You Must Have a Preference: The Impact of No-Preference Communication on Joint Decision Making

Joint Decision Making

Consumers frequently make consumption decisions involving other people, in which they need to reach a decision about offerings that will be jointly consumed. Of course, joint-decision-making processes are often more complex and effortful than individual decision making because they involve multiple parties, each with their own goals, beliefs, and sources of utility. Classic work in the area of joint decision making has focused on the “household” as the primary unit of analysis, investigating how various factors influence major family purchases such as houses, cars, and home furnishings (e.g., Davis 1976). For example, this research examined how purchase outcomes (i.e., the ultimate choice among several options) depend on sex roles and expectations (Qualls 1987), power and authority dynamics (Burns and Granbois 1977; Fisher, Grégoire, and Murray 2011), perceived influence (Davis, Hoch, and Ragsdale 1986), and education (Rosen and Granbois 1983).

Other early research on this topic went beyond the outcomes to examine the process through which families resolve conflict and reach decisions. In such instances, consumers attempt to jointly integrate multiple preference structures to maximize group-level consumption utility (for all parties involved) while also maintaining positive social relationships (Davis 1976; Park 1982). Because it is rarely the case that two consumers’ preferences perfectly align, parties involved in a joint decision employ various strategies to resolve disagreements (e.g., bargaining, aggression; Kirchler 1993; Spiro 1983). Even so, it is difficult to reach a decision that satisfies several distinct preference structures (Lowe et al. 2019), and one or more parties usually ends up revising their preferences or conceding (Aribarg, Arora, and Bodur 2002).

Recently, there has been a resurgence of interest in joint decision making within marketing, with many researchers calling for more work on the topic (Belk 2010; Gorlin and Dhar 2012; Liu, Dallas, and Fitzsimons 2019; Simpson, Griskevicius, and Rothman 2012). By definition, joint decision making involves integrating different tastes, opinions, and preferences of multiple parties with different priorities, needs, and motivations (Liu, Dallas, and Fitzsimons 2019). As such, decision makers often realize that joint consumption choices considerably affect their relationships with co-consumers, which in turn affect their choices (Dzhogleva and Lamberton 2014; Garcia-Rada, Anik, and Ariely 2019). For example, in the context of close friends, consumers may try to select the option that maximizes joint utility within the relationship (Tu, Shaw, and Fishbach 2016); in the context of spouses, consumers may consider dynamic, long-term effects across the relationship time horizon (Su, Fern, and Ye 2003).

Communicating No Preference

In all of this prior literature, successfully arriving at a satisfactory joint decision requires knowledge of the other party's preferences, so that one can balance them with one's own preferences. Thus, an essential first step toward a successful joint decision is to understand the preferences of the other party. Despite this importance, relatively little attention has been given to examining how consumers disclose and learn each other's preferences. Specifically, much of the empirical work in this area has examined contexts in which either two people are already aware of each other's preferences, such as in very close relationships (Park 1982; Spiro 1983; Su, Fern, and Ye 2003), or the other party's preferences are externally imposed by the researcher's or a confederate's explicit communications of preference (Dzhogleva and Lamberton 2014; Raghunathan and Corfman 2006).

Indeed, for higher-involvement joint decisions, such as purchasing a home or a car, consumers would want to clearly communicate their preferences to facilitate an in-depth discussion to resolve differences and reach a satisfactory decision (e.g., Corfman and Lehmann 1987; Park 1982; Qualls 1987). Consistent with this reasoning, the research stream on joint decision making often assumes that all parties in a joint decision will initially exchange information about their individual preferences (Aribarg, Arora, and Bodur 2002; Simpson, Griskevicius, and Rothman 2012). Then, after each party's preferences are communicated, and given that typically such preferences do not perfectly align, the parties employ various strategies to influence each other and resolve disagreements (Kirchler 1993; Spiro 1983) and often end up revising their preferences or conceding (Aribarg, Arora, and Bodur 2002).

However, given that many joint decisions are lower involvement than buying a home or a car, it is reasonable to assume that the parties involved would prefer to ease the decision-making process as much as possible and maintain a positive social exchange. One intuitive way of doing this is for one party simply to state that they have no particular preference among the available options. Such a strategy may appear useful when making mundane day-to-day decisions (e.g., choosing a restaurant, picking which coffee to bring back to the office), where a small increase in consumption utility may not be as important as avoiding potential conflicts. That is, given that disclosing information about one's preferences introduces constraints and potentially reveals a preference mismatch (Corfman and Lehmann 1987; Norton, Frost, and Ariely 2007; Park 1982), communicating no preference may seem appealing.

We note that besides trying to make the decision easier for themselves and for others, consumers may have many additional reasons for communicating no preference to the other party. In some instances, a consumer may truly feel indifferent between various options if the options seem similar to them or offer equal consumption utility. In other cases, while consumers may prefer one option, they may still want to make a positive impression on the other party by relinquishing control and appearing accommodating (Kardas, Shaw, and Caruso 2018). Indeed, a recent study demonstrates that consumers sometimes communicate that they have no preference in order to appear easygoing and likable (Liu and Min 2020). Consumers may also want to delegate a decision to the other party because they believe they have lower levels of expertise or personal involvement (Davis, Hoch, and Ragsdale 1986) or because they want to avoid the burden of responsibility associated with making a decision or feeling at fault if the choice is suboptimal (Steffel and Williams 2017). Nonetheless, our theory—developed further in the next section—is agnostic to the person’s underlying motive for communicating no preference. That is, regardless of the reason, expressing no preference can unintentionally lead to several negative downstream consequences for the joint experience, impacting the decision maker's decision difficulty, ultimate choice, and liking of the co-consumer.

