Perfectionism studies have been conducted in the past three decades showed maladaptive and were linked with a variety of adjustment problems, including psychological distress and suicide ideation (Flett et al., 2023; Hewitt et al., 1997; Hill & Curran, 2016; Hill et al., 1997), multiple mental health disorders, including anxiety and depression (Bieling et al., 2004; Hewitt et al., 1996; Mackinnon et al., 2014; Nepon et al., 2016); and eating disorder (Costra et al., 2004). Perfectionist self-presentation is conceptualized as a stylistic dispositional trait, as articulated by Buss and Finn (1987), and is intricately connected to broader trait perfectionism dimensions, particularly self-oriented, other-oriented (demanding others to be perfect) or socially prescribed perfectionism (feeling that people expect them to be perfect) (Flett et al., 1991; Hewitt & Flett, 1991). It reflects a strong need to appear flawless in the eyes of others, emphasizing the expression of one’s presumed perfection to others while rigorously avoiding the public display or disclosure of any perceived shortcomings or imperfections in public (Hewitt et al., 2003; Sherry et al., 2007), and elevating the rumination about interpersonal events (Nepon et al., 2011). It constitutes a pervasive and stable interpersonal orientation and encompasses three distinct, albeit highly correlated facets. The first, perfectionistic self-promotion (PSP), entails actively promoting a public image of perfection and an expression of perfectionist behaviour to others with flawless abilities and engaging in overt displays of perfectionistic behaviour to elicit admiration and respect. Individuals high in PSP frequently accentuate their strengths, successes, and accomplishments to others, purposefully proclaiming these to others to reinforce an image of exceptional competence. The second one is nondisplay of imperfection, reflects efforts by perfectionistic individuals to conceal and avoid displaying imperfection in order to decrease the likelihood of criticism or disapproval. The third facet, nondisclosure of imperfection in public, similarly represents an avoidant interpersonal style, involving the concealment of personal shortcomings or perceived deficiencies in public contexts (Hewitt, et al., 2003). The latter two facets represent primarily protective or defensive orientations aimed at concealing any imperfections. Collectively, these three dimensions function to minimize the visibility of mistakes, preserve an immaculate public persona, and maintain an idealized self-presentation, while simultaneously avoiding any verbal acknowledgment of errors in interpersonal contexts. Individuals exhibiting these tendencies often harbour a pronounced fear of interpersonal rejection or negative evaluation from others. The nondisplay of imperfection facet involves an avoidance of behavioral displays of imperfection, whereas the nondisclosure of imperfection facet entails a reluctance to verbally admit personal shortcomings (Hewitt, et al., 2003). Previous research (Sherry et al., 2007) has demonstrated that perceiving perfectionistic demands and expectations from others, striving to project an impeccably perfect image, concealing perceived flaws are all positively associated with a wide range of personality psychopathologies. Hewitt et al. (2003) further reported that perfectionistic self-presentation is linked to substantial personal and interpersonal distress. The findings revealed that each facet of perfectionistic self-presentation demonstrated unique associations with various maladaptive psychological and relational outcomes (Mackinnon & Sherry, 2012). Perfectionistic self-presentation was also associated with lower social forms of self-esteem, and excessive concerns with others’ evaluation (Hewitt et al., 2003). The expression of perfection clearly involves lower self-esteem and symptoms of anxiety (MacKinnon et al., 2014), and depression (Besser et al., 2010; Hewitt et al., 2003; McGee et al., 2005). Recent study (Agu & Mmaeml, 2023) revealed that three facets of perfectionistic self-presentation positively predicted anxiety sensitivity on 276 undergraduate students. Past studies showed that people strive for absolute perfection and try to hide their mistakes often feel inadequate and ashamed (Chen et al., 2015; Ferreira et al., 2015; 2018; Flett et al., 2023; Oliveira et al., 2017). Shame is a painful self-conscious and devastating self-conscious human emotion characterized by the negative perception that others see the devaluation of the self as inferior, inadequate or unattractive (Gibert, 2002, 2003; Ho, Fu & Ng., 2004; Tangney & Dearing, 2002). Shame involves a negative introspection and evaluation of self as a failure (Fischer & Tangney, 1995; Tangney, 1991, 1994). It is accompanied with stronger sense of worthlessness and powerlessness and blame themselves harshly after wrongdoings and have very negative self-evaluations (Pinto-Gouveia & Matos, 2011). People in shame become more concerned with hiding his/her self from others and used defensive strategies such as rage, contempt, striving for perfection, and denial to minimize these encounters (Kaufman, 1996). Tangney (2002) regarded perfectionists as strict self-evaluators who broaden the range of outcomes that would be perceived as a failure, which leads perfectionists to feelings of shame. Fedewa et al. (2005) found that negative perfectionism correlated positively with state-shame, and shame-proneness. Positive perfectionism negatively correlated to