Online ISSN : 2189-1338
Print ISSN : 0916-1503
ISSN-L : 0916-1503
Development of Japanese Measures of Reconciliatory Tendencies
The Japanese Trait Forgivingness Scale and the Japanese Proclivity to Apologize Measure
Yohsuke OhtsuboKazuho YamauraAyano Yagi
ジャーナル フリー HTML

2015 年 31 巻 2 号 p. 135-142


Reconciliation processes are influenced by two important dispositional variables: (i) the victim’s disposition to forgive the offender, and (ii) the offender’s disposition to apologize to the victim. We translated extant English measures of each of these dispositions into Japanese using the back-translation method. We then examined the validity of the two resultant measures, the Japanese Trait Forgivingness Scale (J-TFS) and the Japanese Proclivity to Apologize Measure (J-PAM). Consistent with previous findings, J-TFS scores were correlated with agreeableness, neuroticism (inversely), and subjective well-being, while J-PAM scores were correlated with agreeableness and subjective well-being. Interestingly, these two reconciliatory tendencies were positively correlated with each other, even when controlling for agreeableness (i.e., a preference for harmonious social relationships). In addition, two autobiographical recall studies (of actual instances of forgiveness and apology) confirmed the validity of these two measures. The J-TFS predicted the extent to which participants had forgiven a workplace offense inflicted by one of their co-workers, while the J-PAM predicted whether participants had apologized to their victims following a recent transgression.

Occasional disputes are inescapable even within close relationships. Reconciliation is important not only because close relationships are associated with various beneficial effects, such as increases in subjective well-being and decreases in the risk of physical and mental health problems (Argyle, 1987; House, Landis, & Umberson, 1988), but also because unforgiveness (harboring grudges) itself is a stressful state, that is detrimental to one’s health (Witvliet, Ludwig, & Vander Laan, 2001; Worthington, Witvliet, Pietrini, & Miller, 2007). Moreover, apology-making improves health outcomes by reducing self-condemnation and other negative feelings such as guilt and shame (Hall & Fincham, 2005, 2008). Since offenders’ apologies are a major determinant of their victims’ forgiveness, an integrative approach to research on forgiveness and apologies seems imperative (Ohtsubo, 2015). To this end, this research developed Japanese versions of the Trait Forgivingness Scale (TFS: Berry, Worthington, O’Connor, Parrott, & Wade, 2005) and the Proclivity to Apologize Measure (PAM: Howell, Dopko, Turowski, & Buro, 2011; Howell, Turowski, & Buro, 2012).

Forgiveness and Trait Forgivingness

Forgiveness (i.e., the act of forgiving someone) is defined as a psychological process involving a suite of motivational changes that culminate in positive, prosocial states (McCullough, Bono, & Root, 2007). For example, forgiveness entails the replacement of negative emotions with positive, other-regarding emotions (Worthington, 2005). Although forgiveness is strongly influenced by external/situational factors, such as the severity of a transgression and the presence/absence of an apology (see Fehr, Gelfand, & Nag, 2010, for a meta-analytic review), people also differ in their inherent disposition to forgive others (e.g., Berry et al., 2005; Mullet, Houdbine, Laumonier, & Girard, 1998). Roberts (1995) coined the term, forgivingness, to refer to such a forgiving disposition.

There are several scales designed to measure trait forgivingness: Some of them assume that trait forgivingness is a single-factor construct (e.g., Berry et al., 2005; Brown, 2003), while others assume it is a multi-factor construct (e.g., Kato & Taniguchi, 2009; Mullet et al., 1998). Of the various available measures, this study chose to focus on Berry et al.’s single-factor Trait Forgivingness Scale (TFS), which consists of 10 items. Carmody and Gordon’s (2011) autobiographical recall study recently showed that TFS scores predict the extent to which participants had forgiven a betrayal (e.g., infidelity, lying) by a relationship partner. Of note, acts of forgiveness in this study were measured by the Transgression-Related Interpersonal Motivations inventory (TRIM; McCullough, Rachal, Sandage, Worthington, Brown, & Hight, 1998), which is one of most frequently used measures of forgiveness. Therefore, Carmody and Gordon’s results speak to the validity of the TFS.