The Impact on Decision Difficulty

First, we propose that when faced with a no-preference communication, decision makers will experience several sources of difficulty in making a decision. In the case of joint decision making, information about the other party's preferences is a critical component that directly impacts the decision maker's ability to make a selection (Gorlin and Dhar 2012; Simpson, Griskevicius, and Rothman 2012). This is because consumers making joint decisions try to accommodate both parties’ preferences and maximize the joint utility of the dyad (Liu, Dallas, and Fitzsimons 2019). Thus, without specific indication of a co-consumer’s actual preference, or at least an authentic signal of true indifference between the options, decision makers will likely wonder about the co-consumer’s true preferences in an attempt to estimate and predict the group's overall utility from each option. As such, to the extent that decision makers believe that the co-consumer's preferences do exist but are not expressed, their task of choosing for the dyad should become more difficult because they have the additional step of trying to infer the co-consumer's true preference structure.

In addition, prior research relating to the state of suspicion (Campbell and Kirmani 2000; Hilton, Fein, and Miller 1993) has repeatedly demonstrated that wondering about others’ true disposition can lead to a variety of outcomes related to greater cognitive activation (e.g., loss of working memory, excess body movement, decreased eye blinking; for a review, see Bobko et al. [2014]). Such an increase in cognitive effort is explained by one's tendency to generate and consider multiple plausible motivations for a certain behavior (e.g., Fein and Hilton 1994; Hilton, Fein, and Miller 1993; Kim and Levine 2011) and to engage in more complex attributional thought processes (DeCarlo 2005). Notably, this research has postulated that the link between a state of suspicion and greater cognitive activation occurs regardless of whether such suspicion is accompanied by perceptions of potential malintent or positive motivations (e.g., suspicion about others planning one a surprise party or trying to ease the delivery of bad news; DePaulo et al. 1996). Similar processes may ensue in the face of a no-preference communication (as decision makers exert effort to infer the co-consumer's preferences), even if the decision maker believes that the co-consumer expressed no preference for social reasons (e.g., to be accommodating).

In summary, we argue that both a lack of information and the mere state of suspicion activated by perceptions of undisclosed preferences will trigger greater decision difficulty compared with situations in which the other party’s preferences are explicitly communicated.

H_2a: Compared with when the co-consumer's preference is explicitly communicated, no-preference communication increases decision makers’ decision difficulty.

The Impact on Social Utility

Second, we propose that expressing no preference can have social costs for how the co-consumer is viewed by the decision maker. Prior research has demonstrated that self-disclosure and honesty are valued social traits (Van Lange and Kuhlman 1994) that help develop interpersonal relationships (Altman and Taylor 1973). In contrast, hiding personal information results in less positive perceptions and can create a sense of distance between two people (John, Barasz, and Norton 2016). Consistent with this notion, people who are not direct in regular conversations are liked less than those who phrase the same communication more directly (Holtgraves 1986).

Integrating these findings with our prediction that no-preference communication will trigger perceptions of undisclosed preferences, we predict that co-consumers who express no preference will be liked less compared with those who explicitly express their preferences. It is important to note that we expect this effect to hold regardless of the co-consumer's motivation for stating that they have no preference (as long as it triggers perceptions of undisclosed preferences). Even prosocial liars (i.e., people who lie to benefit another person), who have good intentions, are viewed less favorably for not being truthful (Levine and Schweitzer 2015). Moreover, any uncertainty about others’ true motives can cause individuals to see them in a more negative light, even if those motives are clearly not ulterior (Fein and Hilton 1994). Therefore, even if the co-consumer has a prosocial reason for stating no preference (e.g., trying to accommodate the decision maker's desire or ease the decision-making process), we predict that expressing no preference, in and of itself, can lead decision makers to like the co-consumer less for not being willing to disclose what they prefer.

H_4a: Compared with when the co-consumer's preference is explicitly communicated, decision makers who receive a no-preference communication like the co-consumer less.

The Impact on Consumption Utility

Finally, we propose that no-preference communications will have substantial implications for decision makers’ consumption utility. When making an independent decision (i.e., alone), consumers try to maximize their consumption utility and choose in accordance with their own preferences (e.g., Simpson, Griskevicius, and Rothman 2012). In joint decisions, however, consumers often go through an effortful decision-making process to arrive at a compromise option, balancing the self's and the other person's preferences (Liu, Dallas, and Fitzsimons 2019). In fact, consumers in joint decisions often weigh their partner's preferences more heavily than their own (Garcia-Rada, Anik, and Ariely 2019).

We have reasoned that because decision makers assume that co-consumers have undisclosed preferences, they will try to infer those preferences to maximize the joint consumption utility. Although previous work in marketing has examined inferences of another person's preferences either when there is no information present (Weaver and Hamby 2019; Wu et al. 2021) or when the preferences are explicitly communicated (Davis, Hoch, and Ragsdale 1986; Lerouge and Warlop 2006), it has yet to investigate the inferences people make when the other party communicates no preference. Interestingly, whereas simply not knowing other peoples’ preferences (e.g., lack of information) may trigger processes that would lead to predictions of similar preferences (e.g., false consensus effects; Gorlin and Dhar 2012; Mullen et al. 1985; Ross, Greene, and House 1977), we posit that an active communication of no preference is different because it triggers the perception that the other person is not disclosing their true preferences. Specifically, building on this belief that the co-consumer is not disclosing their true preference, we argue that decision makers also infer that the co-consumer's true preference must be dissimilar to their own (otherwise, why wouldn’t they simply say it?). Indeed, prior research has found that when people are uncertain about the accuracy of a communication (Fein and Hilton 1994), they tend to engage in counterfactual thinking (DeCarlo 2005; Hilton, Fein, and Miller 1993; Schul, Mayo, and Burnstein 2004) and construe the other person's dispositions as opposite to their own (Kruglanski 1989; Sitkin and Roth 1993). Consumers often anchor on their encounters with dissimilar others (Sanders and Mullen 1983) and believe that their preferences are more distinct from those of others than they actually are (Davis, Hoch, and Ragsdale 1986; Suls and Wan 1987). Finally, the social nature of everyday joint consumptions may exacerbate the extent to which no-preference communication indicates dissimilarity. That is, conversational norms in which consumers are expected to maintain pleasant conversations (Berger 2014; Grice 1975; Hilton 1995) may contribute to the perception that the co-consumer is withholding their specific preference, because no-preference communication is seen as an attempt to avoid a potential negative exchange from conflicting preferences. Consistent with these points, we found in a pilot study (N = 161) that decision makers inferred that a co-consumer who expressed no preference actually had preferences that were significantly different from (i.e., dissimilar to) their own preferences (i.e., compared with a co-consumer who explicitly expressed having similar preferences; F(1, 105) = 55.17, p < .001; for full study details, see Web Appendix Study 2).

