When ethical failures occur in large corporations, they make the front page news. Executives who breach the public’s trust by misappropriating funds or misrepresenting company earnings can affect the lives of hundreds of thousands of individuals. While these organizational failures are highly visible and salient, they capture only a small proportion of the ethical failures that occur on a daily basis. Alongside these visible failures, smaller ethical failures punctuate our day-to-day organizational lives. Overstating contributions to team outcomes, misrepresenting individual performance, and filtering information are all examples of small-scale ethical failures that affect individual and organizational outcomes. In this research, we focus on a commonplace organizational activity—negotiation—to explore one such small-scale ethical failure, the decision to withhold or misrepresent information. This decision is consequential because at best it prevents negotiators from identifying mutually beneficial outcomes; at worst it violates trust, damaging individuals’ reputations and their long-term relationships (Aquino 1998).

How negotiators should manage information is highly ambiguous. Although negotiators may choose to openly disclose information about preferences and priorities, they may also be motivated to either withhold information (sin of omission) or to deliberately misrepresent their preferences and priorities (sin of commission; Spranca et al. 1991). While openness enables negotiators to craft mutually beneficial solutions, it also exposes them to the risk of exploitation by the other party. Moreover, withholding or misrepresenting information enables negotiators to build their own power and boost their personal outcomes (Murnighan et al. 1999; Steinel and De Dreu 2004). As a result, in planning their information strategy, negotiators are pulled in two directions, raising the question of when negotiators choose to deceive their opponents.

The ambiguity that surrounds information exchange in negotiation implies that individuals’ thresholds for deception are not absolute, but rather are influenced by the negotiation context (Ruedy and Schweitzer 2010). Extending findings in the behavioral ethics domain, which establish the impact of context on ethical judgments (Trevino et al. 2006), negotiation researchers have shown that negotiators’ propensity to deceive increases with increasingly competitive goals, and when power is unequally distributed between negotiators (Aquino 1998; Olekalns and Smith 2007, 2009; Stawiski et al. 2009; Steinel et al. 2010; Steinel and De Dreu 2004). Although both literatures demonstrate that the external context can alter individuals’ ethical thresholds, a key difference between the ethical decisions studied to date and those made in negotiations is that negotiators make their decisions in an interdependent (rather than an independent) context.

An important implication of this shift from an independent to an interdependent decision-making context is that the social context within which decisions are made comes to the forefront in ethical judgments. By social context, we mean the interpersonal context within which negotiations take place, that is, the expectations that negotiators have about their opponents. Because they are interdependent, negotiators need to resolve uncertainty about how their opponents will behave, especially the likelihood that their opponents will act exploitatively. This behavioral prediction is important because it determines whether negotiators will pursue a dominantly self-protective strategy or a more open and collaborative strategy. Interdependence thus has implications for where negotiators set their ethical thresholds: negotiators who anticipate exploitation will be more willing to employ unethical tactics such as deception those who do not anticipate exploitation.

Multiple sources of information help negotiators to resolve this uncertainty about their opponents’ intentions because they establish expectations about how an opponent will behave. This information is especially critical in the opening moments of a negotiation, when negotiators draw on limited information and intuition to form impressions of their opponents (Ambady and Rosenthal 1992; Curhan and Pentland 2007) and, on this basis, to select their strategy. We investigate how one component of strategy, negotiators’ use of deception, is influenced by impressions derived from the social context within which negotiations take place. First, we examine how the behavioral norms conveyed by social roles and opponents’ actions relative to those norms affect negotiators’ ethical thresholds. Our focus is on how the norms associated with gender-based social roles affect negotiators’ use of deception. Second, we examine how the expectations created by negotiators’ first impressions of their opponents affect negotiators’ ethical thresholds. Because of the central role that trust plays in shaping negotiators’ information strategies (Butler 1995, 1999), we focus on initial perceptions of an opponent’s trustworthiness. Taken together, the expectations established by social roles and first impressions help negotiators to assess their opponent’s intentions. This assessment is important because it enables negotiators to determine the extent to which they are vulnerable to exploitation and, consequently, the extent to which they need to protect themselves by deceiving their opponents (Dees and Cramton 1991).

To better understand how social roles and impressions of trustworthiness affect negotiators’ propensity to deceive, we compare levels of deception in response to the confirmation or violation of behavioral norms, and we do so in the context of dyad gender composition. We then consider how perceived trustworthiness further impacts the use of deception. By investigating how dyad gender composition, initial impressions of trustworthiness, and negotiation strategy interact to prime deception, we answer calls for better understanding how multiple variables work together to shape unethical actions (Kish-Gephart et al. 2010; Trevino et al. 2006).