state shame and anxiety. Many researches convinced that shame has connection with psychopathology (Tangney et al., 1995; Tripp & Petrie, 2001). It plays a central role in many mental disorders, such as depression (Kim et al., 2010; Thomas, 2000), anxiety (Fergus et al.,2014; Zhong, 1993; Zhong et al., 2008), eating disorders (Ferreira et al., 2015; 2018; Hewitt et al., 1995; Oliveira et al., 2017). Shame could manifest in several distinct dimensions of the self. A multidimensional conceptualization framework of shame proposed by Andrews and colleagues (2002) delineates three principal forms. Characterological shame entails a global, deeply ingrained negative evaluation of one’s entire self or core identity, often accompanied by the belief that one is inherently flawed or defective (e.g., I am a failure.). It might relate with childhood trauma or abuse, leading to emotional suppression and self-blame, and mental health problems. This form of shame is strongly associated with a range of adverse psychological outcomes; Behavioral shame, by contrast, pertains to negative judgments of specific actions, mistakes, or failures in performance (e.g., I did something bad.). Such shame may prompt individuals to avoid behavioral display of imperfection and conceal perceived shortcomings; Finally, bodily shame involves self-critical feelings and devaluating assessments related to one’s physical appearance; the perceived inability to meet internal or societal bodily ideals can evoke acutely painful experiences of shame (Velotti et al., 2017). Research showed women may adopt maladaptive emotion regulation when experiencing characterological shame (Nyström & Mikkelsen, 2012). Behavioural shame was associated with hostility, lower self-esteem, and guilt (Velotti et al., 2017). Bodily shame was associated with angry feelings, especially for women (Hejdenberg & Andrews, 2011).  Previous research has demonstrated that perfectionistic self-presentation is associated with elevated anxiety symptoms (Hewitt et al., 2003), and has emerged as a robust predictor of daily social anxiety (Mackinnon et al., 2014). Social anxiety is characterized by an intense fear of social or performance situations, particularly those involving unfamiliar individuals or evaluative scrutiny (Beidel & Turner, 2007). It involves experiences of anxiety such as emotional discomfort, fear, apprehension, inferiority, or worry) within interpersonal contexts, including social interactions and situations in which individuals are evaluated or scrutinized by others, accompanied by the belief that such evaluations will be negative or humiliating (American Psychiatric Association, 1994). It arises when people drive a perfect impression to others and have self-doubt about their ability to make perfection (Clark, 2005). Flett et al. (1989) found that perfectionism was correlated with trait anxiety and marginally with neuroticism. Hewitt et al. (2003) found that dimensions of perfectionistic self-presentation are generally related to measures of anxiety and that the nondisplay of imperfection dimension was correlated with anxiety symptom. They also found that perfectionistic self-promotion was consistently correlated with measures of social anxiety. Some researchers have found shame might be one of the reasons to cause social anxiety (Okano, 1994; Li et al., 2003, 2005). Zhong et al., 2002; 2003) supported this aspect that shame plays a core role to influence the development of social anxiety (Gilbert, 2000; Jutwak & Ferrari, 1997; Tangney et al., 1995). Experiencing shame can cause social avoidance, distress and social anxiety disorder (Fergus et al., 2010). Shame involves perceptions that others view oneself as negative personal characteristics and negative evaluations such as a strong sense of worthlessness and powerlessness (Qian et al., 2001). Shame proneness was positively correlated with social anxiety, and the decreasing of shame proneness had good effects on reducing individual’s social anxiety (Li et al., 2005). Shame proneness intervention is effective on the improvement of social anxiety level (Li et al., 2006). Over half of the social anxiety events were related with shameful events (Hirshfeld-Becker, Fredman, Robin, & Rosenbaum, 1999). The essence of social anxiety has been said to be an expectation of negative evaluation by others (Leitenberg, 1990). One theory is that social anxiety occurs when there is motivation to make a desired impression along with doubt about having the ability to do so (Leary, 2001). Excessive levels of nondisplay and nondisclosure of imperfection will appraise social environments as threatening, and have fears of making public errors (Arkin, Appelman, & Burger, 1980). For some social anxious individuals, nondisplay and nondisclosure of imperfection may be a consequence of social anxiety. Feeling of shame is a key factor to cause social anxiety when people interact with others. More shameful events in the reality increased and caused individuals to have anxious behaviours, and they are more likely to avoid from the reality and further have social anxiety problems (Li, Qian, & Zhong, 2005). Researchers have found that shame and social anxiety shared some similarities. They both tend to perceive things negatively and have social withdrawal and avoided behaviours, people with both or either shame-proneness or social anxiety are more susceptible to negative feelings such as anxiety, fear, and depression (Li, et al., 2005; Qian et al., 2001; Tangney, 1995; Tangney et al., 1992).

