi.org/10.1037//0021-9010.84.1.123

42.

Mayer

R. C.

Davis

J. H.

Schoorman

F. D.

(1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709–734. https://doi.org/10.5465/amr.1995.9508080335

43.

McAllister

D. J.

(1995). Affect-and cognition-based trust as foundations for interpersonal cooperation in organizations. Academy of Management Journal, 38(1), 24–59. https://doi.org/10.5465/256727

44.

Pavlou

P. A.

Gefen

(2004). Building effective online marketplaces with institution-based trust. Information Systems Research, 15(1), 37–59. https://doi.org/10.1287/isre.1040.0015

45.

Puranam

(2018). The microstructure of organizations. Oxford University Press.

46.

PytlikZillig

L. M.

Kimbrough

C. D.

(2016). Consensus on conceptualizations and definitions of trust: Are we there yet? In Shockley

Neal

T. M.

PytlikZillig

L. M.

Bornstein

B. H.

(Eds.), Interdisciplinary perspectives on trust: Towards theoretical and methodological integration (pp. 17–47). Springer. https://doi.org/10.13140/RG.2.1.1365.5205

47.

Rizk

Fox

Rosu

(2023). Towards an Ontology of task dependence in organizations. In Proceedings of the 15th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management. Lisbon, Portugal, 2023, https://doi.org/10.5220/0012238300003598

48.

Rousseau

D. M.

Sitkin

S. B.

Burt

R. S.

Camerer

(1998). Not so different after all: A cross-discipline view of trust. Academy of Management Review, 2(3), 393–404. https://doi.org/10.5465/amr.1998.926617

49.

Schilke

Lumineau

How organizational is interorganizational trust?

Academy of Management Review. (in press). https://doi.org/10.5465/amr.2022.0040

50.

Schoorman

F. D.

Mayer

R. C.

Davis

J. H.

(2007). An integrative model of organizational trust: Past, present, and future. Academy of Management Review, 32(2), 344–354. https://doi.org/10.5465/amr.2007.24348410

51.

Smith

E. R.

Semin

G. R.

(2007). Situated social cognition. Current Directions in Psychological Science, 16(3), 132–135. https://doi.org/10.1111/j.1467-8721.2007.00490.x

52.

Spreitzer

G. M.

Mishra

A. K.

(1999). Giving up control without losing control: Trust and its substitutes’ effects on managers’ involving employees in decision making. Group & Organization Management, 24(2), 155–187. https://doi.org/10.1177/1059601199242003

53.

Steiner

I. D.

(1972). Group process and productivity. Academic Press.

54.

Stewart

K. J.

(2003). Trust transfer on the world wide web. Organization Science, 14(1), 5–17. https://doi.org/10.1287/orsc.14.1.5.12810

55.

Thielmann

Hilbig

B. E.

(2015). Trust: An integrative review from a person–situation perspective. Review of General Psychology, 19(3), 249–277. https://doi.org/10.1037/gpr0000046

56.

Tomlinson

E. C.

Schnackenberg

A. K.

Dawley

Ash

S. R.

(2020). Revisiting the trustworthiness–trust relationship: Exploring the differential predictors of cognition‐and affect-based trust. Journal of Organizational Behavior, 41(6), 535–550. https://doi.org/10.1002/job.2448

57.

Tory Higgins

(2000). Social cognition: Learning about what matters in the social world. European Journal of Social Psychology, 30(1), 3–39. https://doi.org/10.1002/(sici)1099-0992(200001/02)30:1<3::aid-ejsp987>3.0.co;2-i

58.

Venkatesh

Bala

(2012). Adoption and impacts of interorganizational business process standards: Role of partnering synergy. Information Systems Research, 23(4), 1131–1157. https://doi.org/10.1287/isre.1110.0404

59.

Weber

E. U.

Blais

A. R.

Betz

N. E.

(2002). A domain‐specific risk‐attitude scale: Measuring risk perceptions and risk behaviors. Journal of Behavioral Decision Making, 15(4), 263–290. https://doi.org/10.1002/bdm.414

60.

Wiener

Mähring

Remus

Saunders

(2016). Control configuration and control Enactment in information systems Projects. MIS Quarterly, 40(3), 741–774. https://doi.org/10.25300/misq/2016/40.3.11