McCullough, Bellah, Kilpatrick, and Johnson (2001) theorized that forgiveness would be associated with two of the Big Five personality traits: agreeableness and neuroticism. Agreeableness is associated with prosocial orientation to others, and forgiving someone is an instance of a pro-social act. People high in neuroticism more readily experience negative affect in general, and anger toward someone’s offense in particular. Supporting McCullough et al.’s conjecture, empirical research has found that trait forgivingness is positively correlated with agreeableness and negatively correlated with neuroticism (Allemand, Job, Christen, & Keller, 2008; Berry et al., 2005; Brose, Rye, Lutz-Zois, & Ross, 2005; Steiner, Allemand, & McCullough, 2012). In addition, people high in trait forgivingness are happier and less depressive (Allemand, Hill, Ghaemmaghami, & Martin, 2012; Brown, 2003) possibly because they tend to have better interpersonal relationships (Lawler-Row & Piferi, 2006).

In our research, we developed a Japanese version of Berry et al.’s (2005) TFS, the Japanese Trait Forgivingness Scale (J-TFS; see Table 1).2) We first examined whether the J-TFS would be positively correlated with agreeableness and subjective well-being, and negatively correlated with neuroticism (Study 1). In addition, by conducting an autobiographical recall study (Study 2), we tested whether the J-TFS would predict participants’ level of forgiveness following a recent transgression by a co-worker.

Table 1 The Japanese Trait Forgivingness Scale (J-TFS) and the Results of the Confirmatory Factor Analysis (CFA) and the Exploratory Factor Analysis (EFA).
以下の文章を読んで、それぞれの文章にあなたがどれくらい賛成するか、あるいは反対するかを評定して下さい。(1=とても反対、2=やや反対、3=どちらとも言えない、4=やや賛成、5=とても賛成)Estimated Path Coefficients (CFA) (1-factor)Factor Loadings (2-factor EFA)
Japanese ItemsF1F2

“R” indicates the reverse-scoring items.

Proclivity to Apologize

Compared to trait forgivingness, individual differences in willingness to apologize for one’s wrongdoing have received limited attention. The Proclivity to Apologize Measure (PAM) is a recently developed measure of trait-level variation in the tendency to apologize (Howell et al., 2011, 2012). In developing the PAM, Howell et al. (2011) theorized that proclivity to apologize would be positively correlated with agreeableness and neuroticism, as these personality traits are correlated with guilt-proneness (Einstein & Lanning, 1998). Proclivity to apologize was also expected to be associated with better interpersonal relationships, and thus with high subjective well-being. In addition, it was predicted that the PAM would be positively correlated with caring for the victim’s well-being. These predictions were confirmed (Howell et al., 2011).

In the same manner as with the J-TFS, we developed a Japanese version of Howell et al.’s (2011) PAM, the Japanese Proclivity to Apologize Measure (J-PAM; see Table 2). We examined whether the above pattern of correlations would be replicable in a Japanese sample (Study 1). Furthermore, we asked the same participants to recall an incident of their own interpersonal wrongdoing and report whether they had apologized to their victim (reported as Study 3 in Ohtsubo & Yagi, 2015). Correlating participants’ J-PAM scores with their apology behaviors, we examined the validity of the J-PAM.

Table 2 The Japanese Proclivity to Apologize Measure (J-PAM) and the Result of the Confirmatory Factor Analysis (CFA).
次の文章にあなたはどれくらい賛成しますか。それとも反対しますか?(1=とても賛成、4=どちらとも言えない、7=とても反対)Estimated Path Coefficients (CFA) (1-factor)
Japanese Items

Study 1



Participants were 192 Japanese undergraduates (84 males and 108 females) who responded to an e-mail advertisement of this study that guaranteed a monetary reward of 1,000 yen in exchange for one hour of participation. Mean age (standard deviation, hereafter abbreviated as SD) of the participants was 19.42 years old (2.06). The relatively large SD was due to one 45-year-old participant: mean age (SD) was 19.29 (0.90) without this participant.


This study was conducted in combination with other unrelated studies. Participants first completed a questionnaire including various individual difference measures, and then engaged in several other tasks. The first questionnaire included the Newcastle Personality Assessor (NPA; Nettle, 2007), a two-item measure of subjective well-being, and the Moral Foundations Questionnaire (MFQ; Graham, Nosek, Haidt, Iyer, Koleva, & Ditto, 2011) along with the J-PAM (translated from Howell et al., 2011) and the J-TFS (translated from Berry et al., 2005). The MFQ was included to assay participants’ caring for other’s well-being. The Japanese versions of the NPA and MFQ, as well as the TFS and the PAM, were developed by the present authors using the back-translation method.