Cross-cultural study

Culture encompasses shared common values and beliefs that shape how individuals interpret the social world, as well as a set of practices that arise from them (Triandis, 1995). Research has documented differences in the cognitive and emotional styles of individuals from Eastern cultures and Western cultures. Two influential theoretical frameworks include individualism-collectivism (Hofstede, 1980, 1994) and independent-interdependent self-construal provide a robust basis for understanding how members of different cultural groups construct their interpersonal relationships (Markus & Kitayanma, 1991). Eastern cultures are generally characterized as collectivistic, prioritizing group interests, hierarchical relationships and a sense of belonging. In contrast, Western cultures tend to be more individualistic, emphasizing autonomy, personal freedom, individual achievement, and self-interests. Markus and Kitayama (1991) conceptualized interdependent self-construal as rooted in close, relational social ties, whereas independent self-construal is characterized as stable, distinct and relatively decontextualized from social relationships. Prior research (Cross & Madson, 1997) has further suggested that Western societies tend to socialize males toward independence and autonomy, while females are more often encouraged to adopt interdependent orientations. In many Asian cultures, such as mainland China, social roles have been shaped by Confucian traditions for over thousand years (Tu, 1985). Within this framework, men have traditionally been expected to lead the household, whereas women are expected to serve, follow and obey male authority. Much of the Confucian tradition continues to influence contemporary Chinese society (Tu, 1985). Consequently, cultural differences and gender role expectations may exert a stronger influence on three facets of perfectionism-particularly nondisplay and nondisclosure imperfection and on psychological functioning among mainland Chinese individuals than among European Canadians. This pattern may be partly explained by evidence suggesting that individuals in Eastern cultures are more likely to engage in self-critical tendencies (Heine et al. 2000), and try to avoid losing face, and Westerners tend to engage in self-enhancement (Chang et al., 2001; Heine et al., 1999). Westerners showed optimistic bias positive to the negative outcomes (Chang & Asakawa, 2003), in contrast, Japanese showed pessimistic bias for negative outcomes, therefore European Canadians might be positive for not showing perfect in front of people than mainland Chinese when they are in negative situations. Chang and Chang (2009) examined the effects of socially prescribed expectation on emotions and cognitions in Asian and European Americans and European American participants. Their findings indicated that priming participants with negative outcome expectancies led to higher levels of negative perfectionisms and lower levels of positive perfectionism. Asian Americans reported lower levels of subjective well-being and higher levels of negative perfectionism than European Americans. Similarly, Xie, Leong and Feng (2007) found that socially prescribed perfectionism was a better predictor of social anxiety among Chinese participants, whereas it was a stronger predictor of trait anxiety among Caucasian participants. Experiences of shame have been observed across nearly all cultures; however, the subjective experience of shame may comprise different components depending on cultural context. Cross-national studies suggest that Asian parents are more likely than European American parents to rely on practices such as love withdrawal, shame, and guilt induction (Wu et al., 2002). Zane and Yeh (2002) found that Asian American college students reported stronger motivation to avoid loss of face compared to their European Americans counterparts.  In Western cultural contexts, shame has been linked to the development of social anxiety (Chorpita & Barlow, 1998). Some cross-national data suggests that shyness or social anxiety may be viewed as normative, consistent with socialization goals, and distinct from other adjustment difficulties in Asian cultures (Chen, Rubin, & Li, 1995). Moreover, there are ethnic differences in the experience of social anxiety, with replicated findings of Asian American college students reporting higher levels of social anxiety symptoms than European American students (Okazaki, 1997, 2002; Okazaki, Liu, Longworth, & Minn, 2002). However, research on Asian Americans suggests that their experience of social anxiety is associated with both subjective distress and functional impairment (Hsu & Alden, 2007). Chinese participants tended to have elevated scores on self-reported anxiety measures than participants from Western countries (Chataway & Berry, 1989; Chen, 1996). China has been referred to as the shame-society by some social and behaviour scientists, and advocates Chinese people should have the sense of shame. Shame has a more important effect on social anxiety in Chinese culture, compared to its effect on Americans (Zhong et al., 2008). The experience of shame should be influenced by culturally-determined moral styles, social norms, and people’s self-concepts (i.e., interdependent self, independent self). Independent self-construal is defined as a bounded, unitary and stable self that is separated from social context, whereas interdependent self-construal refers to a flexible and variable self that emphasizes social contexts such as interpersonal relationships, status and roles (Singelis, 1994). Okazaki (1997) found that independent self-construal as a significant (negative) predictor of social anxiety, but not depression. The positive correlations exist between interdependent self-construal and anxiety, the more interdependent, the greater the tendency that one may monitor other’s evaluation. Therefore, interdependent self-construal might be strongly related with social anxiety for mainland Chinese than European Canadians. Different self-construals might influence various reactions to shame situations, which might different levels of social anxiety and social phobia.

Study hypotheses

In China, a substantial body of research has examined the characteristic of the perfectionism following the revision of the Experience Shame Scale was revised by Qian et al. (2000). However, the perfectionist self-presentation (PSPS) has received relatively little scholarly attention, Zhang et al. (2007) revalidated the PSPS scale in Chinese context and explored its relationship with self-esteem. This project is distinctive in that it is the first to simultaneously investigate perfectionistic self-presentation and experience of shame, as well as their associations with social anxiety. Individuals who experience shame are often characterized by heightened self-blame, tendencies to attribute blame to others, and a strong motivation to avoid shame-eliciting situations. For individuals who strongly endorse perfectionistic self-presentation, such situations may be particularly distressing, imposing greater psychological strain and potentially exacerbating symptoms of social anxiety. If young people cannot recover from this negative emotion quickly, they might seal themselves off from others and may even develop anti-social or violent tendencies (Lutwak & Ferrari, 1997; Tangney, 1994; Anolli & Pascucci, 2005), and mental health problems (Thomas, 2000; Zhong, 1993). Specially, for individuals who engage in perfectionistic self-presentation during interpersonal communication, greater tendencies toward nondisclosure (i.e., concealing imperfections) and nondisplay (i.e., avoiding the display of imperfections), combined with a failure to present themselves as perfect to others, are likely to intensify feelings of shame regarding their behavior in such situations. Furthermore, in the absence of effective coping strategies, situation that expose their flaws may exacerbate emotional distress and heighten levels of social anxiety. To date, no empirical cross-cultural research has examined the relationships among perfectionistic self-presentation, shame, and social anxiety in non-clinical samples of Chinese and European Canadians. Building on prior literature, the present study seeks to integrate these constructs to investigate whether the need to appear perfect, in interaction with shame elicited by perceived wrongdoing, predicts heightened symptoms of social anxiety. In particular, individuals from distinct cultural backgrounds, characterized by differing self-concepts (e.g., interdependent versus independent self-concept) might influence their feeling shame in divergent ways and exhibit varying strength of association among these variables. In summary, the primary aim of this study is to examine the relationship among these factors in non-clinical samples of Chinese and European Canadians. Specifically, we would investigate whether the need to present a flawless image and the experience of shame differ across these two cultural groups in their association with social anxiety. Regression analyses and multivariate analyses of variance (MANOVA) would be employed to test the proposed relational model among the three constructs of perfectionistic self-presentation and to compare and identify their similarities and differences between Chinese and European Canadian models. We hypothesized that all three facets of perfectionist self-presentation would significantly correlate with experience of shame, and social anxiety. In particular, we expect that Chinese participants who disclose perceived wrongdoing in interpersonal contexts would experience heightened shame, which in turn would be positively associated with greater social anxiety. Furthermore, we predicted that interactions between perfectionist self-presentation and interdependent self-construal interactions would significantly elevated levels of shame, and social anxiety. Also, there would be cultural differences between European Canadians and mainland Chinese about perfectionistic self-presentation with either interdependent self-construal or independent self-construal on experiencing shame, and social anxiety. Because people from mainland China with interdependent self-construal might be sensitive for social environment, take more pressures from the surroundings, and focus on more evaluations from others, they might have higher social anxiety, and more shame than Euro-Canadians.

