Communication Efficacy as a Mechanism for the Chilling Effect on Complaint Avoidance A Cross-cultural Comparison of American and Chinese Romantic Relationships

Complaint avoidance is a common response to relational conﬂict; yet, it can cause adverse personal and relational consequences. This study examined cultural (i.e., U.S. versus China) and relational (i.e., relational power, communication efﬁcacy) factors predicting complaint avoidance. Hypothesizing that Americans engage in less complaint avoidance than Chinese (H1), relational power negatively predicts complaint avoidance via communication efﬁcacy across cultures (H2), and culture moderates how power and efﬁcacy predict avoidance (H3), we surveyed 392 college-aged dating individuals (194 Americans, 198 Chinese). Results conﬁrmed H1 and H2. H3 was partially supported: after controlling for relational and conﬂict characteristics, as communication efﬁcacy increased, there were greater decreases in complaint avoidance in Americans than in Chinese. Implications for cross-cultural studies on relational conﬂict are discussed.


Introduction
Conflict is common in romantic relationships because prolonged contact and frequent interactions make couples vulnerable to disagreement over various goals and behaviors.How couples manage conflict is an essential part of their overall relationship history (Peterson 2002), relational satisfaction, and each partner's well-being (Greeff & De Bruyne 2000;Lefcourt 1982).Among many conflict management approaches, avoiding discussion about problematic issues is prevalent.This may be because avoidance is easier and less risky than direct confrontation (Afifi & Olson 2005) and can sometimes facilitate relational harmony (Baxter & Dindia 1990;Rusbult, Verette, Whitney, Slovik, & Lipkus 1991).However, withholding conversations about important relational issues has been associated with relationship dissatisfaction (Merrill & Afifi 2012), increased psychological stress (Slatcher, Robles, Repetti, & Fellows 2010), and negative rumination about relational problems (Cloven & Roloff 1991).Moreover, relational partners lose opportunities for positive relational change when avoiding communication about concerns with one another (Cloven & Roloff 1991).
Considering the adverse influence of complaint avoidance on individual and relational outcomes, an examination of factors predicting complaint avoidance may help us better understand this phenomenon and alleviate its negative effects.Conflict face-negotiation theory (Ting-Toomey 1985; 2017) contends that cultural and relational factors shape communicators' conflict management.Specifically, cultural individualism-collectivism is assumed to exert direct effects on people's preferred conflict strategies (see Ting-Toomey 2017 for a review).People in individualistic cultures tend to use more dominating strategies (e.g., being aggressive, expressing an opinion, and defending a position), whereas people in collectivistic cultures prefer to avoid conflict (Ting-Toomey 2017).Cross-cultural research has considered and continues to identify the U.S. culture as individualistic and Chinese culture as collectivistic, even with the dramatic socioeconomic and political transformations in mainland China since the early 1990s (Gao 2001;Hu, Bernardo, Lam, & Cheang 2018).Thus, this research explores cultural influence on complaint avoidance by comparing romantic relationships in the U.S. and China.
Of course, culture is not a sole determinant of how people manage relational conflict.Relational factors also elucidate why individuals withhold confrontation (Ting-Toomey 1985;2017).This study focuses on relational power, because power is a fundamental relational construct that influences couples' general communication and specific conflict behaviors (Dunbar 2004;Rollins & Bahr 1976).Prior studies showed that powerless individuals often avoid confronting their powerful partners about relational issues due to fear of negative consequences, such as powerful partners' physical and symbolic aggression (e.g., Cloven & Roloff 1993) or termination of relationships (e.g., Rusbult et al. 1991).This phenomenon is known as the chilling effect (Roloff & Cloven 1990).
Although fear of negative outcomes may be one mechanism for the chilling effect, it still does not fully explain why individuals are unwilling to confront relational issues.Afifi, Olson, and Armstrong (2005) stated that when it comes to sensitive and difficult conversations, individuals may lack confidence or feel unable to communicate with others, inhibiting actual actions.Makoul and Roloff (1998) found that individuals' confidence in their ability to confront others determines whether a confrontation actually takes place.Thus, communication efficacy, or an individual's appraisal of his or her ability to successfully engage in communicative behaviors to achieve interpersonal goals (Afifi & Weiner 2004), may be another factor explaining the chilling effect.However, although the negative relationship between communication efficacy and complaint avoidance has been documented (e.g., Worley & Samp 2016), prior research has not examined how relational power predicts people's communication efficacy.Thus, the current study investigates how relational power affects complaint avoidance via communication efficacy.
Although either cultural or relational elements may exert a direct impact on couples' conflict management, culture as a contextual factor may also moderate the way that relational power affects complaint avoidance.Specifically, the chilling effect shows that when people perceive more power over their partners, they will be more likely to confront their partners about relational issues.Yet, when compared to over-powered American individuals, powerful Chinese individuals may be less likely to confront their powerless partners.This may be because, even though they feel they can bring up the issues, powerful Chinese individuals may still prefer to withhold confrontation to save their partners' faces and maintain a harmonious relationship.In other words, culture may moderate the relationship between relational factors and complaint avoidance.However, limited studies have examined this prediction; we know little about how culture influences the process of the chilling effect.To address this gap, we examine how the chilling effect of complaint avoidance differs between American and Chinese relationships.
In the following sections, we first review prior studies on the direct impact of culture on conflict management.Then, we define dependence power and illustrate how power may influence complaint avoidance via communication efficacy.Finally, we explore the way that cultural and relational factors interact to predict couples' conflict management.The hypotheses of the study are formulated as we review previous work and define concepts.

