Motiv Emot (2008) 32:6980 DOI 10.

1007/s11031-008-9084-2

ORIGINAL PAPER

The relation between value priorities and proneness to guilt, shame, and empathy
Mia Silfver Klaus Helkama Jan-Erik Lonnqvist Markku Verkasalo

Published online: 8 March 2008 Springer Science+Business Media, LLC 2008

Abstract The relations of value priorities (the Schwartz Value Survey) to guilt, shame (Tangneys Test of SelfConscious Affect) and empathy (Davis Interpersonal Reactivity Index) were examined in two samples, one of 1519-year-old secondary school students (N = 207), and the other of military conscripts (N = 503). As hypothesized, guilt-proneness was, in both samples, positively related to valuing universalism, benevolence, tradition, and conformity, and negatively related to valuing power, hedonism, stimulation, and self-direction. The results for empathic concern and perspective-taking were similar, but their relation to the opennessconservation value dimension was weaker. Shame and personal distress were weakly related to values, suggesting that voluntary control is less important for these tendencies. In general, self-transcendence and conservation values seem compatible with prosocial tendencies, whereas self-enhancement and openness do not. Keywords Values Guilt Shame Empathy

Two factors have been suggested as the main sources of moral motivation. First, an individuals values, and the moral principles based on those values, have been seen as the motivational basis for morality (Rest 1984). Second, several researchers have emphasized the roles of empathy,

M. Silfver (&) K. Helkama Department of Social Psychology, University of Helsinki, P.O. Box 54, 00014 Helsinki, Finland e-mail: [Link]@helsinki. J.-E. Lonnqvist M. Verkasalo Department of Psychology, University of Helsinki, Helsinki, Finland

guilt, and shame in moral development (Eisenberg 2000; Hoffman 2000; Tangney and Dearing 2002). These emotions are seen as central to moral motivation, and are therefore often referred to as moral emotions. Other emotions have also been linked to morality, for example, anger, disgust, and contempt (Rozin et al. 1999). However, these emotions are directed towards others moral transgressions, and for that reason, they can be regarded as less important when focusing on the sources of an individuals moral motivation. The relation between values and moral emotions has been studied very little, and the two research traditions have remained quite separate. However, there are theoretical reasons to expect that values and moral emotions are related, and therefore research of the topic would enhance our understanding of the nature of moral motivation. Values refer to things that people believe are worth desiring in life. According to Rokeach (1973), values include cognitive, affective, and behavioral components; values are cognitions of the desirable that foster emotions and motivate behavior. In the present study, values are examined using Schwartzs model (1992), which was developed based on the work of Rokeach. Schwartz views values as the criteria people use to select and justify actions and to evaluate people and events. Values are general life goals that guide behavior and the evaluation of behavior of the self and others. Therefore, values can be seen as important antecedents to the self-conscious emotions of guilt, shame, and pride. For example, someone who values tradition is likely to feel guilty over transgression of a traditional norm. In the Schwartz model, values describe an ideal self (e.g., honest) or an ideal state of affairs (e.g., equality), and they are ranked in terms of their relative importance. Values reect basic human needs: biological needs, requisites of coordinated social interaction, and survival and welfare of groups.

123

70

Motiv Emot (2008) 32:6980

Schwartzs model is special because of the way it has been constructed. Values and value items were chosen based on a large quantity of empirical data from different cultures around the world. By performing multidimensional scaling of these data, Schwartz has identied a set of values that are understood similarly in different cultures (see Table 1). This means that there is a cross-cultural consensus on which of these values are compatible and which are in conict. In Schwartzs model, the values form a circle along two axes: conservation versus openness to change, and self-enhancement versus self-transcendence. The values of tradition, conformity, and security represent conservation, whereas self-direction and stimulation represent openness to change. Values of universalism and benevolence represent self-transcendence: the willingness to transcend selsh concerns for the benet of others. Power and achievement represent self-enhancement values: the motivation to enhance ones personal interests. Hedonism is usually considered to include elements of both openness and self-enhancement. Conceiving values as organized in a circular structure implies that the whole set of 10 values should relate to any other variable (e.g., guilt) in an integrated manner. This means that correlations between values and other variables should form a sinusoid curve. If some variable correlates positively with universalism, it should correlate negatively with power, and the correlations should decrease when moving from universalism towards power around the circular model. This is referred to as the integrated hypothesis. Compared to values, the tendencies toward empathy, guilt, and shame are, in some respects, different constructs. They refer to tendencies towards specic emotional reactions in certain situational contexts. Values are conscious goals in life that can elicit emotions, but an emotional

reaction is not included in the denition of value. While certain cognitions can trigger the experience of empathy, guilt, or shame, the triggering processes can also be less conscious and not under voluntary control. This means that a persons conscious value priorities and tendencies towards certain emotional reactions may not always be congruent. Therefore, we dene values and emotional tendencies as separate but interrelated concepts. The term empathy has been used in several different ways within psychological research. Most current approaches dene empathy as a construct that has several different but related components (Davis 1994). Eisenberg (1986) emphasizes differentiating cognitive perspectivetaking and emotional reaction. Perspective-taking, imagining how another person experiences things, can lead to an empathic emotional reaction, but not necessarily: perspective-taking can be seen as a tool that can also be used for malicious purposes. Furthermore, according to Eisenberg, an emotional reaction can take two different forms, sympathy or personal distress. When feeling sympathy, a person responds to another persons emotion with a feeling that is similar to, but not identical with, what the other person is feeling. For example, one may feel concern for someone who is sad or distressed. Davis (1994) has labeled this emotion empathic concern. Personal distress means reacting to another persons emotional state with a negative, self-oriented feeling that does not include concern for the other. Eisenberg maintains that it is crucial to distinguish between sympathy and personal distress, because only sympathy is likely to motivate prosocial behavior, whereas personal distress often leads to avoidant behavior. There is much research supporting this view (Batson 1991; Davis 1994; Eisenberg et al. 2001; Litvak-Miller and McDougall 1997). Several studies indicate that the tendency