Social Roles and Deception

Research has established that men are more likely than women to endorse a range of unethical negotiation strategies, and that they set lower ethical standards for negotiations (Cohen 2009; Kray and Haselhuhn 2012; Ma and McLean Parks 2012; Schweitzer and Gibson 2008). We know less about the role that an opponent’s gender plays in eliciting deception (Volkema and Rivers 2012), although recent research has shown that women elicit more deception than men (Kray et al. 2012). While gender may play a role in eliciting deception, we propose that whether women and men elicit deception is also influenced by the negotiation strategy that they implement. Specifically, because negotiators draw on the behavioral expectations conveyed by gender-based social roles to resolve uncertainty about how their opponents will (and should) behave (e.g., Miles 2010; Stuhlmacher and Linnabery (in press); Wood and Eagly 2012), we investigate whether deception is more likely to be elicited when negotiators employ expectancy disconfirming strategies. Our emphasis on social context also raises the possibility that women and men may respond differently to expectancy violations. Consequently, we also investigate whether negotiators’ use of deception is influenced by whether negotiations take place in same- or mixed-sex dyads.

We have proposed that the behavioral expectations associated with gender-based social roles shape negotiators’ willingness to deceive their opponents. These gender-based stereotypes convey the expectation that women will be warm, communal, and other-oriented, whereas men will be individualistic, ambitious, and self-oriented (Eagly and Karau 2002; Gelfand et al. 2006; Rudman and Phelan 2008). In the context of negotiation, these expectations imply that women will adopt a softer, more accommodating style that preserves the ongoing relationship, whereas men will adopt a harder, more competitive style that promotes self-interest (e.g., Kulik and Olekalns 2012; Pruitt 1981).

Negotiators are most likely to contemplate deception when they encounter expectancy violations, because such violations are weighted more heavily in individuals’ ethical judgments (O’Fallon and Butterfield 2005). According to Expectancy Violation Theory (EVT; Burgoon et al. 1995), and related research, behaviors that violate expectations are subject to greater scrutiny than those that meet expectations (Burgoon et al. 1995; Jett and George 2003; Weick 1995). When behaviors are discrepant with normative expectations such as those established by gender-based social roles, they trigger a reassessment of the other party’s intentions (Baxter and Erbert, 1999; Baxter et al. 1999). This reasoning implies that women who compete and men who accommodate will attract greater scrutiny because their actions violate gender-based behavioral expectations (Kulik and Olekalns 2012). Because such violations typically elicit negative attributions about an opponent’s intentions, negotiators are likely to increase their use of deception when opponents use gender-incongruent strategies.

Based on evidence that dyad gender composition affects behavior (Sutter et al. 2009), we propose that to fully understand the consequences of expectancy violations we need to consider whether negotiations occur in same- or mixed-sex dyads. When negotiations take place in same-sex dyads, gender-based stereotypes provide a highly salient set of behavioral expectations against which to judge negotiators’ actions. Gender-based expectations establish a socially shared meaning system based on salient behavioral norms, and reinforce negotiators’ implicit theories about appropriate strategies (Adair and Brett 2005, for a similar argument in relation to same-culture negotiations). Women and men who violate these norms will therefore invite deception. In the case of violations by women, this is because competition used in the context of cooperation is judged harshly (Hilty and Carnevale 1992). In the case of violations by men, this is because accommodation in the context of a competitive negotiation signals vulnerability. As a result, male negotiators who employ an accommodating strategy are likely to invite exploitation by the other negotiator because softness decreases the perceived opportunity costs of deception (e.g., Elangovan and Shapiro 1998).

When negotiations take place in mixed-gender dyads, the behavioral threshold for assessing expectancy violations is more ambiguous because both male and female gender stereotypes are salient. To move forward, negotiators need to converge on a shared set of expectations (Adair et al. 2006, for a similar argument in relation to cross-cultural negotiations). A plausible alternative set of expectations is provided by a different but salient social role, that of “negotiator” (Stuhlmacher and Linnabery in press). This role conveys the expectation that negotiators will be assertive, dominant, and rational (Kray and Thompson 2004). These expectations align with male-stereotyped behaviors, conveying the expectation that negotiators will employ competitive strategies and pursue self-interest. This negotiator role provides a credible and readily acceptable alternative set of standards that establishes competitive behaviors as normative. Consequently, as was the case in all-male dyads, the use of accommodation both violates expectations and signals softness. We therefore expect that an accommodating strategy used in a mixed-sex dyad will elicit deception. Our reasoning leads to the following hypotheses:

H1a

In dyads with at least one male negotiator, deception will increase when negotiators use an accommodating strategy.

H1b

In all-female dyads, deception will increase when negotiators use a competing strategy.

Initial Impressions of Trustworthiness and Deception

The final social context variable that we consider is an opponent’s trustworthiness. Trust, defined as a confident, positive expectation about the actions of another person (Rousseau et al. 1998), plays a critical role in shaping negotiators’ information strategies. Perceived trustworthiness conveys information about the likelihood that opponents will (or will not) negotiate in good faith. Negotiators are likely to conclude that their opponents will act exploitatively if they perceive them to be untrustworthy. Under these circumstances negotiators might choose to act pragmatically, deceiving their opponents to protect themselves. Indeed, Dees and Cramton (1991) argue that it is unreasonable for negotiators to behave in a trusting way if the other party is not trustworthy. An implication of this principle of moral pragmatism is that negotiators should give accurate information to their opponents only if they believe that the other party is negotiating in good faith.