Accordingly, given decision makers’ motivation to maximize the utility of the joint experience, we argue that they may overly adjust from their own preference to their partner's preferences and choose an option they prefer less (as if co-consumers had directly indicated having dissimilar preferences). That is, as decision makers attempt to choose an option that can adequately balance their own preference and the co-consumer's potentially dissimilar preferences, they may ultimately select an option that lowers their own consumption utility.

H_5a: Decision makers who receive a no-preference communication from a co-consumer choose an option that they prefer less compared with when choosing an option for themselves alone.

Design, Procedure, and Measures

A total of 236 ³ online participants from MTurk (M_age = 41.88 years; 51.7% female) completed the preregistered study (https://aspredicted.org/VFN_5BW). All participants read a scenario in which they were asked to imagine that they were getting dinner with a friend and were trying to decide together which restaurant to go to out of three nearby restaurants. Participants were then randomly assigned to one of two preference communication conditions: their friend either expressed no preference to them (no-preference condition) or explicitly expressed their preference to them (explicit-preference condition) regarding which restaurant to choose. Across the experimental conditions, we varied the phrase used to express no preference with five phrases selected from a pretest (for procedure and results, see Web Appendix A), as well as the phrase used to express an explicit preference with five corresponding phrases. Thus, the study consisted of a two-cell (preference communication: no preference vs. explicit preference) between-subjects design, where those assigned to the no-preference condition imagined hearing one of five no-preference communication phrases (i.e., “Let's go where you want” vs. “I don’t know” vs. “You decide” vs. “I don’t care” vs. “I’ll go wherever”), while those assigned to the explicit-preference condition imagined hearing one of five explicit-preference communication phrases (i.e., “I’m leaning toward option A” vs. “Option A sounds good to me” vs. “I prefer option A” vs. “I like option A” vs. “Let's go to option A”).

As our main dependent variable, participants were asked to evaluate their difficulty in making a joint decision after receiving the co-consumer's preference communication. Specifically, they were asked “To what extent do you feel they make it easier for you to decide?” (1 = “a great deal,” and 7 = “not at all”) and “How difficult would it be to make the decision?” (1 = “not difficult at all,” and 7 = “very difficult”). These two items were correlated (r  =  .59, p < .001) and were averaged to form the decision difficulty measure. Thus, higher scores on this item indicate greater decision difficulty.

Next, we measured perceptions of undisclosed preferences. Participants indicated the extent to which they believed the co-consumer actually had true preferences, but was not disclosing them, using two items (i.e., “Do you think they are hiding their true preferences from you?” [1 = “s/he are definitely not hiding their true preference,” and 7 = “s/he are definitely hiding their true preference”] and “Do you think they are trying to keep their true preference from you?” [1 = “s/he are definitely not trying to keep their true preference from me,” and 7 = “s/he are definitely trying to keep their true preference from me”]). These two items were highly correlated (r  =  .93, p < .001) and were averaged to form the perception of undisclosed preferences measure. The factor analysis for the main measures in this study and in all other studies appear in Web Appendix E.

Results

Decision difficulty

In support of H_2a, a one-way analysis of variance (ANOVA) on decision difficulty revealed a significant main effect of preference communication (F(1, 235) = 60.74, p < .001, η p 2  = .206). Specifically, participants in the no-preference condition reported significantly greater decision difficulty (M = 3.89, SD = 1.67) compared with those in the explicit-preference condition (M = 2.40, SD = 1.24; for a breakdown of all ten phrases, see Figure 2; for contrasts between all ten phrases, see Web Appendix A).

OPEN IN VIEWER

Figure 2.

Decision Difficulty (Study 1).

Discussion

Study 1 provided evidence for our prediction that in a joint decision, no preference negatively impacts decision difficulty for decision makers compared with an explicit-preference communication (H_2a). We also found that the negative impact of no preference on decision difficulty is driven by the decision maker's perception that the co-consumer actually has a preference but is not disclosing it (H_2b). The results were robust across a variety of phrases that consumers report often using in real joint consumption to communicate having no preference.

Design, Procedure, and Measures

We recruited 584 participants on MTurk (M_age = 38.20 years; 45.0% female). Participants were asked to imagine jointly deciding which restaurant to go to for dinner with a friend. This study employed a 2 (preference communication: no preference vs. explicit preference) × 2 (perspective: decision maker vs. co-consumer) between-subjects design. As in Study 1, in the two decision-maker-perspective conditions, participants were told to imagine that “your friend told you that they [do not] have a specific preference for one option over others” (in the explicit- vs. no-preference conditions, respectively). In the two co-consumer-perspective conditions, participants were told to imagine that “you told your friend you [do not] have a specific preference for one option over others” (in the explicit- vs. no-preference conditions, respectively).