Table 1 Values and their contents by Schwartz (single value items in parentheses) Power: social status and prestige, control or dominance over people and resources (social power, authority, wealth, preserving my public image) Achievement: personal success through demonstrating competence according to social standards (successful, capable, ambitious, inuential) Hedonism: pleasure and sensuous gratication for one-self (pleasure, enjoying life, self-indulgent) Stimulation: excitement, novelty and challenge in life (daring, a varied life, an exciting life) Self-direction: independent thought and action-choosing, creating, exploring (creativity, freedom, independent, curious, choosing own goals) Universalism: understanding, appreciation, tolerance and protection for the welfare of all people and for nature (broadminded, wisdom, social justice, equality, world at peace, a world of beauty, unity with nature, protecting the environment) Benevolence: preservation and enhancement of the welfare of people with whom one is in frequent personal contact (helpful, forgiving, honest, loyal, responsible) Tradition: respect, commitment, and acceptance of the customs and ideas that traditional culture or religion provide the self (humble, accepting my portion of life, devout, respect of tradition, moderate) Conformity: restraint of actions, inclinations and impulses likely to upset or harm others, and violate social expectations or norms (politeness, obedient, self-discipline, honoring parents and elders) Security: safety, harmony, and stability of society, of relationships, and of self (family security, national security, social order, clean, reciprocation of favors)

123

Motiv Emot (2008) 32:6980

71

to feel sympathy rather than personal distress requires good coping skills and the ability to regulate ones emotions (e.g., Murphy et al. 1999; Okun et al. 2000). If a vicarious emotion becomes too uncomfortable, it is difcult to concentrate on others and their needs. Concept denition has also been essential in research on guilt and shame. The terms guilt and shame are often used interchangeably, but according to Tangney (1998), they should be dened as two distinct emotions. The essential difference concerns the role of the self. Shame involves global evaluations of the self (I am a horrible person) and often includes imagining how others perceive the self and desiring to change some aspect of the self. Shame also involves feelings of shrinking, being small, worthlessness, and powerlessness. Guilt, on the other hand, concentrates only on a specic behavior (I did a horrible thing). Guilt involves tension, remorse, regret, and concern for how ones behavior affects others. It is generally less painful than shame, because it does not threaten the sense of self. Tangney believes that this difference has important implications for the motivational features of these emotions. An ashamed person is motivated to hide, escape, or strike back, disguising shame by displaying anger towards the person inducing shame. A person who feels guilty, on the other hand, has a desire to confess, apologize, or repair the damage his or her behavior has caused. Accordingly, guilt can be seen as a more moral emotion than shame, both as an emotional state and as a dispositional tendency towards these emotional states. Studies using Tangneys scenario-based measure for guilt and shame-proneness, the Test of Self-Conscious Affect (TOSCA), have shown that shame-proneness was related to several indices of poor psychological and interpersonal adjustment, for example low self-esteem, anxiety, depression, and psychoticism (Averill et al. 2002; Tangney and Dearing 2002; Tangney et al. 1995). In contrast, guiltproneness has been found to relate to positive characteristics, such as interpersonal skills (Covert et al. 2003), perspective-taking (Leith and Baumeister 1998), anger control (Lutwak et al. 2001) and empathy (Fontaine et al. 2001). However, the TOSCA has been criticized for having a narrow denition of guilt and shame. It concentrates only on adaptive aspects of guilt and maladaptive aspects of shame. Several researchers have pointed out that there are also pathological, chronic forms of guilt, as well as adaptive forms of shame, that motivate reparation (Ferguson and Crowley 1997; Eisenberg 2000; Luyten et al. 2002). This should be acknowledged when interpreting results obtained from the TOSCA. The causal relation between values and emotional tendencies is not entirely clear. Because values are dened as abstract concepts, having certain values requires the ability to think in the abstract. The understanding of values is

similar for adolescents and adults, and the hypothesized value structure can be found in samples of adolescents (Verkasalo et al. 1996). However, children probably do not understand value concepts, at least not in the same way as do adults. On the other hand, the earliest forms of empathy, guilt, and shame are observable even in toddlers (Barrett 1998; Hoffman 2000). It is likely that emotional tendencies developmentally precede values. Still, this does not mean that certain values necessarily result from certain emotional tendencies. Values and emotional tendencies may be inuenced by common factors, such as parenting practices (Abell and Gecas 1997; Padilla-Walker 2007). In addition, gender relates to both values and emotional tendencies. Women and girls score higher than men and boys in proneness to empathy, guilt, and shame (e.g., Tangney and Dearing 2002; Ferguson and Eyre 2000; Davis 1994). Furthermore, across different cultures, women tend to value benevolence and universalism more than men, whereas men tend to value power, stimulation, hedonism, achievement, and self-direction more than women (Schwartz and Rubel 2005). Furthermore, the causal relation between values and emotions may be bi-directional. Certain emotional tendencies may affect ones value priorities, but values may also elicit matching emotions. For example, a person who is prone to empathy is likely to value the well-being of others, and a person who values the wellbeing of others pays attention to others emotions, which is likely to foster empathy. Although the direction of causality is uncertain, we hypothesized that values and emotional tendencies are related. Based on the content of the Schwartz values, we expected that the empathy dimensions would be positively related to universalism and benevolence, as these values concern the well-being of others, either of all people or of people with whom a personal relationship has been established. The opposite values, power, achievement, and hedonism, emphasize ones personal interests, and were therefore expected to have a negative relation to empathy. The conservationopenness dimension, on the other hand, is conceptually less directly connected to empathy. It is possible that the goal of maintaining a stable society sometimes conicts with the well-being of individual people. For example, in a war situation, one may be forced to harm others in order to pursue conservation values. Therefore, we did not assume any consistent pattern of connections between empathy and the value dimension of conservationopenness. Consistent with the above analysis, Myyry and Helk ama (2001) and Juujarvi (2003) found that empathy, as measured by Mehrabian and Epsteins (1972) Questionnaire Measure of Emotional Empathy (QMEE), had the strongest connection to the self-transcendenceselfenhancement-dimension. Empathy related positively to