As a part of the thin-slicing process that we described earlier (Ambady and Rosenthal 1992; Curhan and Pentland 2007), negotiators are likely to assess their opponents’ trustworthiness in the opening moments of a negotiation: Meyerson et al. (1996) describe the process of swift trust, that is, trust that is formed rapidly on initially encountering another person. Both thin-slicing generally, and swift trust more specifically, imply that negotiators will draw on sources of information beyond their personal values when deciding whether to deceive the other party.

At the start of negotiations, individuals may thus draw on initial impressions of the other party’s intentions to assess the likelihood that they will be exploited (McKnight et al. 1998). These initial impressions of trustworthiness are critical to how negotiators determine whether or not to give accurate information to their opponents. They provide an important cue for assessing whether opponents will use information that they obtain about a negotiator’s preferences to engage in problem-solving or to gain an advantage in the negotiation. By establishing a baseline of expectations for how opponents will behave in a negotiation, initial impressions of trustworthiness determine negotiators’ willingness to share information with their opponents (Butler 1995, 1999). Consistent with this view, research shows that deception increases when negotiators report low trust in their opponents (Olekalns and Smith 2007, 2009).

The impact of low trustworthiness is likely to be amplified when negotiators violate behavioral norms. Past research shows that first impressions of trustworthiness act as an interpretive filter, influencing how an opponent’s subsequent actions are interpreted (Druckman and Olekalns in press; Druckman et al. 2009). Moreover, low trust in an opponent increases the salience of negative events (Olekalns and Smith 2005). Expectancy violations, which invite scrutiny, also call into question an opponent’s intentions. Concerns about whether an opponent will act in good faith will be further amplified if opponents who have violated behavioral expectations are also perceived to be untrustworthy. We expect that low trust in an opponent will amplify the negative impact of gender-based expectancy violations

H2a

In dyads with at least one male negotiator, deception will increase when negotiators who are perceived as untrustworthy use a counter-normative accommodating strategy

H2b

In all-female dyads, deception will increase when negotiators who are perceived as untrustworthy use a counter-normative competing strategy.

Method

Participants

One hundred and twenty undergraduate and masters-level students (60 females, 60 males) took part in a simulated employment contract negotiation. Participants had an average age of 21 (SD = 2.6 years) and an average of 1.2 years work experience (SD = 1.7). They were recruited through campus-wide advertisements and paid for their participation.

Design

We tested our hypotheses in a Dyad Gender (male–male [MM], male–female [MF], female–female [FF]) x Negotiating Strategy (competing, accommodating) design. Dyads were randomly allocated to either competing or accommodating negotiation strategy. MM, MF, and FF dyads were evenly distributed across the two Negotiating Strategy conditions.

Procedure

Participants negotiated a simulated employment contract face-to-face. Written instructions assigned participants to the role of either an employer or an employee, explained the task, and gave a payoff schedule that described the points awarded for each possible contract. At this point, participants completed the trust questionnaire described below. Once they had completed the questionnaire, instructions were repeated verbally and the negotiation started. Negotiators were required to reach agreement on nine issues: base salary, vacation, moving expenses, performance bonus, job location, start date, signing bonus, contract length, and job assignment. Each negotiator had one indifference issue, an issue that was worth no points, included in the payoff schedule. For employers, this was job assignment; for employees, this was contract length. Past research has shown that such issues trigger deception (Carnevale et al. 2001). The Negotiating Strategy manipulation, as described below, was embedded in the instructions that participants received.

A negotiating strategy manipulation, adapted from Weingart et al. (1996) and Allred (2000) description of negotiating styles was embedded in the instructions. Negotiating Strategy was the same for both individuals in a dyad. We conducted a pilot study to identify the three behaviors most strongly characteristic of a competitive and an accommodating approach. Based on the outcome of this pilot work, we told participants in the competing condition that strategies for effective negotiation included resisting attempts at being persuaded, appearing firm, and pressuring the other person. In the accommodating condition, we told participants that effective negotiators showed empathy, expressed positive expectations or optimism, and created openings for the other negotiator. In addition, participants in the competitive condition were told that their goal was to achieve the best possible outcome for themselves, while the goal for those in the accommodating condition was to strengthen their relationship with the other person. To reinforce the use of their assigned style, participants were asked to write down two things that they could say or do to insure that they either improved their individual outcomes (competing strategy) or strengthened the relationship (accommodating strategy).

Before starting the negotiation, we asked participants whether their goal was to (a) get the best possible deal for themselves, (b) build a strong positive relationship with the other negotiator, or (c) get the best possible deal for themselves and maintain a strong positive relationship with the other negotiator. Negotiators in the competitive condition most frequently identified “get the best possible deal for myself” as their goal whereas negotiators in the accommodating condition most frequently identified “get the best possible deal for themselves and to maintain a strong positive relationship” as their goal, χ 2 (2) = 13.4, p = 0.001.