As our main dependent variable, we measured (actual or expected) decision difficulty using three items. Specifically, participants rated three items, including “To what extent do you feel they [you] were making it easier for you [the other person] to decide?” (1 = “a great deal easier,” and 7 = “not at all easier”), “To what extent do you feel they [you] were making it easier versus more difficult for you [the other person] to decide?” (1 = “a great deal easier,” and 7 = “a great deal more difficult”), and “How much more effort do you think you [the other person] would need to put into making this decision?” (1 = “no more effort,” and 7 = “a lot more effort”). These three items loaded together on one factor and were averaged to form the decision difficulty measure (α = .88). Next, we assessed perception of undisclosed preferences by having participants rate “To what extent will you [do you think your friend will] believe that your friend [you] actually prefer(s) one option over others?” (1 = “not at all,” and 7 = “a great deal”).

Results

Decision difficulty

As we predicted, a 2 (preference communication) × 2 (perspective) ANOVA revealed a significant interaction (F(3, 580) = 10.18, p = .001, η p 2  = .017). In support of H_3a, in the no-preference condition, there was a discrepancy between perspectives: the no-preference communication made the decision significantly more difficult for the decision maker (M = 4.47, SD = 1.61) compared with what the co-consumer anticipated (M = 4.00, SD = 1.59; F(1, 580) = 5.99, p = .015, η p 2  = .010). In the explicit-preference condition, this discrepancy reversed, such that decision makers actually found that the co-consumer’s explicit preference made the decision easier (M = 3.45, SD = 1.74) than the co-consumers expected (M = 3.84, SD = 1.55; F(1, 580) = 4.26, p = .040, η p 2  = .007). ⁴

Looked at another way, replicating Study 1 and supporting H_2a, compared with receiving an explicit preference (M = 3.45, SD = 1.74), decision makers who received a no-preference expression reported significantly greater decision difficulty (M = 4.47, SD = 1.61; F(1, 580) = 38.87, p < .001, η p 2  = .047). However, co-consumers did not expect a difference in decision difficulty as a function of their preference communication (M_{no pref} = 4.00, SD = 1.59 vs. M_{exp pref} = 3.84, SD = 1.55; F(1, 580) = .69, p = .408). In addition, a significant main effect of preference communication emerged, such that explicit preference communication led to lower (anticipated and actual) decision difficulty (M = 3.65, SD = 1.66) compared with no-preference communication (M = 4.24, SD = 1.62; F(3, 580) = 19.08, p < .001, η p 2  = .032), but there was no significant main effect of perspective on decision difficulty (p = .777). Figure 3 displays these results.

OPEN IN VIEWER

Figure 3.

Decision Difficulty (Study 2).

Discussion

Study 2 replicated the findings from Study 1, in that no-preference communication negatively impacts decision makers’ decision difficulty compared with an explicit preference communication (H_2a), and this effect is driven by the perception of undisclosed preferences (H_2b). Importantly, Study 2 also provided evidence for a discrepancy, such that no-preference communication negatively impacts the decision difficulty of decision makers more than co-consumers expect (H_3a–b).

To confirm the robustness of these effects, we also tested the discrepancy between the co-consumer's expectations and the decision maker's actual decision difficulty in two follow-up studies. The first study (Web Appendix Study 1) was a recall study in which we asked participants to recall their own past joint consumption decisions where they had been either a recipient or a communicator of no preference. In the second study (Web Appendix Study 3), we manipulated no-preference communication using the same five ecologically valid phrases from Study 1. Both studies replicated the findings of Study 2, such that no-preference communication negatively impacted the decision difficulty of decision makers more than co-consumers expected (H_3a).

Design, Procedure, and Measures

We recruited 778 participants from MTurk (M_age = 35.19 years; 49.6% female) for this study. Participants were asked to imagine making a joint decision with another person, such as deciding which restaurant to go to, which movie to watch, which food to order, or which gift to buy together for a third party. This study employed a 2 (perspective: decision maker vs. co-consumer) × 3 (perception of undisclosed preferences: control vs. undisclosed preference reinforced vs. undisclosed preference diminished) between-subjects design. We manipulated decision-maker versus co-consumer perspective by asking participants to imagine “hearing that the other person tells you they [telling the other person that you] have no specific preference.”

The second factor we manipulated was the perception that the co-consumer is hiding their preferences. In the control condition, participants did not receive additional information. Thus, participants simply imagined hearing or expressing a no-preference communication. However, participants assigned to the undisclosed-preference-reinforced condition were further told that they “suspect the co-consumer does have a preference,” while those assigned to the undisclosed-preference-diminished condition read that they “believe the co-consumer indeed does not have preferences.” Then, participants indicated their [the decision maker's] difficulty in making the joint decision by rating the same item as Study 2 (i.e., “To what extent do you feel they [you] were making it easier for you [the other person] to decide?”; 1 = “a great deal,” and 5 = “not at all”).

Results

Decision difficulty

A 2 (perspective) × 3 (perception of undisclosed preferences) ANOVA revealed a significant main effect of perspective: decision makers experienced greater difficulty (M = 3.57, SD = 1.22) than co-consumers anticipated (M = 3.11, SD = 1.16; F(5, 777) = 29.52, p < .001, η p 2  = .037). We found no main effect for perception of undisclosed preferences (F(5, 777) = 1.19, p = .304, η p 2  = .003). Importantly, as we predicted, the interaction between perspective and perception of undisclosed preferences was significant (F(5, 777) = 5.60, p = .004, η p 2  = .014; see Figure 4). Replicating the results of Study 2, in the control condition, decision makers felt significantly greater decision difficulty (M = 3.52, SD = 1.21) compared with what co-consumers anticipated (M = 3.16, SD = 1.09; F(1, 772) = 6.12, p = .014, η p 2  = .008). In the undisclosed-preference-reinforced condition, the discrepancy was replicated and amplified, such that decision makers felt even greater decision difficulty (M = 3.84, SD = .995) compared with what co-consumers anticipated (M = 2.99, SD = 1.17; F(1, 772) = 33.03, p < .001, η p 2  = .041). However, when decision makers were told not to be suspicious of undisclosed preferences (undisclosed-preference-diminished condition), the discrepancy effect was attenuated: decision makers’ decision difficulty was not significantly different (M = 3.34, SD = 1.39) than what the co-consumer anticipated (M = 3.17, SD = 1.21; F(1, 772) = 1.38, p = .240, η p 2  = .002).