123

72

Motiv Emot (2008) 32:6980

self-transcendence values and negatively to self-enhancement values. In the present study, we used an empathy measure differentiating sympathy and personal distress. Because personal distress is an involuntary reaction due to poor ability to regulate emotions, we expected personal distress, as compared to sympathy, to have a weaker relation to values. Still, it is likely that someone who values self-enhancement highly would not react to others emotions at all. What is true of the personal distress scale may also be true of TOSCA shame-proneness. It is a maladaptive reaction that relates to different indices of neuroticism, and thus, it most likely relates weakly to value priorities. In line with this view, Roccas et al. (2004) found that neuroticism had a weaker connection to values than the other Big Five personality traits. Guilt-proneness, as measured by the TOSCA, is a broader concept than empathy. Some guilt scenarios describe interpersonal situations and include empathy, but not all of them. Some scenarios concentrate on fullling ones obligations, for example, doing well at school or work. In general, morality can be seen as consisting of treating others kindly and justly, and of conforming to societal norms and rules. Therefore, both self-transcendence and conservation values were hypothesized to be signicant in terms of morality, and they were expected to relate positively to guilt-proneness. In contrast, self-enhancement and openness values highlight ones own interests and independence, which may conict with the interests of others and with societal norms. Therefore, these values were expected to relate negatively to guilt-proneness. Consistent with this view, Jaari (2004) found, in a Finnish adult sample, that universalism, benevolence, conformity, tradition, and security values were positively related to valuing honesty and respect of moral norms (a subscale of the Machiavellism scale by Christie, 1970). In contrast, self-direction, stimulation, hedonism, achievement, and power were negatively related to valuing honesty and respect of moral norms. Laypeoples view of moral values appears to be quite consistent with the above analysis. In a study by Schwartz (2005a), a sample of Israelis was asked which of the Schwartz values they considered moral values. Eighty percent of the respondents labeled all benevolence value items as moral values, and seventy percent of the respondents maintained that all or most of universalism, benevolence, conformity, tradition, and security value items were moral values. Conversely, value items reecting self-enhancement or openness to change were rarely considered moral. However, the cognitive measures of morality relate positively only to self-transcendence values, not to conservation values. Kohlbergs moral reasoning stages have typically been found to relate positively to universalism,

and sometimes also to benevolence and self-direction, but only negative associations have been found for conformity (for a review, see Helkama et al. 2003). Based on these ndings, we hypothesized that valuing conformity or tradition is relevant for constructs that emphasize conforming to norms or rules, whereas valuing benevolence or universalism is related to measures that emphasize independent thinking and understanding the perspectives of others. According to Helkama (2004), different values correspond to different functions of morality: conceptually, valuing conformity/tradition is related to the prevention of antisocial action, and benevolence relates to the promotion of prosocial behavior, whereas universalism, with its focus on justice, most closely relates to solving moral dilemmas. Consequently, all of these values represent important motives for different aspects of moral behavior. Although values are considered important in terms of morality, connections between values and other moral constructs have received surprisingly little attention. However, it could be important to understand how values relate to different personality dispositions, because this knowledge would be useful when discussing the values of organizations or communities. For example, what kinds of implications for students personality development would there be if school systems emphasized self-direction and ignored conformity and tradition values? It is likely that the values of different social groups affect the personalities of the individual members of those groups, but in what way? More research is needed to clarify this. The objective of this study is to examine the connections between value priorities and proneness to guilt, shame, and empathy. Based on the analysis presented above, we formed the following hypotheses. 1. Guilt-proneness will be positively related to selftranscendence and conservation values, and negatively related to openness and self-enhancement values. Empathic concern and perspective-taking will be positively related to self-transcendence values, and negatively related to self-enhancement values. Personal distress and shame-proneness will have a weaker relation to values than have guilt, empathic concern, and perspective-taking.

2.

3.

Method Participants The connections between values and emotional tendencies were studied in two different samples. The rst sample was comprised of adolescents. With respect to values, adolescents have been studied less than university students or

123

Motiv Emot (2008) 32:6980

73

adults, even though adolescence is usually seen as an important period for personality development (Damon 1983). Our sample of adolescents was collected in schools that require a high standard of academic performance from students. Therefore, the sample is not representative of the general Finnish adolescent population; it includes more girls than boys, and the students may represent a higher than average SES. In order to include more males in this study, we collected a second sample from adult conscripts at the Reserve Ofcer School. Military service is compulsory for men in Finland, and so this group should be more representative of the general male population than commonly used student samples. However, only those who perform especially well in military service are chosen for ofcer training, and consequently, this sample is also not a perfect representation of the general Finnish male population. The rst sample included 207 secondary school students, mainly ethnic Finns, 1519-years-old (68% girls, the mean age was 16), representing all grades of lukio (a 3-year school preparing students for university or other higher education). The data was collected in two schools in the Helsinki metropolitan area. Participants completed the questionnaires in class, in the presence of a teacher and the researcher, or in the presence of a teacher who had been instructed by the researcher, without the researcher in attendance. Participants were given the opportunity to ask questions about the questionnaire. Participation was voluntary, and no compensation was provided. No one refused to participate, but data were incomplete for some respondents. The missing values were excluded pairwise. At most, 11% of participants were excluded from the analysis due to incomplete data. The second sample consisted of conscripts at the Reserve Ofcer School in Hamina, Finland. They were scheduled to complete a number of personality measures. The study was conducted in collaboration with the Finnish Defence Forces Education Development Centre and approved by the Finnish Defence Forces Headquarters. A total of 514 (11 of whom were women, mean age 19.7 years) conscripts, from 697 possible participants, gave their informed consent and completed all of the questionnaires. However, we administered two versions of the questionnaire, and only half of the participants (N = 258, 3 women, mean age 19.6 years) received the TOSCA, while the other half receiving an unrelated questionnaire. Due to the small number of women, analyses were restricted only to men. Measures Dimensions of empathy were assessed using the Interpersonal Reactivity Index (Davis 1994), a 28-item test consisting of four seven-item subscales that measure