In summary, the hypotheses of this study were listed as follows:

Hypothesis 1: The relational model encompassing perfectionistic self-presentation, shame, social anxiety, social phobia, interdependent and independent self-construal was expected to differ between Chinese and European Canadian samples. Drawing on existing the theory and empirical evidence, the interpersonal dimension of perfectionism-specifically perfectionistic self-presentation would be hypothesized to exhibit distinct patterns of association with shame, social anxiety, self-construals across cultural groups. More specifically, with respects to the three facets of perfectionistic self-presentation, it was anticipated that perfectionistic self-promotion would be negatively associated with shame, social anxiety, social phobia and interdependent self-construal, and positively linked with independent self-construal in European Canadian participants. In contrast, among Chinese participants, and perfectionistic self-promotion would be positively linked with interdependent self-construal and negatively linked with independent self-construal. Furthermore, nondisclosure and nondisplay imperfections would be hypothesized to be positively related to experience of shame, and social anxiety, social phobia as well as to independent and independent self-construal, in both European Canadian and Chinese samples.

Hypothesis 2: Cultural differences are expected to emerge across the three facets of perfectionistic self-presentations, as well as in levels of shame, and social anxiety or social phobia. Specifically, individuals from Mainland China would be anticipated to report higher levels of perfectionistic self-promotion and greater experiences of shame than their Euro-Canadians. In contrast, Mainland Chinese participants would be expected to report lower levels of nondisplay and nondisclosure imperfections than European Canadians. This may be because European Canadians who strive to appear flawless and actively avoid revealing imperfections may experience greater psychological strain in challenging interpersonal situations, leading to heightened anxiety, fear, and avoidance behaviors. Moreover, shame is generally regarded as a maladaptive emotion in Western cultural contexts, whereas in Eastern cultures such as China, shame is often considered a more adaptive and socially regulating emotion. Consequently, these cultural distinctions are expected to contribute to systematic differences across perfectionistic self-presentation, shame, social anxiety and social phobia related outcomes.

Hypothesis 3: Interdependent self-construal would be expected to function as a mediator in the relationships among perfectionistic self-presentation, experience of shame, and social anxiety or social phobia within the Chinese sample, but not within the European Canadian sample. In addition, the present study examined potential interaction effects among these variables. Gender difference in self-construal would be also anticipated, with females might be expected to endorse higher levels of interdependent self-construal, whereas males might be expected to report higher independent self-construal.

Participants: The sample consisted of 126 Chinese undergraduate students (63 males; 63 females) from a North University of mainland China located in Anshan city. Participants had a mean age of 19.87 years (SD = 0.53). All participants were volunteers to take part in the study and provided informed consent. Upon completion of the surveys, participants received a debriefing sheet in Mandarin and a small monetary gift (approximately $5 Canadian). In addition, 132 European Canadian undergraduate students were recruited from the undergraduate participant pool in York University in Toronto, Canada (70 males; 62 females; Mean age = 20.55; SD = 4.64). All European Canadian participants were born in Canada and identified as being of European descent. They were given the consent form and the package of the surveys online.

Measurements:

1. Demographic information. Students were asked to fill out a questionnaire about their major, birth, gender, which year of the study, religion, and where they were born. Chinese undergraduate students will answer the questionnaires in Mandarin, but all the European Canadian students will answer the surveys in English.
2. Perfectionistic self-presentation Scale (PSPS) (Hewitt et al., 2003). The PSPS is a 27-item measure composed of three subscales: perfectionistic self-promotion (e.g., “I try always to present a picture of perfection”), non display of imperfection (e.g. “I do not want people to see me do something unless I am very good at it”), and nondisclosure of imperfection (e.g., “I should solve my own problems rather than admit them to others”). Participants rate (dis)agreement with items on a 7-point scale, where higher scores indicate greater levels of perfectionistic self-presentation. It has solid reliability and validity.
3. The Experience of Shame Scale (ESS) (Qian et al., 2000). The ESS contains a total of 25 items and measures three areas of shame: characterological shame, behavioral shame (BES), and bodily shame (BOS). Questions in this scale relate to how often they felt shame in different conditions. Subjects were asked to rate each item from ‘‘1’’ (not at all) to ‘‘4’’ (often). Higher scores indicated higher levels of shame proneness. The test-retest reliability was 0.88, and the internal reliability of each subscale measured by the Cronbach a ranged from 0.70 to 0.86 (Qian et al., 2000).
4. Social Interaction Anxiety Scale (SIAS) and Social Phobia Scale (SPS) (Mattick & Clarke, 1998). This set of companion scales is commonly used self-report instruments for social anxiety. They assess the main fears of interaction with others and the fears of being observed while doing routine activities, and avoidance of social phobia, focusing respectively on interaction fears and more specific performance-based fears. Statements are rated on a 5-point Likert scale. Each scale consists of 20 items base on a Likert-type scale ranging from 0 (not at all) to 4 (extremely). For both the SIAS and SPS in various samples, with Cronbach’s alphas ranging from 0.88 to 0.94 for both scales. Also, both scales have demonstrated high test–retest reliabilities.
5. The Self-Construal Scale (Singelis, 1994). Participants’ levels of independent and interdependent in self-construal were measured using the Singelis Self-Construal scale. The scale consists of 24 items, with 12 items reflecting independence and 12 items reflecting interdependence. Each item is rated on a 7-point scale, with answers ranging from 1 (strongly disagree) to 7 (strongly agree). The mean scores range of interdependent self-construal is from 2.00 to 7.00 with higher scores reflecting more interdependence, and the mean scores range of independent self-construal is from 1.00 to 7.00 with higher scores reflecting more independence. Singelis reported coefficient alphas of .70 for the Independence subscale.74 and for the Interdependence subscale.