OPEN IN VIEWER

Figure 4.

Decision Difficulty (Study 3).

Looked at another way, the decision makers who were told not to be suspicious of undisclosed preferences (undisclosed-preference-diminished condition) felt significantly less decision difficulty (M = 3.34, SD = 1.39) compared with those in the undisclosed-preference-reinforced condition (M = 3.84, SD = .99; F(1, 772) = 11.12, p < .001, η p 2  = .041) and directionally less decision difficulty compared with those in the control condition (M = 3.52, SD = 1.21; p = .223). Moreover, those in the control condition felt significantly less decision difficulty (M = 3.52, SD = 1.21) than those in the undisclosed-preference-reinforced condition (M = 3.84, SD = .99; F(1, 772) = 5.40, p = .030, η p 2  = .021).

Discussion

Together, Studies 2 and 3 provide evidence for our discrepancy prediction: compared with what the co-consumers anticipate, expressing no preference increases decision makers’ decision difficulty (H_3a). We also showed, using both mediation and moderation designs, that this discrepancy arises due to decision makers’ perceptions of undisclosed preferences (H_3b). In a follow-up study, we compared no-preference communication with a condition in which a decision maker does not receive any communication from the co-consumer (i.e., no-communication condition) and found support for no-preference communication uniquely triggering the perception of undisclosed preferences. We discuss the results and implications of these findings in the “General Discussion” section.

In the following studies, we return our focus to the other key predictions of our article: comparing the expression of no preference with the expression of an explicit preference and examining how no-preference communication can lead to several negative consequences.

Design, Procedure, and Measures

A total of 120 students at a large North American university (M_age = 20.63 years; 59.5% female) participated in the study for course credit. First, all participants were assigned to a partner (i.e., a participant sitting next to them in the lab) and worked together on a task intended to familiarize themselves with each other. Specifically, they were instructed to pull their chairs closer to their assigned partner and to freely engage in a brief conversation using an abridged version of the relationship closeness induction task (Sedikides et al. 1999; see Web Appendix B), which has been shown to form friendships in lab settings. Next, participants returned to their individual computer stations to start the next task, which was the focal part of our study.

At this point, participants were told that they would be choosing a snack and sharing it with their partner (i.e., the person they just interacted with). Participants learned that they would have to choose one snack to consume together from a selection of four different snacks of similar sizes: Kit Kat, M&M’s, Welch's fruit snacks, and Biscoff (for visual stimuli, see Web Appendix B). All participants were assigned to the “decision maker” role and were led to believe that their partner was assigned to the “co-consumer” role. Specifically, they were told, “You have been assigned the role of telling the lab managers which snack you two will share. Your partner has been assigned the role of communicating to you about their preference, before you tell the lab manager.”

Prior to choosing a snack to share, participants were randomly assigned to one of two preference-communication conditions. Those assigned to the no-preference condition read, “Your partner communicated to you the following: ‘I have no preference—it's your call!’,” while participants assigned to the explicit-preference condition read, “Your partner communicated to you the following: ‘I like [Kit Kats]—but it's your call!’” The specific name of the snack that appeared in the explicit preference condition was counterbalanced evenly across the four available snacks. Notably, we included the explicit statement “it's your call!” in both conditions to control for the degree of delegation expressed within the communication (and thus decision makers’ feelings of responsibility for making the joint decision; Steffel and Williams 2017).

As our main dependent variable, we first measured participants’ difficulty in making the joint decision by asking them to rate “To what extent did you feel that your task partner was making it easier versus more difficult for you to decide?” (1 = “more easy,” and 7 = “more difficult”), consistent with that used in Study 2. In addition, this study measured social utility by asking participants to rate “How much do you like your partner?” (1 = “not like at all,” and 7 = “like a lot”).

As a manipulation check of preference communication, participants also rated “To what extent did your partner explicitly express his or her snack preference?” (1 = “not at all,” and 7 = “to a great extent”). We also measured perceptions of delegation by asking participants “To what extent did your partner delegate the snack decision?” (1 = “not at all,” and 7 = “to a great extent”). Once all participants completed the measures, the partners consumed their candy together and then were thanked and debriefed.

Results

Discussion

The results of Study 4 replicated Studies 1 and 2 and showed that no-preference (vs. explicit-preference) communication negatively impacts decision difficulty. We also found that decision makers liked the co-consumer less when they expressed no preference compared with when they expressed an explicit preference, in support of H_4a. It is noteworthy that this negative effect on social utility emerged in a realistic setting with actual social interactions.

For robustness, and to examine the underlying process for the effect on liking, we conducted a follow-up study (N = 397) with a similar design using a hypothetical scenario (see the details of the study in Web Appendix Study 4). Replicating the results from Study 4, decision makers who received a no-preference communication again felt greater decision difficulty compared with explicit-preference communication (F(1, 396) = 4.33, p = .038) and liked the co-consumer less when they expressed no preference (F(1, 396) = 44.64, p < .001). Importantly, no-preference communication also evoked greater perception of undisclosed preferences (F(1, 396) = 108.01, p < .001), which significantly mediated the effect of preference communication on both decision difficulty (β = −.94, SE = .13, 95% CI: [−1.222, −.710]) and liking of the co-consumer (β = .40, SE = .08, 95% CI: [.259, .580]), in support of H_2b and H_4b, respectively.