dimensions of empathy (empathic concern, personal distress, perspective-taking, and fantasy). For example, the scales include the following items: (a) I often have tender, concerned feelings for people less fortunate than me (empathic concern), (b) When I see someone who badly needs help in an emergency, I go to pieces (personal distress), (c) I believe that there are two sides to every question and I try to look at them both (perspective taking), and (d) I daydream and fantasize, with some regularity, about things that might happen to me (fantasy). Response choices ranged from 0 (does not describe me well) to 4 (describes me very well). The items were translated using back-translation procedure. Cronbachs alphas for the subscales were the following (rst value for the adolescent sample, second value for the conscript sample): empathic concern .76/.79, personal distress .74/ .65, perspective taking .72/.75, and fantasy .79/.80. The fantasy scale was not used in the analysis, because we had no theory-based hypothesis concerning this variable. Value priorities were measured using the Schwartz Value Survey (Schwartz 1992; Helkama et al. 1992). The survey contains 57 single value items, which are followed by a short explanation in parentheses (e.g., equality [equal opportunities for all]). Respondents rated the importance of each value item as a guiding principle in their lives on a 9-point scale. Examples include: -1 (opposed to my values); 0 (not at all important); 7 (of supreme importance). Ten values were formed from the single value items. To control individual differences in scale use, we used proportional sum variables; that is, the values scores were divided by the personal mean of all 57 values (see Verkasalo et al. 1996). The values scores, therefore, represent the relative importance to the person of each value type compared to the other value types, with the mean score for all 10 value scales being 1.00. See Tables 2 and 3 for descriptives. Reliabilities (Cronbachs alphas) of the values were the following (rst value for the adolescent sample, second value for the conscript sample): power .73/.69, achievement .71/.77, hedonism .72/.74, stimulation .82/.79, self-direction .53/.61, universalism .78/.74, benevolence .76/.76, tradition .59/.58, conformity .63/.68, and security .57/.57. The reliabilities fell within the ranges reported by Schwartz (2005b). The guilt and shame measure used in the present study, the TOSCA (the Test of Self-Conscious Affect, Tangney and Dearing 2002) has different versions for adolescents and adults. The adolescent version, the TOSCA-A, was used in the student sample, and the adult version, the TOSCA-3, in the conscript sample. The measure consists of scenarios designed to assess respondents shame, guilt, and defensive reactions. Each scenario is followed by four different responses, representing brief descriptions of shame, guilt, and defensive responses (externalization and

123

74

123
1 _ .42** .38** .23** .24** -.20** -.10 -.36** _ .32** .35** -.22** -.19* .01 .05 -.10 -.33** -.30** -.09 _ .20* -.24** -.24** -.24** -.29** -.20** _ .12 -.17* -.40** -.40** -.28** _ .18* -.15* -.18* -.22** _ -.04 .04 -.21** _ .42** -.02 _ -.17* -.34** -.30** .26** .31** .18* .21** -.01 -.15* .10 .25** -.05 -.08 -.22** -.18* .03 .05 -.07 .13 .07 .04 -.05 .03 .00 .32** .04 .25** -.43** .06 .26** .29** .32** -.53** -.44** -.26** -.13 -.18* -.02 .07 -.21** -.08 -.21* -.13 -.16* -.17* .06 .07 .18* -.19* -.18* -.15* -.03 -.09 -.19* -.12 .43** -.07 _ -.15* .02 -.29** -.21** -.29** _ -.06 .45** .00 _ .49** .15* _ .15* _ _ 2 3 4 5 6 7 8 9 10 11 12 13 14

Table 2 Intercorrelations between values, guilt, shame, empathic concern, personal distress and perspective-taking in the adolescent sample

Mean (sd)

1. TOSCA guilt

.85 (.56)

2. TOSCA shame

2.64 (.72)

3. IRI empathic concern

3.54 (.69)

4. IRI personal distress

2.80 (.63)

5. IRI perspective-taking

2.25 (.59)

6. Power

.40 (.32)

7. Achievement 8. Hedonism

.90 (.26) 1.16 (.34)

9. Stimulation

1.02 (.35)

10. Self-direction

1.12 (.23)

11. Universalism

1.08 (.25)

12. Benevolence

1.26 (.23)

13. Tradition

.52 (.24)

14. Conformity

.94 (.21)

15. Security

.98 (.22)

Note: N = 184204, missing cases were excluded pairwise

* p \ .05, ** p \ .01

Motiv Emot (2008) 32:6980

Motiv Emot (2008) 32:6980 .21**

75

Table 3 Intercorrelations between values, guilt, shame, empathic concern, personal distress and perspective-taking in the conscript sample

detachment), with respect to the specic context. The scenarios describe interpersonal situations, in most cases unintentionally harming a friend, or failure in achievement situations in school or work. The guilt items describe feeling bad about the behavior and willingness to repair the damage the behavior has caused, whereas the shame items include negative self-evaluations and motivation for avoidance. Externalization items describe avoiding responsibility for the situation, and detachment items refer to playing down the signicance of the event. For example, the scenario (from the TOSCA-3), At work, you wait until the last minute to plan a project, and it turns out badly is followed by four responses: (a) You would feel incompetent (shame); (b) You would think There are never enough hours in the day (externalization); (c) You would think Whats done is done (detachment); and (d) You would feel: I deserve to be reprimanded for mismanaging the project (guilt). Another example from the TOSCA-A involves an interpersonal situation: You make a mistake at school and nd out a classmate is blamed for the error, (a) I would think: The teacher does not like the classmate. (externalization); (b) I would think: Life is not fair. (detachment); (c) I would keep quiet and avoid the classmate. (shame); (d) I would feel unhappy and eager to correct the situation. (guilt). Respondents are requested to rate, on a 5-point scale, the likelihood of their responding in each manner indicated. The scenarios were translated using backtranslation procedure. Cronbachs alphas were, in the adolescent sample, 71 for guilt and .77 for shame, and in the conscript sample .76 for guilt and .77 for shame.