Trust measures

After meeting their negotiating partners, but before starting the negotiation, participants completed a 27-item scale measuring trust. We developed our questionnaire using items from several trust scales (Lewicki et al. 1997; McAllister et al. 2006; Mayer and Davis 1999). Although we treated trust as a one-dimensional construct in our hypothesis development, past research has shown that greater insight into the relationship among trust, the negotiating context, and deception can be gained by representing trust as a multi-dimensional construct (e.g., Olekalns and Smith 2007, 2009). We therefore conducted a factor analysis to determine the structure that best fit our trust scale. Eighteen items loaded on one of 4 sub-scales, described below (the remaining 9 items either loaded on multiple factors or formed single-item factors).

We first calculated a Global Trustworthiness score by averaging the rating of opponents on all 18 items. We then calculated an average rating of trustworthiness on each of the four trust sub-scales that we identified through the factor analysis. Two of these sub-components were consistent with the deterrence-based and identity-based sub-components of trust described by Lewicki and colleagues (Lewicki et al. 1997; McAllister et al. 2006). Three items formed a deterrence-based trust sub-scale (e.g., “If this person doesn’t do what he/she says she is going to do, I can get even”), and two items formed an identity-based trust sub-scale (e.g., “This person’s interests are the same as mine”). A further two sub-components were consistent with the benevolence-based and integrity-based sub-components of trust described by Mayer and Davis (1999). Six items formed a benevolence-based trust sub-scale (e.g., “This person is concerned about my welfare”) and seven items formed an integrity-based trust sub-scale (e.g., “This person will try to be fair in his/her dealings with me”). Table 1 shows scale correlations and reliabilities.

Table 1 Scale inter-correlations and reliabilities
Full size table

Deception

After transcribing negotiations, we identified every occasion on which negotiators mentioned the two indifference issues. Drawing on past research, we distinguished between active and passive forms of deception. Deception can be active, as is the case when individuals misrepresent the situation by giving false information, or passive, as is the case when individuals conceal information. These two forms of deception are frequently referred to as sins of commission and sins of omission, respectively (O’Connor and Carnevale 1997; Schweitzer and Croson 1999; Spranca et al. 1991).

Two coders, blind to the study’s hypotheses, coded these utterances as either sins of omission or sins of commission. A reference to an indifference issue was coded as a sin of commission if negotiators claimed that their indifference issue had a high value. For example, in relation to the applicant’s indifference issue of contract length, the applicant says “we need to agree on how long I am going to stay [contract length] because that is very important to me …”. Applying O’Connor and Carnevale’s definition, we coded any use of a negotiator’s indifference issue in a trade-off as a sin of omission, because this action implicitly conveys that the issue has value to the negotiator. For example, the applicant says “well, I’d like to discuss my moving expenses as well, if you’re going to give me a 2 year contract, I’m going to need 80 % for moving expenses”. All other references to the indifference issue (e.g., “Can we talk about professional development days?”) were coded as other. Inter-rater reliability, as measured by Cohen’s κ, was 0.94. We note that not all negotiators use sins of omission or sins of commission. On average, 0.64 sins of commission and 0.95 sins of omission occurred during a negotiation.

Approach to Data Analysis

We used hierarchical linear modeling (HLM) to test our hypotheses. HLM enables us to examine individual behaviors while controlling for dyadic membership (Bryk and Raudenbush 1992; Kenny et al. 1998). Because models with random slopes and intercepts cannot be estimated for dyadic data, our model allowed for a random intercept but fixed the slopes. We modeled our dependent variables using a Poisson distribution because this better captures the distributional properties of low frequency events.

Before testing our hypotheses, we tested the null model for each of our dependent variables (sins of omission, sins of commission) by modeling the intercept with equations that had no independent variables in the model. The null model tests for the presence of significant within dyad interdependence in the use of deception. For both of our measures of deception, we found this to be the case: sins of omission, γ 00 = 1.06, t(61) = 7.13, p < .001 and, sins of commission, γ 00 = 0.72, t(61) = 5.88, p < .001.

We fit two models for global trust, one for each type of deception. In 2-level models, Level 1 predictor variables describe attributes of the individual and Level 2 predictor variables describe attributes of the dyad. Ratings of global pre-negotiation trust were entered into the equation as Level 1 predictor variables. We also entered Role and Individual Gender as Level 1 predictors, to control for the possibility that these variables affected the use of deception. Negotiating Strategy, Dyad Gender, and their interaction, were entered as Level 2 predictor variables. In setting up each model, we specified interactions between our Level 2 predictors and the Level 1 intercept, as well as between our Level 2 predictors and the four sub-components of trust. We then repeated these analyses, replacing global trust with ratings of the four sub-components of pre-negotiation trust.

Results

In our analyses, we first considered whether global trust predicted the use of either sins of omission or sins of commission. We also considered whether the trust subscales differentially predicted sins of omission and commission. We analyze both global and unique trust because identifying which sub-components of trust predict the use of deception gives greater insight into the social context conditions under which deception is elicited.