Procedure

Chinese participants were recruited from undergraduate students in North University in mainland China and were administered by the experimenters in different class. Participants were asked to sign the consent form and complete the questionnaire package in 45 minutes if they agreed to participate. They received a debriefing form and 5$ or equivalent gift from the experimenter.

European Canadian participants were recruited from the undergraduate participant pool at the York University. They answered the questionnaires online through URPP system of York University. After they submit their answers, each participant received a 1% bonus added to his/her final course grade.

Analysis

In this cross-sectional design study, the software’s IBM SPSS Statistics 26.0 were used. The descriptive statistics, MANOVA, Pearson correlation analyses, and regression statistic method were performed separately to study the relationships among all these factors for the Chinese and European Canadian samples.

Relationships between perfectionistic self-presentation, experience of shame, social anxiety, social phobia and interdependent/independent self-construal All of the variables were centered prior to analyses. In this way, the risk of multicollinearity was reduced. Pearson correlations were calculated for each of the independent variables including perfectionistic self-promotion, nondiplay of imperfections, nondisclosure of imperfections, characerological shame, behavioural shame, bodily shame; dependent variables including social anxiety, and social phobia; and controlling variables (interdependence self-construal, independence self-construal). The correlation analyses are first reported with the four cultural groups combined (see Table 1), and then with the two groups separately (see Table 2, Table 3).

** p < .01***. p < .001. † 0.5 < p <.10.

Table 1: Means and standard deviations of perfectionistic self-presentation, shame, loss of face, anxiety, and social phobia by culture group..

* p < .05. ** p< .01*** p < .001. † 0.5< p <.10.

Table 2: Correlations of all the predictors and outcome variables in two cultural groups Chinese students (N = 126) are above the diagonal; Euro-Canadian students (N = 104) are below the diagonal.

* p < .05. ** p < .01*** p <.001. † 0.5< p><.10.

Table 3: Means and standard deviations of perfectionistic self-presentation, shame, anxiety, social phobia, and self-construals by cultural groups.

Correlations between two individual variables and other variables by group

For the correlations in the mainland Chinese group, (see Table 2), perfectionistic self-promotion was positively correlated with nondisplay of imperfections (r=.42, p=.00), with nondiscolsure of imperfections (r=.26, p=.00), with bodily shame (r=.19, p=.04), and with social phobia (r=.20, p=.02). Nondisplay of imperfections was positively correlated with nondiscolsure of imperfections (r=.40, p=.00), with characterological shame (r=.35, p=.00), with behavior shame (r=.21, p=.00), with overall shame (r=.29, p=.00), with social anxiety (r=.21, p=.02), and with social phobia (r=.20, p=.02). Nondiscolsure of imperfections (r=.45, p=.00), was positively correlated with social phobia (r=.21, p=.02), and was negatively correlated with independent self-construal (r=-.20, p=.02) (see Table 2). Overall shame was positively correlated with social anxiety (r=.56, p=.00), with social phobia (r=.64, p=.00), and was negatively correlated with independent self-construal (r=-.22, p=.00) (see Table 2). Characterological shame was positively correlated with behavior shame (r=.64, p=.00), bodily shame (r=.56, p=.00), with overall shame (r=.91, p=.00), with anxiety (r=.49, p=.00), with social phobia (r=.56, p=.00), and with negatively correlated with independent self-construal (r=-.25, p=.01). Behavior shame was positively correlated with bodily shame (r=.60, p=.00), with overall shame (r=.88, p=.00), with social anxiety (r=.52, p=.00), and with social phobia (r=.58, p=.00). Bodily shame was positively correlated with overall shame (r=.76, p=.00), with anxiety (r=.41, p=.00), and with social phobia (r=.50, p=.00) (see Table 2).  Social anxiety was positively correlated with phobia (r=.68, p=.00), and was negatively correlated with independent self-construal (r=-.24, p=.01). Social phobia was negatively correlated with independent self-construal (r=-.26, p=.01) (see Table 2). Interdependent self-construal was positively correlated with independent self-construal (r=.28, p=.00) (see Table 2).  For the correlations in the European Canadian group (see Table 2), perfectionistic self-promotion was positively correlated with nondisplay of imperfections (r=.70, p=.00), with nondiscolsure of imperfections (r=.58, p=.00), with characterological shame (r=.35, p=.00), with behavioural shame (r=.52, p=.00),with bodily shame (r=.43, p=.00), with overall shame (r=.47, p=.00), and with social phobia (r=.24, p=.01). Nondisplay of imperfections was positively correlated with nondisclosure of imperfections (r=.52, p=.00), with characterological shame (r=.42, p=.00), with behavior shame (r=.53, p=.00), with bodily shame (r=.42, p=.00), with overall shame (r=.52, p=.00), with social anxiety (r=.35, p=.02), and with phobia (r=.30, p=.00). Nondisclosure of imperfections was positively correlated with characterological shame (r=.41, p=.00), with behavior shame (r=.41, p=.00), with bodily shame (r=.39, p=.00), with overall shame (r=.45, p=.00), with social anxiety (r=.40, p=.00), with social phobia (r=.44, p=.00), and with independent self-construal (r=-.20, p=.00) (see Table 3).  Overall shame was positively correlated with characterological shame (r=.94, p=.00), with behavior shame (r=.90, p=.00), with bodily shame (r=.78, p=.00), with social anxiety (r=.53, p=.00), with social phobia (r=.42, p=.00). Characterological shame was positively correlated with social anxiety (r=.57, p=.00). Behavioural shame was positively correlated with bodily shame (r=.60, p=.00), with social anxiety (r=.37, p=.00), and with social phobia (r=.29, p=.00). Bodily shame was positively correlated with social anxiety (r=.41, p=.00), with social phobia (r=.36, p=.00) (see Table 2).  Social anxiety was positively correlated with social phobia (r=.78, p=.00), and was negatively correlated with independent self-construal (r=-.42, p=.00). Social phobia was negatively correlated with independent self-construal (r=-.46, p=.00) (see Table 2). 