Cultural Influence on Complaint Avoidance
Conflict face-negotiation theory (FNT; Ting-Toomey 1985; 2017) poses that cultural variability influences cultural members' preference for self-face concerns, other-face concerns, or mutual face concerns.Specifically, people in individualistic cultures, such as the United States, are often assumed to have a higher self-face concern.A concern for self-face is associated with more dominating conflict management strategies, such as acting aggressively, defending a position, and expressing an opinion.In contrast, people in collectivistic cultures, such as China, have more other-face and mutual-face concerns as well as value interpersonal harmony; thus, they tend to avoid a conflict by giving in to other's position or pretending that the conflict does not exist (Oetzel et al. 2001;Oetzel & Ting-Toomey 2003;Ohbuchi, Fukushima, & Tedeschi 1999;Ting-Toomey, Oetzel, & Yee-Jung 2001).When romantic couples dealing with emotional transgressions, Zhang, Ting-Toomey, Dorjee and Lee (2012) found that American individuals tend to overtly express their emotions and that Chinese individuals tend to avoid direct confrontation with their partners.Thus, prior research has provided strong evidence that Chinese individuals are more likely to avoid relational conflict than their American counterparts.We predict that: H1.American individuals engage in less complaint avoidance than Chinese individuals.

Relational Factors Predicting Complaint Avoidance
Culture may influence people's preference for certain conflict strategies; yet, characteristics of the particular close relationship in which interpersonal problems arise often exerts a more proximal impact on how partners manage conflict (Ting-Toomey 2017).Interpersonal conflict research (e.g., Dunbar 2004;Rollins & Bahr 1976;Roloff & Cloven 1990) has indicated that a person's perceived ability to influence a relational partner's decisions and outcomes, or relational power, plays a crucial role in conflict communication.Thus, this study examines how relational power predicts individuals' complaint avoidance.
In intimate relationships, a major source of relational power is dependence power, which reflects the control that an individual who is less dependent on the relationship for valuable resources accrues in the eye of his or her more dependent partner (Lawler & Bacharach 1987).Specifically, dependence power is a composite judgment based on three relational perceptions, including an individual's perceived commitment to the relationship, the perceived commitment of a partner to the relationship, and the perceived alternatives that a partner has to the relationship.When an individual is less committed to the relationship, has a more committed partner, and has equivalent or better alternatives to current romantic partners, that individual accrues dependence power over his or her more dependent partner.In contrast, a person perceives less dependence power if he or she is highly committed to a relationship but has a less committed partner who has significant relational alternatives (Cloven & Roloff 1993).
The chilling effect summarizes the role of dependence power in shaping people's conflict management, such that perceived power disadvantage tends to quell people's expression of interpersonal complaints (Roloff & Cloven 1990).Social exchange theory (Emerson 1976) indicates that when a person is more committed and believes there are only poor alternatives to the current relationship, he or she will likely perceive the costs of leaving the current ones as higher than the potential benefits of other relationships and thus, want to stay in the relationship.Fear of losing the current relationship, in turn, makes people endure a powerful partner's costly behaviors (Rusbult & Martz 1995;Rusbult et al. 1991).Moreover, the worry that powerful partners may respond aggressively also inhibits direct confrontation (Cloven & Roloff 1993;Jewkes 2002).Thus, fear of powerful partners' negative consequences has been identified as the primary reason for the chilling effect.However, even though fear of negative outcomes is likely to inhibit complaint expressions, it still does not fully explain why people are unwilling to confront their partners.Makoul and Roloff (1998) showed that people's confidence in their ability to confront others determines whether a confrontation actually occurs.Afifi et al. (2005) found that when individuals feel unable to communicate a secret with a potentially aggressive family member, they tend to avoid self-disclosure and continue to conceal the secret.Similarly, when confronted with a powerful partner who may abandon the relationship or respond negatively, powerless individuals may consider if they are able to bring up the issues with their partners.Thus, communication efficacy, or people's perceived ability to successfully achieve interpersonal goals via communication, may be another mechanism for the chilling effect.Indeed, Worley and Samp (2016) found that lack of communication efficacy increases complaint avoidance.
People do not develop communication efficacy in a vacuum; rather, their assessment of the current relationship is likely to impact their efficacy.According to Bandura (1977), people's perceived abilities to perform certain behaviors is grounded in their anticipated outcomes of those behaviors.Williams (2010) also showed that if people predict more likelihood of successes of a communicative behavior, they will feel more confident to enact that behavior.Because dependence power entitles powerful individuals to control their powerless partner' outcomes and decisions, powerful people may anticipate more positive outcomes via direct confrontation.In other words, powerful individuals may expect that their desires and wishes will be satisfied via complaint expressions and thus, feel more confident to voice their opinions.Indirectly supporting this reasoning, Falbo and Peplau (1980) observed that over-powered individuals tend to utilize more direct influence strategies, such as reasoning and persuading; whereas under-powered partners prefer more indirect influence strategies like withdrawal.Thus, power advantage seems to boost people's communication efficacy.Taken together, we predict: H2.Communication efficacy mediates the negative relationship between dependence power and complaint avoidance, such that dependence power is positively associated with communication efficacy, which in turn negatively predicts complaint avoidance.