14

.20** .14** .17** .05 _ -.22** -.33** -.31** -.25** -.33** -.33** -.37** -.32** -.53** -.15 .16** -.03 -.07 .14** .24** _ -.29** .02 _ .20** -.35** .36** -.38** .20** .14** _ -.43** .41** .17** -.05 _ -.20** .32** -.11* -.23** -.04 -.10* -.05 _ -.08 .01 .05 -.09 _ _ _ .19** -.02 -.43** -.303** -.21** -.14 .10* -.06 -.25**

_ -.03 -.14* .07 -.02 -.08 -.12** -.25** -.25** .01 -.07

10

11

12

13

Results
.50** -.21** .37** .31** .06 .13** -.28** -.13** -.20** -.12** -.13*

Pearson correlation coefcients for the adolescent sample are shown in Table 2. As expected, guilt-proneness was positively related to universalism, benevolence, tradition, and conformity, and negatively related to power, hedonism, stimulation, and self-direction. The results for empathic concern were similar, but it was unrelated to conformity and tradition. As expected, shame and personal distress had fewer signicant correlations with values than guilt, empathic concern, or perspective-taking. We tested the integrated hypotheses that relate each moral emotion to the full set of values by correlating the predicted with the observed order of correlations between moral emotions and values. For empathic concern, personal distress, and perspective taking, the Spearman correlations between the observed and hypothesized order of correlations were .86 (p \ .01), .78 (p \ .01), and .98 (p \ .001). Similarly, for guilt, a Spearman correlation of .71 (p \ .05) conrmed the integrated hypothesis. However, for shame, the integrated hypothesis was not supported (r = .21, ns).

-.19**

.20**

.00 .21**

-.04

-.10

.01

.05

.00

.02

-.01

.01 Note: N = 503, for TOSCA guilt and shame N = 255 1.10 (.24) -.11

.37**

.30**

-.24**

-.22**

.27**

.22** .14* 1.27 (.30) .53 (.30)

Mean (sd)

-.08

-.13*

-.12*

.14*

3.43 (.58)

3.75 (.47)

3.44 (.48)

2.20 (.46)

2.45 (.48)

3. IRI empathic concern

4. IRI personal distress

5. IRI perspective-taking

1.01 (.32)

1.04 (.26)

1.01 (.27)

.53 (.45)

.91 (.42)

.96 (.26)

.17*

2. TOSCA shame

10. Self-direction

1. TOSCA guilt

7. Achievement

9. Stimulation

6. Power

15. Security

* p \ .05, ** p \ .01

11. Universalism

12. Benevolence 13. Tradition

14. Conformity

123

76

Motiv Emot (2008) 32:6980

The correlation coefcients for the conscript sample are shown in Table 3. The guilt-proneness variable was clearly skewed to a very high average. Therefore, logarithm transformation was used to improve normality of the distribution. The pattern of correlations was very similar to the pattern found in the adolescent sample. Furthermore, the ndings concerning the differences between guilt and shame and personal distress and other empathy-dimensions were replicated: shame and personal distress had weaker connections to values than guilt, empathic concern, or perspective-taking. The integrated hypotheses specied that the correlations between the moral emotions and the whole set of 10 values would follow the motivational circle of values. As in the adolescent sample, the integrated hypothesis was conrmed for guilt (r = .94, p \ .001), empathic concern (r = .90, p \ .001), and perspective taking (r = .79, p \ .01). However, contrary to the results for the adolescent sample, it was also conrmed for shame (r = .88, p \ .001), but rejected for personal distress (r = .47, ns). We used regression analyses to assess how strongly values predicted tendencies for guilt, shame, empathic concern, perspective-taking, and personal distress. However, we rst looked at the effects of gender. In the adolescent sample, gender was a signicant predictor for all the dependent variables, and girls scored higher than did boys. Girls also scored signicantly higher on universalism, F(1, 186) = 23.54, p \ .001, and benevolence F(1, 186) = 15.93, p \ .001, whereas boys scored signicantly higher in power, F(1, 186) = 17.10, p \ .001, achievement, F(1, 186) = 6.94, p \ .01, stimulation, F(1, 186) = 4.67, p \ .05, and security, F(1, 186) = 17.80, p \ .001. The associations between values and emotions differed somewhat between genders, but the differences were too small to demonstrate statistical signicance. Therefore, the effect of gender was controlled in the rst step of the hierarchical regression analyses. Neither grade level nor school was related to the dependent variables, and consequently these variables were not controlled. Two models were calculated for each dependent variable, so that the opposite values, having high negative correlations with each other, did not appear in the same model, and the problem of multicollinearity was avoided. The rst model included self-enhancement and openness values, whereas the second model included self-transcendence and conservation values. The percentages of variance accounted for by gender and by values for each dependent variable are presented in Table 4. In the conscript sample, there was no need to control any variables, and therefore, the regression analyses were performed entering the values in the rst step. The results of the regression analyses are presented in Table 5. In both samples, values explained more about the variance of guilt than of shame, according to hypothesis 3. In the adolescent sample, the results for the stronger

models were R2 = .32, F(6, 181) = 14.23, p \ .001 for guilt and R2 = .15, F(6, 181) = 5.24, p \ .001 for shame. In the conscript sample, the results were R2 = .13, F(5, 249) = 7.33, p \ .001 for guilt and R2 = .05, F(5, 249) = 2.75, p \ .05 for shame. Personal distress had a weaker relation to values than empathic concern or perspective-taking, as hypothesized. In the conscript sample, the results were R2 = .21, F(5, 497) = 26.71, p \ .001 for empathic concern, R2 = .14, F(5, 497) = 15.80, p \ .001 for perspective-taking, and R2 = .03, F(5, 497) = 3.48, p \ .01 for personal distress. In the adolescent sample, the differences are better observable in R2 change gures in Table 4, because these models include the effect of gender, but the results were R2 = .23, F(6, 180) = 9.40, p \ .001 for empathic concern, R2 = .14, F(6, 179) = 5.00, p \ .001 for perspective-taking and R2 = .18, F(6, 176) = 6.25, p \ .001 for personal distress.