Sins of Omission

We first tested whether global trust predicts the use of sins of omission. Our analysis showed that neither Role, γ 10 = −0.27, t(114) = −1.23, ns, nor Individual Gender, γ 20 = −0.30, t(114) = −1.11, ns, predicted the use of sins of omission. We found that Dyad Gender predicted the use of sins of omission, γ 02 = −0.37, t(58) = −2.18, p < .05. Sins of omission were used least often in FF dyads (M = 0.58, SD = 1.28), moderately in MM dyads (M = 0.70, SD = 0.93) and most often in MF dyads (M = 0.88, SD = 2.05). Dyad Gender interacted with Negotiating Strategy (H1), γ 03 = 0.54, t(58) = −3.75, p = .001, to affect the use of sins of omission. Figure 1 shows that sins of omission occurred most frequently when an accommodating strategy was used in a MF dyad and least frequently when FF dyads used an accommodating strategy. With respect to sins of omission, we have partial support for H1. This hypothesis predicted that an accommodating strategy would increase deception in dyads with at least one male negotiator, whereas a competing strategy would increase deception in all-female dyads. As can be seen in Fig. 1, negotiators in male–female dyads use more deception when opponents accommodate than when they compete; and in all-female dyads negotiators use more deception when opponents compete than when they accommodate.

Fig. 1
figure 1

Average use of sins of omission as a function of dyad gender and negotiating strategy

Full size image

This two-way interaction was qualified by a 3-way interaction among Dyad Gender, Negotiating Strategy, and Global Trust, γ 33 = −0.16, t(114) = −3.31, p < .01. To interpret this and subsequent 3-way interactions, we used a median split to classify pre-negotiation trust as high (above median) or low (equal to or below median). As can be seen in Fig. 2, deception was unaffected by strategy when trust was high. However, when trust was low and negotiators accommodated, deception decreased in same-gender dyads but increased in mixed-gender dyads, providing partial support for the hypothesis that, in dyads with at least one male negotiator, low trust combined with an accommodating strategy would increase the use of deception (H2a).

Fig. 2
figure 2

Average use of sins of omission as a function of global trust, dyad gender, and negotiating strategy

Full size image

Our second analysis tested whether the interaction among Dyad Gender, Negotiating Strategy, and Global Trust could be isolated to a specific sub-component of trust. As well as replicating the effects described above for our independent variables, this analysis showed a significant 3-way interaction among Benevolence-based Trust, Dyad Gender, and Negotiating Strategy, γ 33 = −1.14, t(102) = −2.60, p < .001. As shown in Fig. 3, negotiators in FF dyads were most likely to use sins of omission when both parties competed and benevolence-based trust was low (H2). Conversely, in MF dyads, sins of omission occurred most frequently when benevolence-based trust was low and the other party accommodated (H2). Contrary to H2, in MM dyads, sins of omission were used least frequently when benevolence-based trust was low and negotiators accommodated. Our findings show that although low trust has a larger impact in sins of omission than high trust, the decision to deceive or not is shaped by the use of a competing strategy in FF dyads, whereas it is shaped by the use of an accommodating strategy in MM and MF dyads. This pattern replicates the pattern of deception we observed in relation to Global Trust, suggesting that the effects of Global Trust on the use of sins of omission can be isolated to the assessment of an opponent’s benevolence.

Fig. 3
figure 3

Average use of sins of omission as a function of benevolence-based trust, dyad gender, and negotiating strategy

Full size image

Sins of Commission

We first tested whether global trust predicts the use of sins of commission. Our analysis showed that neither Role, γ 10 = −0.30, t(114) = −1.08, ns, nor Individual Gender, γ 20 = 0.41, t(114) = 1.29, ns, predicted the use of sins of omission. We found no significant effects for our independent variables, Dyad Gender and Negotiating Strategy, or their interaction. However, we did find a 3-way interaction among Dyad Gender, Negotiating Strategy, and Global Trust, γ 33 = −0.14, t(114) = −2.23, p < 0.05. As shown in Fig. 4, sins of commission were unaffected by Global Trust or Negotiating Strategy in MM dyads. In FF dyads, deception increased when trust was high and this effect was more pronounced when negotiators implemented an accommodating strategy. Conversely, in MF dyads, deception increased when trust was low and negotiators implemented an accommodating strategy.

Fig. 4
figure 4

Average use of sins of commission as a function of global trust, dyad gender, and negotiating strategy

Full size image

Our second analysis tested whether the interaction among Dyad Gender, Negotiating Strategy and trust could be isolated to a specific sub-component of trust. Consistent with H2, identity-based Trust interacted with Dyad Gender and Negotiating Strategy to influence the use of sins of commission, γ 63 = -0.72, t(101) = −4.45, p < .001. As shown in Fig. 5, sins of commission were highest when negotiators in FF accommodated and identity-based trust was high or when negotiators in MF dyads accommodated and identity-based trust was low (H3c). Sins of commission were unaffected by trust or strategy in MM dyads.

Fig. 5
figure 5

Average use of sins of commission as a function of identity-based trust, dyad gender, and negotiating strategy

Full size image

Deterrence-based Trust interacted with Dyad Gender to affect sins of commission, γ 52 = 0.74, t(101) = 2.44, p < .05. This 2-way interaction was qualified by a 3-way interaction between Trust, Dyad Gender, and Negotiating Strategy, γ 53 = 0.73, t(101) = 2.43, p < .05. As shown in Fig. 6, sins of commission were used most frequently in MF dyads when deterrence-based trust was low and negotiators accommodated (H2). In FF dyads, sins of commission decreased when deterrence-based trust was low and negotiators competed; and, in MM dyads sins of commission were unaffected by trust or strategy.