Cultural Influences on the Chilling Effect
Although either cultural or relational factors can directly influence couples' compliant avoidance, couples' cultural background may guide the way that relational factors affect conflict management.Specifically, the chilling effect shows that powerful individuals are less likely to withhold their confrontation about relational issues.However, when compared to people in individualistic cultures, powerful collectivists may still engage in some sort of complaint avoidance in dealing with relational conflict.Gudykunst and colleagues (1996) indicated that cultural individualism-collectivism influences people's communication styles, which are especially apparent in linguistic expressions.Individualists tend to communicate directly, whereas collectivists tend to communicate indirectly as they rely more on the shared knowledge to produce and interpret messages.Because collectivists are more likely to recognize the existence of power discrepancy and to accept powerful persons getting their way (Bochner & Hesketh 1994;Holtgraves & Yang 1992), powerful individuals in collectivistic cultures may feel there is no need to overtly express their opinions (see Dunbar 2004;Huston 1983).In other words, while powerful collectivists can confront their powerless partners, direct expressions seem unnecessary, in that their confidence and ability to successfully influence powerless partners is the shared knowledge and implicitly present.
American and Chinese couples also view their close relationships differently, which may in turn influence how relational factors predicts complaint avoidance.For instance, Chinese couples tend to report a higher level of commitment and loyalty in their romantic relationship than their U.S. counterparts (Gao 2001).Also, Chinese couples report greater consensus on emphasizing couple unity, maintaining relationship harmony, and investing in the relationships than U.S. couples (Epstein, Chen, & Beyder-Kamjou 2005).Compared to their powerful U.S. counterpart, powerful Chinese individuals may be more willing to avoid direct confrontation, because they perceive complaint avoidance as a means to demonstrate their commitment to the relationship, protect powerless partners' faces, and foster relationship well-being.In short, although a power advantage can encourage open expressions of conflict, Chinese individuals may be more hanxu, or reserved, in using their power.Similarly, while higher communication efficacy can motivate open discussions about relational problems, collectivists may reserve their abilities to actually confront their partners.Therefore, we predict: H3a.Culture moderates the negative relationship between dependence power and complaint avoidance, such that as dependence power increases, there will be greater decreases in complaint avoidance in Americans than in Chinese.
H3b. Culture moderates the negative relationship between communication efficacy and complaint avoidance, such that as communication efficacy increases, there will be greater decreases in complaint avoidance in Americans than in Chinese.
In contrast to valuing open expressions in the U.S. culture, direct expressions of feelings and thoughts are often discouraged in Chinese society (Huang 1981).Chinese people are socialized to control displays of emotions (Wu 1996), use implicit verbal and nonverbal messages to remain reserved, avoid showing-off, and to foster interpersonal harmony (Gao, Ting-Toomey, & Gudykunst 1996).Thus, cultural collectivism may lower people's confidence that they can confront their partners about relational issues.Yet, no studies have examined how culture may influence the way that relational power predicts communication efficacy.Thus, we ask the following research question.