Trait forgivingness and proclivity to apologize

The questions that comprise the J-TFS and the J-PAM are reported in Tables 1 and 2. The J-TFS was accompanied by a 5-point scale (higher scores indicate higher trait forgivingness). The J-PAM was accompanied by a 7-point scale (higher scores indicate a greater proclivity to apologize). The internal consistency of the J-TFS was slightly low in Study 1: Cronbach’s α coefficient=.69 (but see Study 2). The J-PAM was associated with a reasonable level of internal consistency: Cronbach’s α coefficient=.78. The J-TFS and J-PAM scores were obtained by averaging the responses to the relevant item. Therefore, the theoretical range of the J-TFS score was 1 to 5, while that of the J-PAM score was 1 to 7.

Newcastle Personality Assessor and subjective well-being

The NPA is a twelve-item measure of the five-factor personality traits (with two or three items for each trait). The NPA was followed by two subjective well-being items (i.e., “In general, I consider myself a happy person,” “Compared to most of my peers, I consider myself happy”), which were adopted, with some modification, from Shimai, Otake, Utsuki, Ikemi, & Lyubomirsky’s (2004) Japanese Subjective Happiness Scale (SHS). The NPA and subjective well-being items were accompanied by a 5-point scale (1=“very uncharacteristic” to 5=“very characteristic”).

Cronbach’s α coefficients were reasonably high for neuroticism (.82) and subjective well-being (.85), but only modest for agreeableness (.44) and openness to experience (.58). We omitted extraversion (.27) and conscientiousness (.22) from the subsequent analyses due to their low reliability.

Moral Foundations Questionnaire

The MFQ was developed by Graham et al. (2011) to measure moral concerns in five moral domains. The primary interest of this study was its Care subscales, which measure respondents’ concern for others’ suffering and well-being. The MFQ has two sections. In the moral relevance section, respondents evaluate the moral relevance of several issues (e.g., “whether or not someone suffered emotionally”) from 0 (not at all relevant) to 5 (extremely relevant). In the moral judgment section, respondents rate their level of agreements with several contextualized moral judgments (e.g., “compassion for those who are suffering is the most crucial virtue”) from 0 (strongly disagree) to 5 (strongly agree). Each section consisted of three Care items. We averaged the responses to the six items to obtain the Care score (Cronbach’s α=.63).

Autobiographical recall of interpersonal transgression

Participants briefly described a recent instance of making someone angry. They rated the seriousness of their transgression on a 4-point scale and pre-transgression closeness to their victim using the Inclusion of Other in the Self scale (IOS scale: Aron, Aron, & Smollan, 1992). Participants were then asked whether they apologized to their victim. In this study, two participants did not recall any such events (N=190 for the analyses including this dichotomous apology-making variable).

Results and discussion

The descriptive statistics for the variables of interest, and the correlation matrix of those variables are shown in Table 3. The means (SD) of trait forgivingness were 2.98 (.58) for males and 2.93 (.61) for females, t(190)<1, ns (cf. the means for European American college students were 3.63 and 3.49 for males and females, respectively; the means for Asian American college students were 3.32 and 3.14 for males and females, respectively; Berry et al., 2005). The means (SD) of proclivity to apologize were 5.11 (.81) for males and 5.39 (.92) for females. Females were slightly more likely to apologize than males, t(190)=2.19, p=.030. The comparative mean for Canadian college students, not separated by sex, was 4.92 (Howell et al., 2011).3)

Table 3 Descriptive Statistics for the Variables of Interest and the Correlation Matrix of these Variables (N=192).
Mean(SD)1. J-TFS2. J-PAM3. A4. N5. O6. WB
1. J-TFS2.95(.57)
2. J-PAM5.26(.88).34***
3. Agreeableness3.66(.64).36***.41***
4. Neuroticism3.69(.93)−.33***−.02.07
5. Openness2.71(.95).12−.00.02.18*
6. Well-being3.65(.94).25***.21**.20**−.29***−.03
7. Care (MFQ)3.58(.65).10.26***.39***.08−.01.07

* p<.05, ** p<.01, *** p<.001.

The correlational pattern of the J-TFS was comparable with previous studies. Trait forgivingness was correlated positively with agreeableness and well-being, and negatively with neuroticism. The correlational pattern of the J-PAM partly deviated from the previous findings: the J-PAM was positively correlated with agreeableness, well-being and caring, but bore no correlation with neuroticism. In sum, the correlational pattern for the J-TFS and J-PAM was mostly comparable with previous studies. Interestingly, as shown in Table 1, J-TFS and J-PAM scores were positively correlated with each other (r=.34).