Fig. 6
figure 6

Average use of sins of commission as a function of deterrence-based trust, dyad gender, and negotiating strategy

Full size image

These patterns replicate the pattern of deception we observed in relation to Global Trust in MM and MF dyads, suggesting that the effects of Global Trust on the use of sins of commission in these dyads can be isolated to the assessment of an opponent’s identity and the ability to impose sanctions. In FF dyads, the link between identity-based trust and sins of commission mirrored the pattern we observed for Global Trust, suggesting that the effects of Global Trust on the use of sins of commission in these dyads can be isolated to the assessment of an opponent’s shared values. The impact of deterrence-based trust on sins of commission did not, however, mirror the impact of Global Trust suggesting that negotiators in all FF dyads interpret information about punitive capability (ability to impose sanctions) differently to negotiators in MM and MF dyads.

Supplementary Analysis

To clarify the effects that we observed in mixed-sex dyads, we undertook a further analysis of only those dyads. We did this to determine whether the effects in mixed-sex dyads reflect a dyad-level or an individual-level dynamic. Our analyses showed that individual’s gender did not affect the use of either sins of commission or sins of omission in its own right (γ10 = 2.52, t(29) = 1.09, ns, sins of commission; γ10 = −0.19, t(29) = −0.83, ns, sins of omission) or in interaction with negotiating strategy (γ11 = 2.24, t(29) = 0.89, ns, sins of commission; γ11 = −0.39, t(29) = −1.62, ns, sins of omission). These results suggest that characteristics of the negotiating dyad rather than characteristics of the individual influence the use of deception.

Negotiators’ Outcomes

Although we had no hypotheses about negotiators’ outcomes, the scoreable nature of our negotiation task allows us to test whether using deception affects negotiators’ outcomes. To calculate negotiators’ outcomes, we calculated the points they obtained for the final contract. Using the SPSS procedure Mixed Models to control for dyad membership, we found that sins of omission, F(1, 121) = 4.5, p < 0.05, but not sins of commission, F(1,121) = 1.47, ns, predicted negotiators’ outcomes. To better understand this effect, we then separated integrative from distributive issues. Our first analysis showed that the use of deception did not affect negotiators’ outcomes across distributive issues (sins of omission, F(1,121) = 0.56, ns; sins of commission, F(1,121) = 0.27, ns). Sins of omission, F(1,121) = 7.23, p < 0.005, but not sins of commission, F(1,121) = 2.16, ns, predicted negotiators’ outcomes on integrative issues: the more sins of omission, the higher negotiators’ profits. These results suggest that negotiators who strategically pair an indifference issue with an integrative issue in a trade-off are able to claim a greater share of available value.

Discussion

Social context, which is shaped by both dyad- and individual-level factors, plays an important role in negotiators’ decisions to use deception. In this research, we clarified the relationships between social context and deception. We focused on three features of social context: dyadic gender composition, the gender-congruence of negotiation strategies, and initial impressions of trustworthiness. Our analyses showed that these three variables interacted to determine not only whether negotiators deceived the other party but also whether they chose to deceive by withholding or misrepresenting information.

We also demonstrated that as negotiators increased their use of sins of omission, they enhanced their ability to claim value. Negotiators who used sins of omission to obtain concessions on integrative (nonzero-sum) issues were able to claim a greater share of resources than those who did not. The same effect was not apparent in relation to distributive (zero sum) issues, suggesting that negotiators benefit from strategic deception only when they use it to replace low-value issues with no-value issues in their trade-offs (sin of omission).

Social Context and Deception

The decision to withhold information (sin of omission), but not the decision to misrepresent information (sin of commission), was affected by dyad composition and strategy choice. Our analysis showed that the decision to withhold information was affected by dyad composition when negotiators accommodated but not when they competed. This effect was most pronounced in dyads with at least one female negotiator: whereas all-male dyads’ use of deception was stable across negotiating strategies, all-female dyads decreased and mixed-dyads increased their use of deception when the other party accommodated. Consistent with our hypotheses, these findings suggest that all-female and mixed-sex dyads use different reference points for establishing behavioral expectations. In all-female dyads, accommodation is judged against gender stereotype-based expectations, consequently reducing concerns about exploitation (Simpson and Van Vugt 2009; Wang and Yamagishi 2005). In mixed-sex dyads, the same behavior is compared to stereotype-based expectations of an effective negotiator, and encourages the pursuit of self-interest via deception.

The 3-way interactions among Dyad Gender, Negotiating Strategy and Trust that we observed showed that negotiators focus on different components of trustworthiness in deciding whether to withhold or misrepresent information: whereas evidence of benevolence affected the use of sins of omission, evidence of shared values (identity-based trust) and the ability to sanction opponents (deterrence-based trust) affected the use of sins of commission. These interactions also showed that the same information was used differently depending on dyad gender composition.