RQ. Does the positive association between relational power and communication efficacy differ in Americans and Chinese?

Participants
Data were collected from young adults who were currently in an unmarried romantic relationship in the United States and China.The U.S. participants were recruited from a research subject pool at a large, southeastern public university.The university's Institutional Review Board (IRB) approved this study.The Chinese respondents were recruited from communication classes at a large public university in Southwestern China.The questionnaires were translated by the first author into Chinese with the English version attached to ensure the integrity of the translation.The university's Division of Student Affairs, which serves as the IRB at the university, approved this study.

Procedures
After providing informed consent, participants were instructed to briefly write a conflict scenario that they encountered recently with their partners.A conflict was defined as "a disagreement between partners and a situation when one partner's behavior negatively influences another's ability to achieve his or her goals."Participants were instructed that the conflict could be issues that both partners were aware of or that only one partner knew.This task helped promote reflection on relational conflict and was followed by questions pertaining to participants' perceptions of their relational power, cognitive appraisals of the problem, complaint avoidance and demographic information.U.S. students received a small amount of extra course credit upon completion of an online questionnaire.Chinese students completed paper-andpencil surveys and were not compensated.

Measures
Except for the measure of relationship length, this study utilized a series of 7-point Likert-type scales (1 = strongly disagree, 7 = strongly agree) to assess the study variables.Unless otherwise specified, items were reverse-coded (if applicable) and averaged, with higher numbers indicating higher levels of the variable.Table 1 summarizes the mean and standard deviation for each variable.(Cloven & Roloff 1993), we adapted the formative index developed by Samp and Abbott (2011) to operationalize dependence power.The index was based upon three 7-point Likert-type scales reported in Solomon and Samp (1998).Four items assessed an individual's commitment to his or her relationship (e.g., "I would like this relationship to last a lifetime.").This scale had good reliability among the full, Chinese, and American samples, Cronbach's α Full = .83,α U.S. = .84,α China = .82.Perceived partners' commitment to the relationship was examined by another four items (e.g., "My partner is committed to our relationship," α Full = .83,α U.S. = .83,α China = .80).Perceived partner's access to relational alternatives was also measured via four items (e.g., "If our relationship were to end, it would be easy for my partner to find a new dating relationship that was good or better," α Full = .78,α U.S. = .80,α China = .74).
To compute the index, the measures for individuals' commitment and partner relational alternatives were first reverse-coded.The three measures were then summed to obtain individuals' dependence power scores, with higher numbers indicating that the persons felt less committed, considered having a more committed partner, and perceived better relationship alternatives (i.e., greater dependence power) (M Full = 13.10,SD Full = 1.63;M U.S. = 13.43,SD U.S .= 1.35;M China = 12.77, SD China = 1.82).To remain consistent with prior research using the formative index of dependence power (e.g., Samp & Abbott 2011), the three scales were summed, rather than averaged.