Cultural and gender differences on interdependent/independent self-construal 

We performed a series of Multivariate Analyses of Variance (MANOVAs) to look at mean differences among two different cultural group (MC: mainland Chinese; EC: European- Canadians) on interdependent self-construal, and independent self-construal. The means, standard deviations and F values for interdependent self-construal, and independent self-construal were shown in Table 3. Significant effects were explored using Bonferroni post-hoc pairwise comparisons with a .05 significance level. A 4 (cultural group) by 2 (gender) MANOVA showed that significant differences were found among two groups of cultures on the dependent variables interdependent self-construal, and independent self-construal, the Wilks’ Lambda multivariate test of overall differences among cultural groups was significant, Wilks’ lambda = .50, F (18, 209) = 11.47, p = .00, delta square = .50. There was also a significant overall main effect of gender on the dependent variables, Wilks’ Lambda = .87, F (18, 209) = 1.67, p = .04, delta square = .13. There were no significant overall interactions of cultural group by gender on the dependent variables, the Wilks’ Lambda = .92, F (18, 209) = 1.01, ns, delta square = .08. Analyses of variance (ANOVA) on each dependent variable were conducted as follow-up tests the MANOVA, the results showed that the cultural groups differed significantly on interdependent self-construal (see Table 4), F (1, 226) = 99.44, p=.00, delta square = .31. Post-hoc analyses revealed that mainland Chinese (M = 66.61) were higher than European Canadians (M = 54.69) on interdependent self-construal. There was no a significant effect of cultural group on independent self-construal, F (1, 226) = .81, ns. For the effect of gender on self-construals, the effect of gender on interdependent self-construal was not significant, F (1, 226) = .43, ns, delta square = .00. The effect of gender on independent self-construal was significant, F (1, 226) = 4.02, p = .05, delta square = .02. Women (M = 55.53) reported less independent self-construal than men (M = 58.14).  There was no interaction of culture and gender on interdependent self-construal, F (1, 226) = .64, ns, delta square = .00. 

Effects of culture on perfectionistic self-presentation, shame, social anxiety and social phobia

According to our first and second hypothesis, perfectionism and emotion variables among two groups should differ cross-culturally. We performed a two-way multivariate Analyses of Variance (MANOVAs) to look at mean differences among two different cultural group (MC: mainland Chinese; EC: European- Canadians) by gender for these variables (perfectionism self-promotion, nondisplay imperfectionism, nondisclosure imperfectionism, shame, social anxiety, and social phobia) and then ran a second MANOVA to look at the effects of culture and gender on these variables. The overall means, standard deviations and F values for the MANOVAs for the dependent variables are shown in Table 3. Significant effects were explored using Bonferroni post-hoc pairwise comparisons with a .05 significance level. A 4 (cultural group) by 2 (gender) MANOVA showed that significant differences were found among two groups of cultures on the dependent variables (perfectionism self-promotion, nondisplay imperfectionism, nondisclosure imperfectionism, shame, social anxiety, and social phobia), the Wilks’ Lambda multivariate test of overall differences among cultural groups was significant, Wilks’ lambda = .79, F (6, 249) = 10.88, p = .00, delta square = .21. There was also a significant overall main effect of gender on the dependent variables, Wilks’ Lambda = .95, F (6, 249) = 2.15, p = .04, delta square = .05. There was no significant overall interactions of cultural group by gender on the dependent variables, the Wilks’ Lambda = .97, F (6, 249) = 1.01, ns, delta square = .03.  Analyses of variance (ANOVA) on each dependent variable were conducted as follow-up tests the MANOVA, the results showed that the cultural groups differed significantly on perfectionistic self-promotion (see Table 3), F (1, 254) = 7.20, p=.01, delta square = .03. Post-hoc analyses revealed that mainland Chinese (M = 44.35) were higher than European Canadians (M = 41.07) on perfectionistic self-promotion.  There was no significant effect of cultural group on nondisplay imperfectionism, F (1, 254) = .00, ns, delta square = .00. There was a marginally significant effect of cultural group on nondisclosure imperfectionism, F (1, 254) = 3.00, p = .08, delta square = .01. Post-hoc analyses revealed that mainland Chinese (M = 22.75) were marginally lower than European Canadians (M = 24.18) on nondisclosure imperfectionism. There was no significant effect of cultural group on overall shame, F (1, 254) = .04, ns, delta square = .00. There was significant effect of cultural group on anxiety, F (1, 254) = 5.06, p = .03, delta square = .02. Post-hoc analyses revealed that mainland Chinese (M = 23.12) were lower than European Canadians (M = 26.71) on social anxiety. There was a significant effect of cultural group on social phobia, F (1, 254) = 10.75, p=.00, delta square = .03. Post-hoc analyses revealed that mainland Chinese (M = 22.42) were higher than European Canadians (M = 17.22) on social phobia.