Design, Procedure, and Measures

A total of 165 students at a large North American university (M_age = 19.83 years; 53.9% female) participated in this study. Participants were first asked to rank-order five movie genres (comedy, action, drama, science fiction, and romance) according to their preferences. After completing this ranking, participants were randomly assigned to one of four conditions. In three of the conditions, participants were told that they would be paired with another student to watch a movie clip together. Participants were also told that they would be the one in charge of selecting which movie clip to watch, and that they were allowed to exchange a short message with their partner before making their selection. Participants then wrote a brief message to send to their partner. We implemented this task to make it believable that their partner wrote the communication message they would receive subsequently. Thus, similar to Study 4, all participants in these three conditions took on the role of the decision maker, and we manipulated what the co-consumer expressed in the message they received.

In the no-preference condition, participants read a message ostensibly sent by their partner saying, “I have no preference, it's your call!” In the explicit-similar-preference condition, decision makers read a message stating “I like [highest-ranked option] the best, but it's your call!,” where [highest-ranked option] was filled in with the genre the decision maker had ranked earlier as their most preferred (first out of five options). In the explicit-dissimilar-preference condition, decision makers received a message stating “I like [lowest-ranked option] the best, but it's your call!,” where [lowest-ranked option] was filled in with the genre the decision maker had ranked earlier as their least preferred (fifth out of the five options). In a fourth (alone) condition, participants were not told anything about a partner or a communication message and instead learned that they would watch a movie clip by themselves.

Our dependent variable in this study was participants’ choice of which movie to watch. Participants were asked to choose one out of five movie options (Movies A to E). The five movie choice options that participants received were anchored based on their most-preferred and least-preferred movie genres that they previously indicated. For example, if a participant had indicated that their most-preferred movie genre was comedy and their least-preferred movie genre was action, participants received five movie choice options that varied in the degree to which each movie was related to comedy and action (for an example of the choice options, see Figure 5; for details of the coding procedure, see Web Appendix C).

OPEN IN VIEWER

Figure 5.

Choice Dependent Variable Setup (Study 5).

After the data collection, we later coded Movie A as 1, Movie B as 2, Movie C as 3, Movie D as 4, and Movie E as 5. Thus, a movie choice coded closer to 1 indicates that the decision maker chose a movie that was more similar to their most-preferred genre, whereas a movie choice coded closer to 5 indicates that the decision maker chose a movie that was more similar to their least-preferred genre. After choosing a movie to watch with their partner, all participants actually watched that movie clip and were told that their partner was simultaneously watching the same movie clip (except in the alone condition, where there was no reference to a partner).

Results

A one-way ANOVA revealed a significant effect (F(3, 161) = 18.79, p < .001, η p 2  = .259). As we predicted, participants in the no-preference condition chose movies that they preferred less (M = 3.80, SD = 1.29) compared with participants in the explicit-similar-preference condition (M = 2.57, SD = .93; F(1, 84) = 24.06, p < .001, η p 2  = .223) and the alone condition (M = 2.56, SD = 1.32; F(1, 88) = 19.96, p < .001, η p 2  = .185). That is, although participants received a message saying their partner had no particular preference, they still deviated from their most-preferred topic, suggesting that they did not take the no-preference communication at face value. In fact, participants in the no-preference condition even chose movies that were equally less preferred as those chosen in the explicit-dissimilar-preference condition (M = 4.29, SD = 1.47; F(1, 85) = 2.77, p = .100, η p 2  = .032). Figure 6 displays these results. For the full distributions of movie choices by condition, see Web Appendix C.

OPEN IN VIEWER

Figure 6.

Movie Choice (Study 5).

Discussion

The results of Study 5 demonstrated the impact of no-preference communication on consumption utility: decision makers ultimately chose an option they preferred less, compared with when they were consuming alone (H_5a) and compared with when they received an explicit communication of similar preference (H_5b). Moreover, decision makers who received a no-preference communication actually chose an option as if their partner had explicitly expressed a dissimilar preference (H_5b). These choice patterns support our theorizing that receiving a no-preference communication leads the decision maker to try to accommodate the co-consumer's undisclosed preferences, which they infer to be dissimilar to their own. As a result, decision makers end up shifting away from their own most-preferred option rather than choosing an option that would maximize their own consumption utility. ⁶

Design, Procedure, and Measures

A total of 531 students at a large university (M_age = 23.45 years; 54.0% female) participated in this study remotely. The study was designed to utilize an online trivia game that two people can play together virtually. Similar to the setup of Study 5, participants were first asked to rank-order four trivia topics (finance, politics, technology, and entrepreneurs) according to their own preferences. After completing this ranking task, all participants were informed that they would be playing a virtual game with another student (i.e., their game partner). Specifically, they were told that they would play the trivia game together as a team by collaborating and taking turns narrowing down and ultimately selecting an answer. The game was described as a fun shared experience. Next, to increase realism, participants waited a few seconds while the system ostensibly matched them with another student. Then, they said hello to each other by sending and receiving a greeting message. We implemented this task to make it believable that their partner wrote the communication message that they would receive subsequently.

As in the previous two studies, all participants were assigned to the role of the decision maker. Participants were told that they would be the one in charge of selecting the trivia topic of the game they would play together. However, before choosing a trivia topic for the joint game, they were allowed to exchange another short message with their partner. Participants were randomly assigned to one of three conditions, in which we manipulated what the co-consumer expressed in the message. Specifically, in the no-preference condition, participants received a message stating, “Hey, I don’t have a preference. It's your call!” In the explicit-similar-preference condition, participants received a message stating, “Hey, I like [highest-ranked option] the best, but it's your call!,” where [highest-ranked option] was filled in with the topic the decision maker had ranked earlier as their most preferred. In the explicit-dissimilar-preference condition, decision makers received a message stating “I like [lowest-ranked option] the best, but it's your call!,” where [lowest-ranked option] was filled in with the topic the decision maker had ranked earlier as their least preferred. Thus, in the explicit-similar-preference condition, participants were led to believe that their preferences matched their partner’s. However, in the explicit-dissimilar-preference condition, the participant and their partner's preferences were mismatched.