Communication efficacy.
We assessed communication efficacy based on a 7-point scale established by Afifi et al. (2005).Participants indicated their level of agreement with four statements regarding their ability to communicate the conflict they specified to partners (e.g., "I don't know what to say when I try to communicate about this conflict to my partner," reverse-scored).This scale had high reliability (α Full = .93,α U.S. = .93,α China = .92)(M Full = 4.85, SD Full = 1.63;M U.S. = 5.36, SD U.S .= 1.49;M China = 4.35, SD China = 1.16).

Complaint avoidance.
A measure developed by Worley and Samp (2016) was utilized to assess complaint avoidance.Participants responded to the following prompt, "When communicating the conflict with my partner," using a 7-point Likert-type scale that included four statements.A sample statement was "I 'hold my tongue' rather than telling my partner what I really think about the problematic issue."The complaint avoidance scale had good reliability (α Full = .89,α U.S. = .91,α China = .86)(M Full = 2.97, SD Full = 1.14;M U.S. = 2.67, SD U.S. .= 1.18;M China = 3.26, SD China = 1.01).

Problem severity.
Participants were asked to write their own conflict scenarios, of which the perceived severity varied.Solomon and Samp (1998) indicated that appraisals of problem severity might constitute another mechanism underlying the chilling effect, such that powerless partners tend to perceive a problem as trivial, which in turn increases complaint avoidance (Solomon & Samp 1998).Thus, to control the influence of problem severity on complaint avoidance, participants were instructed to indicate how severe the problems were after the writing task.Problem severity was measured on a 7-point scale adapted from Samp and Solomon (1999).There were six items (e.g., "This issue is very serious in my relationship.").This scale had good reliability (α Full = .89,α U.S. = .86,α China = .90)(M Full = 3.30, SD Full = 1.45;M U.S. = 2.97, SD U.S .= 1.38;M China = 3.61, SD China = 1.45).

Relationship length.
Theoretical frameworks, such as social penetration theory (Altman & Taylor 1973) and the relational turbulence model (Solomon & Knobloch 2004), indicate that relationship length may influence people's perceptions and performance of communication with their partners.

Preliminary Analyses
Variable means were compared between the United States and the Chinese sample.A Chi-Square test of independence examined whether gender distribution differed between the two groups.No significant differences were observed, χ 2 (1, 392) = .15,p = .696.Independent sample t-tests compared variable means between U.S. and Chinses participants.Results (see Table 1) showed that the U.S. sample was younger than the Chinese sample.Also, U.S. participants reported higher dependence power, higher communication efficacy, lower problem severity, and lower complaint avoidance than Chinese respondents.Relationship length, however, was not different between the two groups, t(390) = 1.19, p = .238.Overall, t-tests supported H1 that Americans engaged in less complaint avoidance than Chinese participants.
Next, variable means were compared between the male and female participants.For the combined sample, male participants were slightly older (M = 20.77,SD = 2.46) than female respondents (M = 20.12,SD = 1.46), t(390) = 3.23, p = .001.No other mean differences between men and women were found.For the U.S. sample, male participants perceived their conflict scenarios to be less sever (M = 2.74, SD = 1.30) than female respondents (M = 3.17, SD = 1.42), t(192) = -2.20,p = .029.For the Chinese sample, men were slightly older (M = 20.04,SD = 1.61) than women (M = 19.64,SD = 1.22), t(196) = 2.89, p = .004.No other mean differences were observed for either the U.S. or the Chinese sample.One-way ANOVA also showed that variable means did not differ as a function of participants' sexual orientation and ethnicity for both the combined and separate samples.