The Effects of gender on perfectionistic self-presentation and emotions variables

The effect of gender on perfectionism self-promotion was not significant, F (1, 254) = .00, ns, delta square = .00. The effect of gender on nondisplay imperfection was not significant, F (1, 254) = .16, ns, delta square = .00. The effect of gender on nondisclosure imperfection was significant, F (1, 254) = 7.74, p = .01, delta square = .02. Women (M = 22.31) reported less nondisclosure imperfection than men (M = 24.61). The effect of gender on overall shame was not significant, F (1, 254) = .12, ns, delta square = .00. The effect of gender on anxiety was not significant, F (1, 254) = 2.22, ns, 2 = .00. The effect of gender on social phobia was not significant, F (1, 254) = .06, ns, delta square = .00. 

The interaction of culture and gender on perfectionistic self-presentation and emotion variables

There was no interaction of culture and gender on perfectionistic self-promotion, F (1, 254) = .03, ns, delta square = .00. There was no interaction of culture and gender on nondisplay imperfectionism, F (1, 254) = .23, ns, delta square = .00. There was marginally interaction of culture and gender on nondisclosure imperfectionism, F (1, 254) = 3.32, p= .07, delta square = 01. There was no interaction of culture and gender on shame, F (1, 254) = .05, ns, delta square = .00. There was a marginal interaction of culture and gender on anxiety, F (1, 254) = .04, ns, delta square = 01. There was no interaction of culture and gender on phobia, F (1, 226) = .55, ns, delta square = .00. 

Regression Analyses 

Since there were some significant correlations between interdependent/ independent self-construal and types of perfectionistic self-presentation, experience of shame, anxiety, and phobia, we included these as a potential mediator in the regression analyses. To test how these self-related variables were correlated to the different types of perfectionisms, shame, social anxiety and phobia. Therefore, a series of regression analyses were performed. Linear multiple regressions were conducted with types of perfectionisms, shame, anxiety, and phobia as the dependent variable, and culture, interdependent/ independent self-construal as the independent variables. Analyses were performed using SPSS REGRESSION for evaluation of assumptions. Data were screened and no outliers among the cases were found. All assumptions were met, and the data showed a normal distribution. We recoded mainland Chinese culture into a dummy variable, which compared mainland Chinese (MC) against another type of cultural group European Canadian (EC). Dummy coded variable and gender were added in the first step. The second step included interdependent/ independent self-construal. The third step included the two-way interactions between all variables.

Three Facets of Perfectionistic self-presentation

Perfectionistic self-promotion

For the perfectionistic self-promotion, the first step including the dummy coded cultural variables and gender significantly predicted perfectionistic self-promotion, R2ch= .03, Fch (2, 255) = 3.64, p < .05. The dummy variables isolating Mainland Chinese (β= .17, p < .01) was significant. The effect of gender was not significant (β = .00, ns). Thus, participants of Chinese descent generally engaged in more perfectionistic self-promotion than European Canadians, maybe the residence of the country matter.  When interdependent and independent self-construal were added in the second step, the equation was marginally significantly improved, R2ch = .02, Fch (2, 253) = 2.90, p= .06. In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step did not improve the fit of the equation above that of the equation with the main effects, R2ch = .01, Fch (2, 251) = 1.06, ns. The final equation model (step 2) with just the main effects marginally significantly predicted 3% of the variance in perfectionistic self-promotion, R2adj= .03, F (2, 253) = 2.90, p = .05. These results suggest that mainland Chinese undergraduate students are more likely to have more perfectionistic self-promotion than European Canadians, regarding to their cultural environment.  There were not enough evidence to show interdependent and independent self-construal mediated between culture and perfectionistic self-promotion, and no significant interactions with these variables.

Nondisplay imperfections

For the nondisplay imperfections, the first step including the dummy coded cultural variables and gender not significantly predicted nondisplay imperfections. R2ch = .00, Fch (2, 255) = .08, ns. The effect of gender was not significant (β = -.02, ns). Thus, participants of two descents generally have no difference to engage in more nondisplay imperfections, maybe the residence of the country does not matter.  When interdependent and independent self-construal were added in the second step, the equation was marginally significantly improved, R2ch = .03, Fch (2, 253) = 2.93, p= .06. Interdependent self-construal increased nondisplay imperfections. In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step did not improve the fit of the equation above that of the equation with the main effects, R2ch = .01, Fch (2, 251) = .79, ns.

Nondisclosure imperfections

For the nondisclosure imperfections, the first step including the dummy coded cultural variables and gender significantly predicted nondisclosure imperfections, R2ch = .04, Fch (2, 255) = 5.68, p < .01. The dummy variables isolating Mainland Chinese (β =-.11, p = .08) was marginally significant. The effect of gender was significant (β = -.17, p < .01). Thus, participants of Chinese descent generally engaged in less nondisclosure imperfections than European Canadians, maybe the residence of the country matter. When interdependent and independent self-construal were added in the second step, the equation was significantly improved, R2ch = .02, Fch (2, 253) = 7.77, p = .00. Interdependent self-construal marginally increased nondisclosure imperfections (β = .13, p = .07), independent self-construal decreased nondisclosure imperfections (β = -.25, p = .00). In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step marginally improved the fit of the equation above that of the equation with the main effects, R2ch = .02, Fch (2, 251) = 2.64, p = .08. The final equation model (step 2) with just the main effects marginally significantly predicted 8% of the variance in nondisclosure imperfections, R2adj= .08, F (2, 253) = 7.77, p < .01.  These results suggest that mainland Chinese undergraduate students are less likely to have nondisclosure imperfections than European Canadians, regarding to their cultural environment.  The relationship between mainland Chinese and nondisclosure imperfections was marginally significant (β = -.11, p = .07). When controlling for interdependent self-construal, the mainland Chinese dummy variable was significant but decreased (β = -.17, p < .01). But the relationship between interdependent self-construal and nondisclosure imperfections was not significant (β = -.01, ns). Therefore, interdependent self-construal was not a partial mediator between the mainland Chinese dummy variable and nondisclosure imperfections. Although the relationship between independent self-construal and nondisclosure imperfections was significant (β = -.18, p = .00), however, the relationship between mainland Chinese and independent self-construal was not significant (β = -.01, ns), therefore, independent self-construal was not a partial mediator between Chinese dummy variable and nondisclosure imperfections.