We were better able to predict the use of sins of omission than the use of sins of commission. Our analysis showed that, in dyads with at least one female negotiator, expectancy violations combined with low-benevolence-based trust increased sins of omission. However, in all-male dyads, low benevolence-based trust reduced sins of omission when negotiators accommodated. Negotiators’ decisions to use sins of commission were more variable. As we predicted, in mixed-sex dyads, expectancy violations combined with low deterrence- and low identity-based trust increased sins of commission. However, in all-female dyads, expectancy violations and low deterrence-based trust decreased sins of commission, and expectancy confirmation and high identity-based trust increased sins of commission. In all-male dyads, neither trust nor strategy affected sins of commission. These findings yield four insights into the role of social context in shaping deception.

First, the use of deception in mixed-sex dyads was more predictable than in same-sex dyads. Consistent with our hypotheses, negotiators in mixed-sex dyads increased their use of deception when opponents who accommodated were seen as untrustworthy. It supports our argument that, in mixed-sex dyads, the competitive style linked to the stereotypes of effective negotiators sets the behavioral norms. It further demonstrates that expectancy violations, in combination with low trust, prime deception. The pattern that emerged in mixed-sex dyads also suggests that negotiators in mixed-sex dyads act pragmatically, increasing their use of deception as they collect evidence that they are vulnerable to exploitation (moral pragmatism; Dees and Cramton 1991).

Second, unlike negotiators in mixed-sex dyads, negotiators in same-sex dyads responded to negative information about opponents (expectancy violation, low trust) by decreasing their use of deception. In all-male dyads, an expectancy violation combined with low benevolence-based trust reduced sins of omission, and in all-female dyads an expectancy violation combined with low deterrence-based trust reduced sins of commission. When contrasted with mixed-sex dyads, the overall pattern demonstrated that the same evidence may trigger different consequences depending on dyad gender composition. Theories of ethical behavior more generally propose that, in deciding on a course of action, individuals weigh the benefits of unethical action relative to its costs (Allingham and Sandmo 1972). One interpretation of our findings is that dyad gender composition shifts the relative weight assigned to these two aspects of the decision-making process. Whereas negotiators in mixed-sex dyads, armed with evidence of ill intentions, react offensively to prevent exploitation, negotiators in same-sex dyads armed with the same evidence react defensively to limit the potential costs of deceiving their opponents.

Third, in their decisions to misrepresent information, negotiators in all-male dyads appear less sensitive to social context than negotiators in all-female and mixed-sex dyads. Whereas negotiators in all-female and mixed-sex dyads adjusted their use of sins of omission in response to trustworthiness and negotiating strategy, all-male dyads did not. Croson and Gneezy (2009) report similar results in their review of the economic games literature, suggesting that overall men are less likely to factor variations in context into their decisions. It is however surprising that this insensitivity emerges in relation to the more serious form of deception, that is, sins of commission (Spranca et al. 1991). On interpretation of this finding is that men may be more utilitarian in their decision-making (Schminke et al. 1997). In an “ends justify the means” world, nuances of context are unlikely to affect ethical decisions.

Fourth, unlike negotiators in mixed-sex dyads, those in all-female dyads did not use information about strategy and trustworthiness in a consistent way. As we have already highlighted, low deterrence-based trust in combination with an expectancy violation (competing strategy) reduced sins of commission. However, high identity-based trust in combination with an expectancy confirmation (accommodating strategy) increased sins of commission. This latter finding fits better with an opportunistic approach to deception than with a pragmatic approach. It suggests that, faced with a benign environment in which the potential costs of detection are low, all-female dyads are willing to pursue self-interest (also Olekalns and Smith 2007, 2009). More generally, these patterns parallel findings that women not only use a greater range of fairness principles, but also use conditional fairness principles (Croson and Gneezy 2009; Miller and Ubeda 2012) demonstrating that the decision to misrepresent relies on multiple decision criteria.

Implications for Theory and Practice

Our findings demonstrate the benefits of treating trust as a multi-dimensional construct. Although in general the sub-components of trust behaved in the same way as global trust, this was not the case in all-female dyads. Moreover, because we were able to isolate the role of specific trustworthiness cues in triggering either sins of omission or commission, we are better able to specify the conditions under which negotiators may elicit both forms of deception. Whereas benevolence-based trust affected sins of omission, identity- and deterrence-based trust affected sins of commission. Both identity- and deterrence-based trust afford behavioral control, indirectly through shared values (identity) or directly through the ability to impose sanctions (deterrence). Consequently, our findings suggest that whereas sins of omission may be elicited by judgments about the individual, sins of commission are elicited by judgments about behavioral predictability. We conclude that, in combination with an opponent’s choice of strategy and dyad composition, negotiators use sins of omission when they believe the other negotiator is ill-intentioned but use sins of commission when they believe they lack behavioral guarantees.