Multiple Regression Analyses
As an extension of simple linear regression, multiple regression is used to predict the value of one dependent variable based on the values of multiple independent variables and the interactions among those independent variables.Multiple regression also allows researchers to determine the overall fit (variance explained) of the regression model and the relative contribution of each of the predictors (i.e., independent and interaction variables) to the total variance explained (Cohen, 2003).Thus, we utilized multiple regression to test our hypotheses and explore our research question.
Participants reported their complaint avoidance based on self-reported conflict scenarios that varied in problem severity.Also, the duration of their relationships was different, which may affect perceptions about their relationships communication.Thus, we controlled problem severity and relationship length in all regression analyses.A participant's current country was a binary variable (0 = U.S., 1 = China).Other demographic variables such as age and gender were not significant covariates in this study.H2 predicted that communication efficacy mediates the negative relationship between dependence power and complaint avoidance.We followed Baron and Kenny's (1986) steps for mediation analysis.First, complaint avoidance was regressed on dependence power (Model 2 in Table 2).Dependence power was negatively associated with complaint avoidance, β = -.20,p < .001.Second, communication efficacy was regressed on dependence power (Model 1 in Table 2).Dependence power positively predicted communication efficacy, β = .20,p < .001.Third, complaint avoidance was regressed on dependence power and communication efficacy (Model 3 in Table 2).Communication efficacy negatively predicted complaint avoidance, β = -.35,p < .001.Dependence power also negatively predicted complaint avoidance, β = -.13,p = .007,but the absolute value of the regression coefficient was smaller than that of the model without communication efficacy, of which the β is -.20.Thus, for the combined sample (N = 392), controlling for relationship length, problem severity, and country, communication efficacy partially mediated the negative relationship between dependence power and complaint avoidance.H2 was supported.
In addition, after controlling for relational characteristics (e.g., relationship length, dependence power) and conflict features (e.g., problem efficacy) (see Model 2 and 3), country (0 = U.S., 1 = China) positively predicted complaint avoidance.The result provided additional support for H1 that U.S. individuals engaged in less complaint avoidance than Chinese people.
H3 predicted that as relational power (a) and communication efficacy (b) increase, there will be greater decreases in complaint avoidance in American participants than in Chinese respondents.We followed Dawson's (2014) approach to examine whether country (0 = U.S., 1 = China) was a significant moderator of the power-avoidance relationship (H3a) and efficacy-avoidance association (H3b).First, to examine that country interacted with dependence power to predict complaint avoidance, an interaction term between power and country was computed and entered in step 3 of Model 2 (see Table 2).Entering the interaction term did not significantly increase the variance explained by the model, ΔR 2 = .00,p = .931;and the interaction effect was not significant, β = -.03,p = .931.Thus, the negative relationship between dependence power and complaint avoidance did not differ between the U.S. and the Chines sample.H3a was not supported.
To answer RQ that asked whether the positive association between dependence power and communication efficacy differ across cultures, another regression model examined if country interacted with power to predict efficacy.The interaction between dependence power and country was computed and entered in step 3 of Model 1 (see Table 2).Entering the interaction term did not significantly increase the variance explained by the model, ΔR 2 = .00,p = .328;and the interaction effect was not significant, β = .32,p = .328.Thus, the positive relationship between dependence power and communication efficacy did not differ across the two cultures.
A final regression model investigated whether country interacted with communication efficacy to predict complaint avoidance.Control variables were entered in step 1, country, dependence power, and communication efficacy were entered in step 2, and an interaction term between country and communication efficacy was computed and entered in step 3 (Model 3 in Table 2).Entering the interaction term significantly increased the variance explained by the model, ΔR 2 = .03,p < .001;and the interaction effect was significant, β = .56,p < .001.Thus, communication efficacy and country interacted to predict complaint avoidance, suggesting potential cultural influences on the chilling effect.
To further explain the cultural difference, we split the sample by countries and conducted two regression models to examine the mediation effects across cultures.First, when communication efficacy was regressed on dependence power, positive relationships were observed for both the U.S. (B = .15,SE = .07,β = .14,p = .021)and the Chinese (B = .22,SE = .05,β = .25,p < .001)sample (this model is not presented in Table 2).In the second regression model that predicts complaint avoidance, control variables were entered in step 1, dependence power was entered in step 2, and communication efficacy was entered in step 3.
For the U.S. sample (Model 4.A in Table 2), dependence power negatively predicted complaint avoidance in step 2, β = -.15,p = .026.However, after adding communication efficacy in the model, the prior negative relationship between power and avoidance became non-significant, β = -.07,p = .235.Communication efficacy, however, was negatively associated with complaint avoidance, β = -.51,p < .001.Thus, for the U.S. sample, communication efficacy fully mediated the negative relationship between dependence power and complaint avoidance.