We finally examined whether the J-PAM would predict actual apology-making behavior. Proclivity to apologize was significantly correlated with the dichotomous measure of apology-making; the point-biserial correlation was .19, df=188, p=.008. The relatively weak correlation is understandable given that apology-making ought to be more strongly determined by external variables, such as transgression seriousness. A logistic regression analysis including the participants’ sex, closeness to the victim, self-rated transgression seriousness as well as proclivity to apologize confirmed this conjecture. Although the strongest determinant of apology was transgression seriousness, β=.77 (standard error=.24), p=.001, proclivity to apologize remained significant even after controlling for these variables, β=.42 (SE=.21), p=.043 (cf. the effects of sex and closeness were not significant). This result confirms the validity of the J-PAM.

Study 2


Study 2 was conducted as a part of a larger internet-based survey. Respondents were 310 employees (155 males and 155 females) of Japanese companies. Their ages ranged from 23 to 68 (mean=46.52 years old, SD=12.05). The respondents recalled an incident in which they were offended or upset by one of their co-workers (either a supervisor or a subordinate). Among other measures, their state forgiveness at the time of the survey was assessed by the TRIM,4) and their trait forgivingness was assessed by the J-TFS. The J-TFS was associated with a reasonable level of internal consistency in Study 2 (Cronbach’s α=.78). The TRIM consisted of 18 items and assessed motivations of revenge, avoidance, and benevolence. A recent study revealed that this measure has a uni-dimensional structure (McCullough, Luna, Berry, Tabak, & Bono, 2010). Therefore, we averaged the 18 items to obtain the single index of forgiveness (Cronbach’s α=.89).

Results and discussion

The mean (SD) of trait forgivingness were 2.89 (.47) for males and 2.84 (.62) for females, t(287.00)=0.73, ns (equal variance not assumed). A 2 (Study 1 [students] vs. Study 2 [company employees])×2 (sex) ANOVA revealed no significant effects of sample or sex, although the student sample (Study 1) was associated with a slightly higher level of forgivingness than the employee sample (Study 2), F(1, 498)=3.35, p=.068. We refrain from over-interpreting this marginally significant difference.

First, replicating the result of previous studies (Mullet et al., 1998; Steiner et al., 2012), trait forgivingness was positively correlated with age, r=.15, df=308, p=.007. Second and more importantly, trait forgivingness was positively correlated with state forgiveness (i.e., the extent to which the respondents had forgiven their co-worker’s offense), r=.27, df=308, p<.001. This J-TFS×TRIM correlation remained significant (β=.28, SE=.06, p<.001) after controlling for the effects of age and sex. The effects of age and sex were not significant (βage=−.05, βsex=.06). These results confirmed the validity of the J-TFS.

General Discussion

This research developed Japanese versions of two reconciliatory tendency measures: the J-TFS to measure the disposition to forgive others, and the J-PAM designed to measure the disposition to apologize. Consistent with previous studies, trait forgivingness was positively correlated with agreeableness and subjective well-being, and negatively with neuroticism among a Japanese sample. Moreover, the J-TFS predicted the respondents’ actual forgiveness at the workplace, which can be considered as strong evidence for the validity of the Japanese TFS.

The J-PAM was positively correlated with agreeableness, subjective well-being, as well as moral concerns of care. These correlations were consistent with previous findings. However, the J-PAM was not correlated with neuroticism. The lack of this correlation might be due to collectivism in Japan. According to Einstein and Lanning (1998), neuroticism is associated with anxiety-based guilt. Transgressors in collectivistic cultures, however, might be more prone to feel anxiety regardless of their levels of neuroticism, because their behaviors starkly contradict the collectivistic value of social harmony (Triandis, 1995). Future studies involving samples from individualistic and collectivistic countries are needed to test this possibility.

An interesting finding of the present research was the positive correlation between trait forgivingness and proclivity to apologize. This correlation remained significant after controlling for agreeableness. In other words, the observed correlation cannot be accounted for by a general tendency to value harmonious relationships. Instead, there seems to be a more specific disposition to peacefully resolve interpersonal conflicts. However, the present study employed the NPA, a truncated measure, to assess the participants’ personality traits. The author of this measure, Nettle (2007) himself recommended the use of longer measures to achieve greater validity. Therefore, whether the J-TFS×J-PAM correlation is robust and, if so, whether agreeableness, at least partly, accounts for this relationship should be investigated in future studies.