Results

Discussion

Using a real online joint consumption setting, Study 6 further demonstrated the negative impact of no-preference communication on decision makers, as well as the underlying mechanism leading to these negative consequences. Specifically, no-preference communication triggered perceptions of undisclosed preferences, which led decision makers to choose an option they prefer less, compared with when they received an explicit communication of similar preferences (H_5b–c). The results of Study 6 also provided additional insight into the effect on decision difficulty, which was greater when decision makers received a no-preference communication than when they received a similar-preference communication, but less than when they received a dissimilar-preference communication. Finally, this study revealed another negative consequence of no-preference communication on the decision makers’ consumption utility. Specifically, decision makers who received a no-preference communication enjoyed the joint consumption experience significantly less than those who received a communication expressing explicitly similar preferences. Interestingly, no-preference decision makers reported equal levels of enjoyment as those who received an explicit communication of dissimilar preferences.

We note that while in both Studies 5 and 6, those in the no-preference communication condition chose an option they preferred significantly less compared with those in the explicit-similar-preferences condition, the comparison with the explicit-dissimilar-preferences condition yielded slightly different patterns across studies. In Study 5, participants who received a no-preference message deviated from their most-preferred option as if the message had indicated that their partner had dissimilar preferences. However, in Study 6, this deviation, though significant and noticeable, was not as severe. This difference across the two studies could be a result of different contexts (movie vs. game selection), or it could speak to the notion that the decision maker is trying to balance the two parties’ preferences (which results in them choosing something they prefer less, but not always something completely different from what they prefer). All in all, additional research may be warranted to examine what factors influence the degree to which consumers are willing to deviate from their most-preferred choice during joint decision making.

Contributions and Implications

Our findings make several important contributions. First, although prior research on joint decision making has primarily focused on how the parties resolve preference conflicts to reach a joint decision (e.g., Park 1982; Spiro 1983), it has largely assumed that the parties were already aware of each other's preferences (e.g., Simpson, Griskevicius, and Rothman 2012). However, the outcomes of joint decisions in which one party does not have a preference (or at least communicates not having one) remain relatively understudied. Notably, some recent work documents consumers’ inclination to state having no preference (despite what their consumption partner actually prefers; Liu and Min 2020) as well as the interpersonal costs of being perceived as having no preference (Lopez, Woolley, and McGill 2021). The current article adds to this emerging body of research in several ways. First, we focus on how no-preference expressions impact joint decision making and the consumption experience. In particular, we show that such no-preference communications can have a negative effect on the recipient (i.e., the decision maker) by increasing their experienced decision difficulty and leading them to choose options that they ultimately enjoy less. Our research also reveals the negative effects of such expressions on liking of the person communicating no preference, and it demonstrates how all of these downstream consequences are driven by perceptions that this individual has undisclosed preferences. In addition, our work extends prior work on joint consumption that separately examined decisions jointly made by the dyad and decisions individually made by one partner on behalf of a dyad (e.g., Gorlin and Dhar 2012), by showing that no-preference communication is a distinct case that essentially transforms a joint decision into an individual decision made by the decision maker on behalf of the dyad.

Second, documenting these unexpected negative consequences of no-preference communication in joint decisions on both consumption and social dimensions adds to a recent and growing body of literature on the impact of seemingly benign verbal communication in relationships. Although mundane expressions (e.g., saying “thank you”) may seem to be unimportant, they can in fact significantly impact interpersonal relationships (e.g., Chaudhry and Loewenstein 2019; Lambert and Fincham 2011; Maio et al. 2008; Park et al. 2019) and create tangible business benefits as well (e.g., Abeler et al. 2010).

Third, the current work contributes to the study of consumer doubt and suspicion. In the business literature (e.g., marketing and management), most of the research concerning consumers’ suspicion has focused on firm-to-consumer contexts (e.g., Campbell and Kirmani 2000; Friestad and Wright 1994) and salesperson-to-consumer interactions (e.g., DeCarlo 2005; DePaulo and DePaulo 1989; Tuk et al. 2009). We make a novel contribution to this investigation by examining when messages might not be taken at face value in consumer-to-consumer contexts. To the best of our knowledge, this is the first article in marketing to examine how suspicion among consumers may impact consumption decisions and utilities.

Finally, our findings extend previous work examining the inferences people make about others in the absence of any communication (Mullen et al. 1985; Ross, Greene, and House 1977; Weaver and Hamby 2019; Wu et al. 2021). Whereas simply not knowing others’ preferences may trigger processes that lead to predictions of similar preferences (e.g., the false consensus effect), our work introduces how an active communication of no preference is different because it triggers the perception of undisclosed preferences. To further examine these different contexts, we ran an additional study (N = 218; Web Appendix Study 6) that compares an individual who receives an explicit no-preference communication with an individual who makes a joint decision without receiving any communication from their partner (i.e., a “false consensus” context). In support of our theorizing, the individual who did not receive any communication was significantly less likely to infer that their partner had undisclosed preferences (p = .002) and experienced significantly less decision difficulty (p = .025), compared with the decision maker who received a no-preference communication.