Experience of shame

For experience of shame, the first step including the dummy coded cultural variables and gender not significantly predicted experience of shame. R2ch = .00, Fch (2, 255) = .08, ns. The effect of gender was not significant (β = -.02, ns). No cultural difference in experience of shame.  When interdependent and independent self-construal were added in the second step, the equation was significantly improved, R2ch = .06, Fch (2, 253) = 7.44, p = .00. In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step did not improve the fit of the equation above that of the equation with the main effects, R2ch = .01, Fch (2, 251) = .03, ns. For the subscale bodily shame, the first step including the dummy coded cultural variables and gender significantly predicted bodily shame, R2ch = .10, Fch (2, 255) = 13.70, p = .00. The dummy variables isolating Mainland Chinese (β = -.29, p = .00) was significant. The effect of gender was significant (β = .13, p = .03). Thus, participants of Chinese descent generally engaged in less bodily shame than European Canadians, maybe the residence of the country matter.  When interdependent and independent self-construal were added in the second step, the equation was significantly improved, R2ch = .03, Fch (2, 253) = 3.58, p = .03. Independent self-construal decreased bodily shame (β = -.16, p = .01). In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step did not improve the fit of the equation above that of the equation with the main effects, R2ch = .00, Fch (2, 251) = .10, ns. The final equation model (step 2) with just the main effects significantly predicted 16% of the variance in perfectionistic self-promotion, R2adj= .11, F (2, 253) = 3.58, p = .03. These results suggest that mainland Chinese undergraduate students are less likely to have social bodily shame European Canadians.  The relationship between mainland Chinese and bodily shame was significant (β =-.29, p = .00). When controlling for interdependent and independent self-construal, the mainland Chinese dummy variable was significant but decreased (β = -.33, p = .00). However, the relationship between mainland Chinese and independent self-construal was not significant too (β = -.01, ns), therefore, independent self-construal was not a partal mediator between the culture and bodily shame.

Social anxiety

For social anxiety, the first step including the dummy coded cultural variables and gender significantly predicted social anxiety, R2ch = .03, Fch (2, 255) = 3.79, p < .05. The dummy variables isolating Mainland Chinese (β = -.14, p < .05) was significant. The effect of gender was not significant (β = -.09, ns). Thus, participants of Chinese descent generally engaged in less social anxiety than European Canadians, maybe the residence of the country matter. When interdependent and independent self-construal were added in the second step, the equation was significantly improved, R2ch = .09, Fch (2, 253) = 12.97, p = .00. independent self-construal decreased social anxiety (β = -.32, p < .001). In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step did not improve the fit of the equation above that of the equation with the main effects, R2ch = .01, Fch (2, 251) = .67, ns. The final equation model (step 2) with just the main effects significantly predicted 10% of the variance in social anxiety, R2adj= .10, F (2, 253) = 12.97, p = .00. These results suggest that mainland Chinese undergraduate students are less likely to have social anxiety than European Canadians. The relationship between mainland Chinese and social anxiety was significant (β = -.14, p = .03). When controlling for interdependent and independent self-construal, the mainland Chinese dummy variable was significant but decreased (β = -.19, p = .01). Although the relationship between independent self-construal and social anxiety was not significant (β = -.27, p = .00), however, the relationship between mainland Chinese and independent self-construal was not significant (β = -.01, ns). Therefore, independent self-construal was not a partial mediator between the mainland Chinese dummy variable and social anxiety.

Social phobia

For social phobia, the first step including the dummy coded cultural variables and gender significantly predicted social phobia, R2ch = .04, Fch (2, 255) = 5.47, p < .01. The dummy variables isolating Mainland Chinese (β = .20, p < .001) was significant. The effect of gender was not significant (β = -.01, ns). Thus, participants of Chinese descent generally engaged in more social phobia than European Canadians, maybe the residence of the country matter. When interdependent and independent self-construal were added in the second step, the equation was significantly improved, R2ch = .13, Fch (2, 253) = 19.63, p = .00. Interdependent self-construal increased social phobia (β = .15, p < .05), independent self- construal decreased social phobia (β = -.38, p < .001). In the third step, we entered the interaction variables for culture, gender, interdependent and independent self-construal. This step did not improve the fit of the equation above that of the equation with the main effects, R2ch = .00, Fch (2, 251) = .41, ns. The final equation model (step 2) with just the main effects significantly predicted 16% of the variance in social phobia, R2adj= .16, F (2, 253) = 19.63, p = .00. These results suggest that mainland Chinese undergraduate students are more likely to have social phobia than European Canadians. The relationship between mainland Chinese and social phobia was significant (β= .20, p = .00). When controlling for interdependent and independent self-construal, the mainland Chinese dummy variable was significant but decreased (β = .13, p = .04). however, the relationship between independent self-construal and social phobia was not significant (β = -.27, p = .00). The relationship between mainland Chinese and independent self-construal was not significant too (β = -.01, ns), therefore, independent self-construal was not a partial mediator. However, the relationship between mainland Chinese and interdependent self-construal was significant (β = .46, p = .00). The relationship between interdependent self-construal and social phobia was marginally significant (β = .11, p = .08). A Sobel test showed that the drop in the predictive power of the mainland Chinese variable was significant, Z = 1.68, p = .09. Thus, interdependent self-construal marginally partially mediated the effect of culture on social phobia, independent self-construal was not a partial mediator between the mainland Chinese dummy variable and social anxiety.