Discussion Consistent with the rst hypothesis, in both samples, guilt-proneness in transgression situations was positively related to valuing universalism, benevolence, tradition, and conformity. In line with Schwartzs model (1992), in both samples, the opposite values of power, hedonism, stimulation, and self-direction were negatively related to guilt-proneness. However, achievement related negatively to guilt-proneness only in the conscript sample. It can be concluded that emphasizing ones personal interests or independence is related to low guilt-proneness, whereas valuing others well-being or the stability of the social system is related to high guilt-proneness. However, it must be remembered that, in some situations, self-direction may motivate moral behavior: not conforming may be the moral thing to do when societal norms can be considered immoral, e.g., racial inequality. Compared to guilt-proneness, shame-proneness had a weaker connection to values in both samples. In the adolescent sample, shame correlated negatively with hedonism and stimulation and positively with achievement, whereas, in the conscript sample, it correlated negatively with achievement and positively with tradition. Although there seems to be a weak connection between shame and the opennessconservation value dimension, in general, shame had a clearly weaker connection to values than guilt, which supports Tangneys view that guilt and shame-proneness, as measured by the TOSCA, are very different tendencies (Tangney and Dearing 2002). The TOSCA shame-proneness can be seen to reect the inability to control ones emotions, because it has been found to relate to several pathological emotional tendencies: depression, anxiety, and low self-esteem (Tangney and Dearing 2002).

123

Motiv Emot (2008) 32:6980 Table 4 Summary of regression analysis for variables predicting guilt, shame, empathic concern, personal distress, and perspective-taking in the adolescent sample

77

Guilt Model 1 1. Step Gender R (adjusted R ) 2. Step Gender Power Achievement Hedonism Stimulation Self-direction R2 change R2 (adjusted R2) Model 2 1. Step Gender R2(adjusted R2) 2. Step Gender Universalism Benevolence Tradition .19** .24** .25** .20** .17* .08 .19** .27 (.24) .28** .08 (.07) .17* -.23** -.02 -.23** -.17* -.26** .24** .32 (.30)
2 2

Shame

Empathic concern

Personal distress

Perspective-taking

.28** .08 (.07)

.22** .05 (.05) .23** -.01 .22* -.10 -.17* -.04 .10** .15 (.12)

.34** .11 (.11) .24** -.27** -.06 -.12 -.02 -.19** .12** .23 (.21)

.35** .12 (.11) .30** -.19* .09 -.04 -.08 -.04 .04 .16(.13)

.17* .03 (.02) .07 -.24** -.07 -.11 -.11 .05 .10** .13 (.10)

.22** .05 (.05) .25** .00 -.04 .16* -.02 -.02 .03 .08 (.05)

.34** .11 (.11) .21** .17* .23** .12 .05 -.08 .11** .22 (.20)

.35** .12 (.11) .29** .20** -.06 .13 -.05 -.06 .06* .17 (.15)

.17* .03 (.02) .01 .27** .18* -.02 .06 -.06 .11** .14 (.12)

Note: The values are standardized beta coefcients. N = 184204, missing cases were excluded pairwise * p \ .05, ** p \ .01 Table 5 Summary of regression analysis for variables predicting guilt, shame, empathic concern, personal distress, and perspective-taking in the conscript sample

Conformity Security R2 change R2(adjusted R2)

Guilt Model 1 Power Achievement Hedonism Stimulation Self-direction R2 (adjusted R2) Model 2 Universalism Benevolence .23** .14* .06 .13* -.07 .13 (11) -.15* -.11 -.14* -.01 -.06 .10 (.08)

Shame

Empathic concern

Personal distress

Perspective-taking

-.11 -.24** -.01 .02 -.05 .05 (.03) .06 -.04 .24** -.06 .04 .06 (.04)

-.20** -.09* -.11* -.12* -.08 .15 (.14) .32** .26** .00 .08* -.04 .23 (22)

.07 -.11* .12* -.11* -.09 .04 (.03) .02 -.07 .16** -.08 .00 .03 (.02)

-.17** .06 -.14** -.06 -.02 .07 (.06) .29** .10* -.09* .13** -.10* 14 (.14)

Note: The values are standardized beta coefcients. N = 503, for TOSCA guilt and shame N = 255 * p \ .05, ** p \ .01

Tradition Conformity Security R2 (adjusted R2)

Therefore, it is not surprising that shame was weakly related to values that are dened as conscious goals. Even if a person reports being more likely to react by feeling shame than guilt, it probably does not reect unwillingness to react

by feeling guilty, but more likely reects that the individuals own belief is that he or she would be unable to do so. The results for empathy and values were largely consistent with the hypotheses. As expected, the strongest

123

78

Motiv Emot (2008) 32:6980

connections were found in the self-transcendenceselfenhancement dimension. Universalism and benevolence were positively related to empathic concern and perspective-taking in both samples. However, as expected, values explained less variance in personal distress than in empathic concern or perspective-taking in both samples. Similar to shame, the tendency toward personal distress reects an inability to cope with emotions (Eisenberg 2000), and as an involuntary reaction, it is less connected to conscious value priorities than empathic concern or perspective-taking. The results were largely similar in both samples. For the most part, the relations were in the same direction. However, an interesting difference was that personal distress and empathic concern correlated positively in the adolescent sample, but negatively in the conscript sample. Consequently, in the adolescent sample, personal distress and empathic concern had similar correlations with values, although personal distress had a weaker connection to values than empathic concern. In the conscript sample, the correlations with values were clearly different for empathic concern and personal distress; personal distress was weakly but positively related to tradition and negatively related to self-direction, but it was unrelated to the self-transcendenceself-enhancement dimension. It is possible that empathic concern and personal distress become more differentiated with age, and for adolescents, these tendencies are more intertwined than for adults. Another possibility is that this difference is related to gender, because the conscript sample included only men. However, this seems unlikely, because in the adolescent sample, there were no statistically signicant gender differences in the correlations. It is clear that more research is needed before any conclusions can be made. An interesting nding was that self-direction was negatively related to both guilt and empathy, whereas universalism was clearly positively related to both of these emotional tendencies. This, to some extent, conicts with the Schwartz model, because self-direction and universalism are adjacent values, and therefore, their correlations with other variables should not differ to a signicant degree. The nature of the measured concepts can perhaps provide an explanation: proneness to guilt or empathy does not require a great deal of complex independent thinking. However, measures of cognitive complexity involving moral dilemmas have been found to relate positively to both universalism and self-direction (Helkama 2004). In the adolescent sample, gender differences were consistent with previous research (Davis 1994; Ferguson and Eyre 2000; Tangney and Dearing 2002; Schwartz and Rubel 2005). Girls scored higher than boys in guilt, shame, empathic concern, personal distress, and self-transcendence