To the extent that women in all-female dyads used shifting criteria for determining when to deceive the other party, we concluded that women inhabit a more complex social world than men. Negotiators in all-female dyads adjusted their use of deception based on unique constellations of trust and strategy. The more nuanced decisions made by negotiators in all-female dyads may be better understood if we consider the possibility that these criteria combine to provide an assessment of behavioral latitude and the likely costs of deception (Jones 1991). High identity-based trust increases tolerance for expectancy violations (Lewicki and Weithoff 2000), and accommodation emphasizes relationship maintenance. In combination, these cues are likely to decrease the perceived costs of misrepresentation thereby increasing behavioral latitude. In contrast, low deterrence-based trust, which reduces behavioral guarantees, and competition, which emphasizes an individualistic orientation, are likely to increase the perceived costs of misrepresentation thereby reducing behavioral latitude.

Our findings highlight the need for negotiators to identify and assess all potential contextual triggers for deception before starting a negotiation. These triggers can be relatively subtle, based on initial trust in the other person, as well as the degree of perceived similarity between negotiators (in this case, same or different gender). Negotiators need to be aware that the signals they convey about their trustworthiness may prime the other party to deceive them. In particular, negotiators need to be aware that their opponents make trust judgments along multiple dimensions, and that the assessment of negotiators on each dimension of trust shapes opponents’ use of deception. However, low trust does not necessarily trigger deception: low trust comes into play only when negotiators violate behavioral expectations. These findings suggest that negotiators should take care to prevent expectancy violations while simultaneously signaling trustworthiness to the other party.

At the same time, negotiators need to be aware of opportunism. Our results (also Olekalns and Smith 2007, 2009) show that deception can be elicited under the best of circumstances: when the other party poses no threat and when negotiators believe the other party shares their goals and values. In this research, we found that when women in all-female dyads reported high identity-based trust and encountered an accommodating strategy, they increased their use of misrepresentation. These results extend the finding that women are more likely to elicit opportunistic deception (Kray et al. 2012), suggesting that opportunism is encouraged in benign environments. Our results suggest that women in all-female dyads should be especially alert to the impressions that they convey to the other party and may benefit from actively downplaying their similarities.

Limitations and Future Directions

We used a simulated employment contract negotiation to test our hypotheses. Although simulations yield results comparable to real-life negotiations (Donohue et al. 1984; Herbst and Schwarz 2011), our simulation provided negotiators with a highly structured environment. Even under such structured conditions, we were able to demonstrate that three social context variables—negotiating strategy, dyad gender composition, and initial trust—worked together to elicit or constrain deception in negotiation. Given that behavioral latitude increases as contextual ambiguity increases, we believe there is merit to exploring how these variables affect deception under conditions of greater uncertainty.

The idea that different motives prime deception in men and women underpinned our explanations of the gender differences in deception that we observed. Although there is some evidence that men act out of greed, whereas women act out of fear (Simpson and Van Vugt 2009; Wang and Yamagishi 2005), this assumption would benefit from further testing. Moreover, our results suggest that women may act out of either motive whereas men may not. Knowing how these motives shape the use of deception, and micro-ethical actions more generally, would increase our understanding of the conditions that trigger unethical behavior.

Extending this idea, the patterns we observed may reflect the different ethics embraced by women and men, who are more inclined to formalism and utilitarianism, respectively. The finding that, in all-male dyads, neither trust nor strategy affected the use of deception points to a utilitarian world view on the part of men in which the violation of behavioral norms is not seen as a moral issue (Schminke et al. 1997). Consistent with these different world-views, we note that deception was higher in dyads with at least one male negotiator than in all-female dyads. More broadly, it may be that women espouse an “ethic of community”, focusing on such values as loyalty, whereas men espouse an “ethic of autonomy”, focusing on personal welfare (Shweder et al. 1997). Better understanding the values that women and men bring to negotiations will increase our understanding of, and ability to predict, the factors that trigger or constrain unethical actions on the part of men and women.

Finally, we showed that dyad gender composition matters when individuals make ethical decisions. Although our focus was on dyadic negotiations, our findings raise the question of how diversity affects ethical action. For example, Abdolmohammadi et al. (1997) show that all-male groups make more ethical judgments than might be predicted by their individual opinions, whereas all-female groups make less ethical judgments than might be predicted by their individual opinions. These findings point to the need to consider how the demographic make-up of negotiating teams or decision-making groups affects individuals’ (un)ethical actions.

Conclusion

Deciding what to tell another negotiator is critical to how a negotiation unfolds. While giving accurate information helps the problem-solving process, it also increases the risk of exploitation. We verified this assertion by showing that negotiators who withhold information are able to claim more value than those who do not. We also showed that negotiators’ propensity to deceive varied as a function of three social context variables: the gender composition of the negotiating dyad, whether strategies were gender congruent or incongruent, and initial trust in the other party. A key finding was that whereas both trust and strategy affected deception in all-female and mixed-sex dyads, neither variable affected deception in all-male dyads. As the number of female negotiators increased, the criteria that trigger deception become more complex. In mixed-sex dyads there was clear evidence of a morally pragmatic stance: negotiators increased deception when trust was low and their behavioral expectations were violated. Although, for the most part, negotiators also adopted a morally pragmatic stance in all-female dyads, we observed some evidence of opportunism in these dyads. We concluded that negotiators in all-female dyads more carefully assess the behavioral latitude afforded by social context, and the associated costs of deception, before misrepresenting information.