Post Hoc Analysis
Using Mplus 7, we tested the indirect effect from dependence power to complaint avoidance through communication efficacy.For the combined sample, the standardized indirect effect was -.07, with a 95% confidence interval (CI) ranging from -.132 to -.054, p <.001.For the U.S. sample, the standardized indirect effect was -.07, 95% CI ranged from -.143 to -.016, p = .015.The direct effect of dependence power on complaint avoidance was not significant, β = -.07,p = .228.For the Chinese sample, the standardized indirect effect was -.03, 95% CI ranged from -.097 to -.001, p = .047.The direct effect of dependence power on complaint avoidance was significant, β = -.23,p = .002.
In short, U.S. individuals reported less complaint avoidance than their Chinese counterparts (H1).
Communication efficacy fully mediated the negative relationship between dependence power and complaint avoidance for the U.S. sample, and partially mediated the relationship for the full and Chinese sample (H2).Moreover, while the relationships between power and efficacy (RQ) and between power and avoidance (H3a) did not differ across cultures (H3a), the negative association between communication efficacy and complaint avoidance was stronger in U.S. participants than in the Chinese respondents (H3b).

Discussion
Complaint avoidance is a common response to the relational conflict between romantic partners.While avoidance may be easier, less risky (Afifi & Olson 2005) and help to maintain relational harmony (Rusbult et al. 1991), withholding complaints can result in adverse individual and relational outcomes, such as higher psychological stress (Slatcher et al. 2010), more negative thoughts (Cloven & Roloff 1991), and less relationship satisfaction (Merrill & Afifi 2012).This study attempted to better understand factors predicting complaint avoidance.Specifically, we examined cultural influences on complaint avoidance and found that, after controlling for relational and conflict characteristics, Chinese participants reported more complaint avoidance than U.S. respondents.In addition, we investigated how relational power predicted complaint avoidance via communication efficacy.Results showed a partial mediation for the full and Chinese sample, and a full mediation for the U.S. sample.Finally, the cultural impact on the chilling effect was also explored.As communication efficacy increased, U.S. individuals were less likely to withhold complaints than their Chinese counterparts, indicating that culture indeed moderates relational factors' effects on conflict management.
Aligning with past research, we observed that culture influences people's perceptions of their relationships and conflict management.For instance, U.S. participants perceived themselves having greater dependence power in their relationships than their Chinese counterparts.Prior studies (Gao 2001;Lin & Rusbult 1995) found that Americans are less committed to their dating relationships than Chinese.Less commitment to a relationship grants people dependence power (Cloven & Roloff 1993); thus, U.S. participants' lower commitment relative to their Chinese counterpart may explain why they perceived greater power in the current study.In addition, the U.S. sample perceived greater communication efficacy and engaged in less complaint avoidance than the Chinese, even after controlling for relational characteristics (e.g., relational length, dependence power) and problem severity (see Model 1 and 2 in Table 2).The principle of being hanxu, or reserved, in Chinese culture and the lack of training on overt verbal and nonverbal expressions (Epstein et al. 2005) may contribute to Chinese participants' lower communication efficacy.Moreover, people in collectivistic cultures have more other-face and mutual-face concerns and thus, tend to favor conflict avoidance to maintain relational harmony (Ting-Toomey 2017).