values. Being caring and emotional is associated with the feminine gender role, and thus, girls may be more willing to describe themselves as such than are boys. However, it is possible that this difference appears on the behavioral level as well: if boys do not want to be perceived as caring and emotional, they may avoid expressing such emotions in their behavior (Bybee 1998). The samples included in the present study consisted of adolescents and young adults, whose degree of cognitive development can be expected to be sufcient to enable them to deliberate on their life goals. It is likely that, in these samples, there are bi-directional inuences between values and emotions. Thinking about certain values can elicit emotions, and having certain emotional tendencies can affect the way value priorities are chosen. The direction of causality may depend on the emotional tendency in question. For example, a person prone to neurotic tendencies, such as shame and personal distress, may prefer stability instead of excitement, because he or she is unable to cope with strong emotional stimuli. Consistent with this assumption, proneness to shame and personal distress correlated positively with tradition in the conscript sample and negatively with stimulation in the adolescent sample. An example of an opposite causal relation might be that someone who believes that conformity is important feels guilty over breaking societal norms, whereas someone who considers independence more important than conformity does not feel guilty in a similar situation. However, this connection was somewhat stronger in the adolescent sample than in the conscript sample. There are also some limitations that must be addressed. The participants in this research were mainly young ethnic Finns; therefore, the ndings of this study should be replicated in different samples in order to be generalized across different groups. The measure of guilt and shameproneness that was used in testing is based on a certain denition of these emotions, and there are types of guilt and shame that this measure does not cover. The TOSCA does not include chronic, unresolved guilt feelings that could perhaps have a similar relationship to values as has the TOSCA measure of shame. Furthermore, it does not include shame that leads to positive interpersonal behavior: shame is always assumed to be connected to the avoidance of responsibility. Only some of the possible contexts for feeling guilt and shame are included in the TOSCA; there are no scenarios in which the transgression would be intentional, or scenarios in which behavior harms a distant out-group instead of friends. In addition, due to the crosssectional nature of the data, it is not possible to infer the direction of causality between values and emotional tendencies. In spite of these limitations, the results were largely consistent with the hypotheses derived from previous

123

Motiv Emot (2008) 32:6980

79

theories and studies of values and proneness to guilt, shame, and empathy. Research on values has not yet focused on the connections between values and moral tendencies, and this study was planned to meet this need. Consistent with previous ndings, this study supported the view that the TOSCA guilt and shame scales measure different constructs that have different connections to other personality dispositions. Guilt was found to relate to value priorities more clearly than shame. Proneness to shame, as it is conceptualized in the TOSCA, seems to be an involuntary tendency that is not readily compatible with any motivational goals, similar to the Big Five neuroticism factor (Roccas et al. 2004). In a similar way, personal distress had a weaker relation to values than other empathy dimensions. Poorly managed shame and personal distress are undesirable reactions by nature. No one wants to experience them, regardless of which values one prefers. Therefore, it is possible that these emotional tendencies impede prosocial behavior even when a person considers it important. In summation, this study supported the view that self-transcendence values, universalism and benevolence, are important for prosocial tendencies, whereas valuing power, achievement, or hedonism appears to be to some extent incompatible with considering the well-being of others important. The results for the conservationopenness value dimension were less clear, but in general, selfdirection and stimulation appear to relate negatively to moral emotions, whereas conformity and tradition are unrelated or positively related to moral emotions, although the correlations were small, especially for empathy. Furthermore, self-direction has been found to relate positively to moral reasoning (Helkama et al. 2003). Still, there is evidence that prosocial reasoning has a weaker connection to prosocial behavior than has empathy (Hardy 2006), which may indicate that emotion is a more important motivational force in terms of moral behavior than is reasoning. In conclusion, conformity and tradition values seem more likely to support prosocial tendencies than self-direction and stimulation values. It is possible that conformity and tradition play a more important role in moral motivation in more traditional societies, whereas, in modern societies, morality is mainly based on universalism and benevolence values. More intercultural research is needed to increase our understanding of how values motivate moral behavior in different environments. Still, the essential message of this study is that in modern cultures like Finland, an environment that emphasizes self-enhancement values is likely to discourage prosocial behavior, whereas emphasizing self-transcendence values is likely to encourage it. The opennessconservation dimension seems to be less important, but still, conformity and tradition are more compatible with prosocial tendencies than are selfdirection and stimulation. In addition to determining which

values motivate moral behavior, more research is needed to clarify the developmental processes involving values, emotions, and moral behavior.
Acknowledgment This study was nancially supported by the Academy of Finland, grant number 210618.