From an applied perspective, consumers’ online communications are more prevalent than ever and can often replace face-to-face social interactions (particularly in light of the COVID-19 pandemic). While communications that occur in-person could potentially include nonverbal cues (e.g., facial expressions, body language) that would help alleviate decision makers' suspicion, online communication tends to be less informative and engaging for consumers (Eelen, Özturan, and Verlegh 2017). Whether through text messages, emails, or social media, consumers’ online communications are often short and quick, making it even more difficult for decision makers to alleviate their suspicion. Moreover, it is often difficult to determine whether word of mouth in this channel is incentivized, which can have important implications for the perceived sincerity and effectiveness of the message (e.g., Barasch, Berman, and Small 2016; Tuk et al. 2009; Verlegh et al. 2013). Thus, the online environment may be a particularly relevant context for further understanding the negative impact of no-preference communication.

Finally, for marketers, much of consumers’ day-to-day joint consumption entails low-involvement contexts, such as choosing which TV show to watch or picking a snack to share with a friend. Our findings suggest that no-preference communications may have broad negative consequences for consumption utility, which can hurt experiences with the consumed product or service. Understanding this potential communication pitfall and its impact could help companies mitigate some of its negative effects. Our findings suggest that marketers could encourage consumers in joint consumption to explicitly communicate their preferences, especially when they know that there is a match in taste. For example, a TV streaming service could offer sharing tools for friends or family members to access each others’ preferences, thus decreasing the likelihood of undisclosed preference perceptions and helping consumers make choices that can best accomodate both parties.

Future Directions

Future research could more broadly examine how the nature of the relationship between the two parties influences our findings. For example, because our theory would predict that decision makers who receive a no-preference communication should be particularly negatively impacted when they care about satisfying the other person's preferences, decision difficulty may be amplified when the communication comes from close friends or family members (Moore, Fitzsimons, and Fitzsimons 2020). Alternatively, relationship closeness could reduce decision makers’ uncertainty about the other person's preferences (i.e., making those preferences seem less “hidden”) and potentially moderate the no-preference effects. To explore this notion, we ran a follow-up study (N = 550; Web Appendix Study 7) that tested three different types of closeness relationships in the role of the co-consumer: (1) a new person that one just met, (2) a friend, and (3) a spouse. The results revealed a robust effect of no preference on decision difficulty (p < .001) and no interaction (p = .559), such that no-preference communication actually increased decision makers’ decision difficulty across all three types of relationships. Although we did not observe moderation by relationship closeness in this preliminary study, we speculate that it might be particularly difficult to eliminate perceptions of undisclosed preference, even for close relationships, in mundane everyday decisions (e.g., choosing a restaurant), because decision makers do not necessarily know the co-consumer's preference (from a variety of desirable options) on that particular occasion. Future work could also investigate the role of relationship closeness when no preference is communicated in high-involvement decisions, as well as how this factor influences consumption and social utilities. For example, it would be interesting to test how ongoing interactions with close others (i.e., in which the decision maker sees how indecisive co-consumers actually react to items chosen for them on one or more occasions), might affect the decision maker's perspective over time.

It may also be fruitful to examine how individual differences could impact our findings. For instance, “selfish” consumers may be less inclined to balance their own and others’ preferences, whereas “altruistic” consumers may be more likely to prioritize maintaining the relationship over maximizing their own consumption experience (Dzhogleva and Lamberton 2014; Lowe et al. 2019). Ironically, our theorizing suggests that altruistic individuals might experience greater costs, either in the role of the co-consumer who expresses no preference and is liked less or in the role of the decision maker who chooses options that lower their consumption utility. It might also be interesting to explore how our documented effects on the decision maker depend on relevant personal factors: higher levels of expertise or stronger preferences on the decision maker's part might make them less susceptible to the negative effects of perceiving undisclosed preferences, whereas a higher need for control might exacerbate their frustration from not knowing what the other party prefers. Further, understanding which consumers are more likely to question the authenticity of messages received from others versus which would be more convinced by the sincerity of no-preference expression would shed more light on how to mitigate consumers’ suspicion.

Another promising avenue for research could be to examine the effects of different phrases that consumers use to communicate having no preference. In Study 1, we found a robust effect across a variety of phrases that consumers report often using in real joint consumption. However, it is possible that other ways of articulating no preference would alleviate the negative effects of this communication, particularly if the phrase added credibility to the fact that the individual was genuinely indifferent among the options, which would reduce the perception of undisclosed preferences (as demonstrated conceptually in Study 3). For example, a co-consumer could preface their no-preference communication about where to go for dinner (e.g., “Let's go where you want”) with “I just had a big lunch, so I probably won’t eat much.” Alternatively, no-preference phrases that are more positive in tone (e.g., “I love all of these options!”) or that help the decision maker reach a decision (e.g., “I don’t care, but I just read a review about restaurant X”) could potentially mitigate the negative effects of this communication strategy on decision difficulty, liking, and choice.

We further suggest that no-preference communication can engender a myriad of social consequences beyond the ones we document in this work. In a supplemental study (Web Appendix Study 8), we document some additional inferences people make about co-consumers who express no preference in joint decisions (e.g., they are perceived to be significantly more annoying [p = .003] and less helpful [p = .001]). That being said, while we demonstrate mostly negative consequences of no-preference communication in our article, we acknowledge that this communication strategy could simultaneously have positive effects on other interpersonal outcomes (e.g., inferences about the co-consumer's cooperativeness, increased trust in the relationship; Levine and Schweitzer 2015), particularly in contexts where the decision maker focuses on the co-consumer's potentially good intentions (i.e., their desire to accommodate the decision maker's preferences). Future work should examine these interesting possibilities, especially how different lay theories and attributions for the behavior (e.g., personal vs. situational) might influence these outcomes.

Finally, and more broadly, we highlight the need for more work on how consumers communicate their preferences to each other in joint decision contexts. Whereas the current research focused on one common expression, other preference-signaling communications or behaviors may be important and worthwhile to examine.