In addition to its direct effects on conflict management, culture also influences how relational factors impact complaint avoidance.Specifically, after controlling for dependence power and appraisals of problem severity, as communication efficacy increased, Americans were more likely to confront their partners about relational issues than the Chinese.In addition, communication efficacy fully mediated the relationship between dependence power and complaint avoidance in the American sample, whereas it partially mediated the association in the Chinese sample.As Americans place more value on open and direct communication in their relationships (Adams, Anderson, & Adonu 2004) and strive to achieve personal goals in conflict situations (Ohbuchi et al. 1999), the results suggest that as long as they believe they can talk about a problematic issue, U.S. individualists will likely bring it up.In contrast, individuals in collectivist cultures emphasize relational harmony; thus, even when they have the ability to confront a partner, they tend to withhold direct expressions relative to individualists.In other words, while communication efficacy may be the determinant factor influencing Americans' conflict management, other considerations, such as relationship maintenance and face protection, may also significantly affect Chinese couples' complaint avoidance.
This research should be interpreted with several limitations.First, we did not collect data on participants' fear of consequences of confrontation.As fear of negative outcomes is an assumed primary mechanism for the chilling effect; a more thorough examination should have included both communication efficacy and fear of consequences in the model.Moreover, this study primarily focused on young adults' emotional dependence on their partners.For older individuals, other forms of dependence, especially financial dependence, may play crucial roles in complaint avoidance.In addition, we did not account for the potential influences of westernization on younger generations of Chinese and did not measure dimensions of cultural ideology, such as individualism versus collectivism.It is a widespread consensus in cross-cultural studies that the U.S. is an individualistic culture and China is collectivistic culture; thus researchers often directly compare U.S. with China to examine cultural differences on people's communication and relationships (e.g., Epstein et al. 2005;Zhang et al. 2012).But a direct measure of cultural ideology will help control the possible impacts of westernization and globalization on how young adults view romantic relationships and communicate about interpersonal conflict.Finally, this study focused on the romantic relationship of young college students, which limits the generalizability of the results.Data were cross-sectional, which precludes the claim of causality.Future investigations should explore the degree to which these results are replicable across other close relationships and cultures.
In summary, this study further supported the influence of cultural expectations on romantic couples' conflict management.We also identified the role of communication efficacy in the chilling effect.In the U.S. sample, communication efficacy fully mediated the negative relationship between relational power and complaint avoidance; whereas in the Chinese sample, communication efficacy is a partial mediator.
The results indicate that while the phenomenon of the chilling effect is consistent across cultures, the underlying mechanisms may be different.Moreover, this study found that culture also exerts its influences on how relational factors impact complaint avoidance, such that as communication efficacy increase, complaint avoidance decreases in a faster rate in U.S. participants than in Chinses respondents.This result raises the concern that most current studies on relational conflict are conducted in North America and thus, the tenet of theories and results of empirical research may not apply to other cultural contexts.Based on the results of this study, we call for more research on how cultural and relational factors interact to influence expectations about communication and the management of conflict in close relationships.

Table 1 :
Descriptive Statistics and t-tests for U.S. and Chinese Sample.Note.The final column of the table reports the independent sample t-tests for each variable with 390 degrees of freedom.Consistent with the conceptualization of dependence power as a multidimensional construct composing of individuals' commitment, partners' commitment, and partners' alternatives *** p < .001.3.3.1.Dependence power.
Thus, we included relationship length as another covariate.Participants responded to a single-item question, "In months, how long have you been together with your partner?" (M Full = 17.82,SD Full = 15.61;M U.S. = 18.76,SD U.S. .= 15.97;M China = 16.90,SD China = 15.24).