References
Abell, E., & Gecas, V. (1997). Guilt, shame and family socialization: A retrospective study. Journal of Family Issues, 18, 99123. Averill, P. M., Diefenbach, G. J., Stanley, M. A., Breckenridge, J. K., & Lusby, B. (2002). Assessment of shame and guilt in a psychiatric sample: A comparison of two measures. Personality and Individual Differences, 32, 13651376. Barrett, K. C. (1998). The origins of guilt in early childhood. In J. Bybee (Ed.), Guilt and children (pp. 7590). San Diego: Academic Press. Batson, C. D. (1991). The altruism question: Toward a socialpsychological answer. Hillsdale: Erlbaum. Bybee, J. (1998). The emergence of gender differences in guilt during adolescence. In J. Bybee (Ed.), Guilt and children (pp.113125). San Diego: Academic Press. Covert, M. V., Tangney, J. P., Maddux, J. E., & Heleno, N. M. (2003). Shame-proneness, guilt-proneness and interpersonal problemsolving: A social cognitive analysis. Journal of Social and Clinical Psychology, 22, 112. Damon, W. (1983). Social and personality development: Infancy though adolescence. New York: Norton. Davis, M. H. (1994). Empathya social psychological approach. Boulder: Westview Press. Eisenberg, N. (1986). Altruistic emotion, cognition and behavior. Hillsdale: Erlbaum. Eisenberg, N. (2000). Emotion, regulation and moral development. Annual Review of Psychology, 51, 665697. Eisenberg, N., Zhou, Q., & Koller, S. (2001). Brazilian adolescents prosocial moral judgment and behavior: Relations to sympathy, perspective-taking, gender-role orientation and demographic characteristics. Child Development, 72, 518534. Ferguson, T. J., & Crowley, S. L. (1997). Measure for measure: A multitrait-multimethod analysis of guilt and shame. Journal of Personality Assessment, 69, 425441. Ferguson, T. J., & Eyre, H. L. (2000). Engendering gender differences in shame and guilt: Stereotypes, socialization, and situational pressures. In A. Fischer (Ed.), Gender and emotion: Social psychological perspectives (pp. 254276). Cambridge: Cambridge University Press. Fontaine, J. R. J., Luyten, P., De Boeck, P., & Corveleyn, J. (2001). The test of self-conscious affect: Internal structure, differential scales and relationships with long-term affects. European Journal of Personality, 15, 449464. Hardy, S. A. (2006). Identity, reasoning, and emotion: An empirical comparison of three sources of moral motivation. Motivation and Emotion, 30, 207215. Helkama, K. (2004). Values, role-taking and empathy in moral development. New Review of Social Psychology, 3, 103111. Helkama, K., Uutela, A., Pohjanheimo, E., Salminen, S., Koponen, A., & Rantanen-Vantsi, L. (2003). Moral reasoning and values in medical school: A longitudinal study in Finland. Scandinavian. Journal of Educational Research, 47, 399411. Helkama, K., Uutela, A., & Schwartz, S. (1992). Value systems and political cognition. In G. Breakwell (Ed.), Social psychology of

123

80 political and economic cognition (pp.732). Surrey: Academic Press/Surrey University Press. Hoffman, M. L. (2000). Empathy and moral developmentimplications for caring and justice. New York: Cambridge University Press. Jaari, A. (2004). Self-esteem, sense of coherence and values (In Finnish). Unpublished doctoral dissertation, University of Helsinki, Finland. Juujarvi, S. (2003). The ethic of care and its development. A longitudinal study among practical nursing, bachelor-degree social work and law enforcement students. Unpublished doctoral dissertation, University of Helsinki, Finland. Leith, K. P., & Baumeister, R. F. (1998). Empathy, shame, guilt and narratives of interpersonal conicts: Guilt-prone people are better at perspective-taking. Journal of Personality, 66, 137. Litvak-Miller, W., & McDougall, D. (1997). The structure of empathy during middle childhood and its relationship to prosocial behavior. Genetic, Social & General Psychology Monographs, 123, 303324. Lutwak, N., Panish, J. B., Ferrari, J. R., & Razzino, B. E. (2001). Shame and guilt and their relationship to positive expectations and anger expressiveness. Adolescence, 36, 641654. Luyten, P., Fontaine, J. R. J., & Corveleyn J. (2002). Does the Test of Self-Conscious Affect (TOSCA) measure maladaptive aspects of guilt and adaptive aspects of shame? An empirical investigation. Personality and Individual Differences, 33, 13731387. Mehrabian, A., & Epstein, N. (1972). A measure of emotional empathy. Journal of Personality, 40, 525543. Murphy, B. C., Shepard, S. A., Eisenberg, N., Fabes, R. A., & Guthrie, I. K. (1999). Contemporaneous and longitudinal relations of dispositional sympathy to emotionality, regulation, and social functioning. Journal of Early Adolescence, 19, 6697. Myyry, L., & Helkama, K. (2001). University students value priorities and emotional empathy. Educational Psychology, 21, 2540. Okun, M. A., Shepard, S. A., & Eisenberg, N. (2000). The relations of emotionality and regulation to dispositional empathy-related responding among volunteers-in-training. Personality and Individual Differences, 28, 367382. Padilla-Walker, L. M. (2007). Characteristics of motherchild interactions related to adolescents positive values and behaviors. Journal of Marriage and Family, 69, 675686.

Motiv Emot (2008) 32:6980 Rest, J. R. (1984). The major components of morality. In W. M. Kurtines & J. L. Gerwitz (Eds.), Morality, moral behaviour, and moral development (pp.2438). New York; Wiley. Roccas, S., Sagiv, L., Schwartz, S. H., & Knafo, A. (2004). The big ve personality factors and personal values. Personality and Social Psychology Bulletin, 28, 789801. Rokeach, M. (1973). The nature of human values. New York: Free Press. Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). The CAD triad hypothesis: A mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). Journal of Personality and Social Psychology, 76, 574586. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 165). New York: Academic Press. Schwartz, S. H. (2005a). Universalism values and the inclusiveness of our moral universe (in Finnish). In A.-M. Pirttila-Backman, M. Ahokas, L. Myyry, & S. Lahteenoja (Eds.), Arvot, moraali ja yhteiskunta [Values, Morality and Society] (pp.237257). Helsinki: University Press. Schwartz, S. H. (2005b). Robustness and fruitfulness of a theory of universals in individual human values. In A. Tamayo & J. B. Porto (Eds.), Valores e comportamento nas organizaco es [Values and behavior in organizations] (pp. 5695). Petropolis, Brazil: Vozes. Schwartz, S. H., & Rubel, T. (2005). Sex differences in value priorities: Cross-cultural and multimethod studies. Journal of Personality and Social Psychology, 89, 10101028. Tangney, J. P. (1998). How does guilt differ from shame? In J. Bybee (Ed.), Guilt and children (pp. 119). San Diego: Academic Press. Tangney, J. P., Burggraf, S. A., & Wagner, P. E. (1995). Shameproneness, guilt-proneness, and psychological symptoms. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions: The psychology of shame, guilt, embarrassment and pride (pp. 343 367). New York: Guilford. Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New York: Guilford. Verkasalo, M. V., Tuomivaara, P., & Lindeman, M. (1996). 15-Yearold pupils and their teachers values, and their beliefs about the values of an ideal pupil. Educational Psychology, 16, 3547.

123