The present study did not examine whether trait forgivingness predicts health outcomes. One of the important reasons to study forgiveness/forgivingness is its relation to positive health outcomes (Worthington et al., 2007). Although we showed that both trait forgivingness and proclivity to apologize are positively correlated with subjective well-being, it is important to investigate whether they are also related to positive health outcomes in the Japanese population. Furthermore, it is important to study what mechanisms may mediate the relationship between reconciliatory tendencies and positive health outcomes (Lawler-Row & Piferi, 2006).

Peaceful reconciliation is obviously important in the realm of interpersonal relationships. Moreover, it may even have beneficial effects on both physical and mental health. The present research developed two Japanese tools to measure reconciliatory dispositions (trait forgivingness and proclivity to apologize). It is our hope that these measures will be used in the future research to cultivate our understanding of Japanese reconciliatory processes and their cross-cultural similarities and differences.

1)  We are grateful to Jack Berry, Andrew Howell, Michael McCullough, and Everett Worthington for their kind permissions to report the Japanese versions of their measures. We also thank Adam Smith and Molly Vallor for their invaluable assistance. This research was generously supported by grants from the John Templeton Foundation and the Japan Society for the Promotion of Science to Yohsuke Ohtsubo (JSPS KAKENHI Grant No. 26590132) and to Kazuho Yamaura (JSPS KAKENHI Grant No. 23730593). Some variables included in the questionnaire of Study 1 are not reported in the article. The full results including all variables are reported as supplementary materials of this article at the Kobe University Repository (http://www.lib.kobe-u.ac.jp/kernel/).

2)  There are already two Japanese measures of forgivingness, the Japanese Dispositional Forgiveness Scale (J-DFS; Ishikawa & Hamaguchi, 2007) and the Forgiveness of Others Scale (FOS; Kato & Taniguchi, 2009). The J-DFS includes items assessing self-forgiveness, as opposed to other-forgiveness, disposition. As a result, the DFS consists of 23 items, while the TFS consists of only 10 items. The FOS was designed to measure two aspects of forgivingness (the forgiving trait and the vengeful/grudge-holding trait), while TFS is a single-factor scale.

3)  A series of confirmatory factor analyses with the robust maximum likelihood estimation method were applied to the J-TFS and J-PAM in order to test whether the previously reported single-factor structure would be found in the Japanese versions. For the J-PAM, the single-factor structure fitted the data modestly well: χ2(20)=39.37, p=.006, the Comparative Fit Index (CFI)=.93, RMSEA=.071, SRMR=.050 (see Table 2 for the path coefficients). On the other hand, the single-factor structure did not fit the J-TFS data (Studies 1 and 2 combined): χ2(35)=256.56, p<.001, CFI=.70, RMSEA=.112, SRMR=.096 (see Table 1 for the path coefficients). An exploratory factor analysis (EFA) using the promax rotation suggests that it has a two-factor structure (see Table 1 for the factor loadings). However, the two-factor structure might reflect some response biases because all the standard items loaded on the first factor, and all the reverse-scoring items loaded on the second factor (cf. Kato & Taniguchi, 2009). Therefore, in this study, we decided to adhere to the single-factor norm prescribed by Berry et al.’s original study.

4)  We developed the Japanese version of TRIM using the back-translation method (see Appendix). There are two other Japanese versions of TRIM. One was developed by Hashimoto and Karasawa (2012). Examining the factor structure, they dropped seven items. Therefore, Hashimoto and Karasawa’s TRIM has 11 items. There is an 18-item Japanese TRIM developed by N. Takata, which can be found in Michael McCullough’s website (http://www.psy.miami.edu/faculty/mmccullough/). Not noticing Takata’s version, we independently developed our version. In our impression, the three versions are similar to each other, and any one of the three versions is suitable for future studies.

Table A1 The Japanese Transgression-Related Interpersonal Motivations (TRIM) Inventory.

A: Avoidance Motivations, R: Revenge Motivations, B: Benevolence Motivations (1次元の尺度として利用する場合は、Avoidance項目とRevenge項目の得点を逆転して、Benevolence項目の得点と合成)回答は5段階で測定(1=全く当てはまらない、2=当てはまらない、3=どちらともいえない、4=当てはまる、5=非常によく当てはまる)

© 2015 The Japanese Society of Social Psychology