Voluntary settlement and its consequences on predictors of happiness: the influence of initial cultural context

Hokkaido—a northern island of Japan that was settled by ethnic Japanese during the late nineteenth century and the early twentieth century—may remain to be a hybrid of interdependent culture of the mainland Japan and independent culture associated with frontier settlement. We thus anticipated that contemporary Hokkaido residents would exhibit either independent or interdependent psychological profiles depending on the types of behaviors that were required in a given situation. As expected, happiness was associated with positive disengaging emotions (e.g., pride in the self)—an independent profile—in situations that required personal goal pursuit and interpersonal influence; however, happiness was associated with positive engaging emotions (e.g., feelings of closeness)—an interdependent profile—in situations that required interpersonal harmony and adjustment. In contrast, such situational dependency was not observed for either mainland Japanese or Americans. For mainland Japanese happiness was associated with positive engaging emotions whereas for Americans happiness was associated with positive disengaging emotions.


INTRODUCTION
Over the last two decades substantial effort has been made to investigate cultural variations across broad regions that are glossed as East (encompassing much of Asia) and West (North America and Western Europe) (e.g., Markus and Kitayama, 1991;Nisbett et al., 2001). The growing body of evidence suggests that this broad distinction between East and West can serve as a reasonable first approximation in investigating cultural variations on the globe. It also suggests that independence vs. interdependence (or individualism vs. collectivism) (Triandis, 1989;Markus and Kitayama, 1991) is a key dimension to foster the cultural variations. Nevertheless, few investigations pay attention to a vast amount of variation that is likely to exist within each broad cultural area. Thus, concerted effort toward understanding within-culture variation is very much called for.
It is reassuring, then, that over the last several years researchers have examined a number of variables that may be linked to within-culture variations including social class (e.g., Snibbe and Markus, 2005;Kraus et al., 2009), power (e.g., Miyamoto and Wilken, 2010), residential mobility (e.g., Oishi, 2010), daily economic activities (e.g., Uskul et al., 2008), urbanism (e.g., Kashima et al., 2004;Yamagishi et al., 2012), and infectious disease prevalence (Fincher et al., 2008). One such variable that may prove to be particularly important in understanding the nature of new cultural forms when people migrate new lands of frontier and initiate their society is frontier settlement. In the present work, we will focus on Hokkaido-a northern island of Japan that was settled by ethnic Japanese during the second half of the nineteenth century and the first half of the twentieth and examine the hypothesis that contemporary culture of Hokkaido is a hybrid of culture of mainland Japan (which features interdependence) and culture of frontier (which is likely to be dominated by needs of, as well as desires for, independence).

HOKKAIDO: JAPAN'S NORTHERN FRONTIER
Hokkaido is located at the northern Japan, close to Russia, where an indigenous people called Ainu formed a hunter-gathering society. It was a wilderness area until the middle of the nineteenth century. After that, the new Meiji government sent numerous jobless Samurai warriors into Hokkaido to settle and take control over Hokkaido and prepare defense against a perceived threat from Russia. This governmental initiative was followed by the migration of numerous farmers and peasants, who voluntarily moved to Hokkaido to settle and to seek land. Immigration continued throughout the next several decades, transforming Hokkaido into an integral part of Japan with a population of nearly three million by the middle of the twentieth century. Thus, Hokkaido is similar to North American in terms of the history of voluntary settlement.
Voluntary settlement is a factor generating the ethos of independence, which consists of collective beliefs and practices of independent agency (Turner, 1920). According to the voluntary settlement hypothesis (Kitayama et al., 2009;Kitayama and Bowman, 2010), frontier ecological conditions are usually harsh and social institutions are often primitive; consequently, frontier settlers are more likely to develop a sense of independent agency, which is represented in orientation toward personal goal pursuits and personal choice. Supporting the hypothesis, Kitayama et al. (2009) demonstrated that North Americans were quite independent compared to Western European groups of the UK and Germany, assessed by several measurements including dispositional attribution and symbolic self-inflation. In contrast, mainland Japanese were less independent than North Americans and the two groups of Western European.
Similar effect of voluntary settlement would be expected in Hokkaido sharing the history of voluntary settlement with North America. Indeed, Kitayama et al. (2006a) demonstrated that Hokkaido-born Japanese in Hokkaido are more similar to North Americans than to mainland Japanese in terms of three indicators of independent agency (personal goal pursuits, personal choice, and the dispositional lay theory of social behavior). Moreover, Yamawaki (2012), which created an index of Japanese collectivism, showed that Hokkaido is the second lowest of the 47 prefectures of Japan in terms of the index. Only Tokyo, the largest metropolitan city of Japan, exceeded Hokkaido. This suggests that not only urbanism (e.g., Kashima et al., 2004;Yamagishi et al., 2012) but also the history of voluntary settlement would contribute to the regional variation in the level of collectivism.
However, North America and Hokkaido are very different in terms of the initial cultural context in which voluntary settlement occurred. While settlers of North America were mostly Western European origin, settlers of Hokkaido were ethnic Japanese. Voluntary settlement in Hokkaido is thus unique because it occurred in the context of the predominant mainland-Japanese ethos, which is embedded in a larger cultural context of interdependence. This raises an important research question whether in Hokkaido the idea of interdependence fostered by the predominant mainland-Japanese ethos is likely to coexist with the idea of independence generated by voluntary settlement.
Culture is conceptualized as a collective-level phenomenon that is composed of both socially shared meanings such as ideas and beliefs and scripted behavioral patterns of norms and practices (Kitayama and Uskul, 2011). It can evoke and guide different sets of psychological responses. If independence and interdependence coexist in Hokkaido, they would facilitate the corresponding psychological responses, that is, a tendency toward personal goal pursuit fostered by independence and a tendency toward achievement of interpersonal harmony fostered by interdependence. In addition, these psychological tendencies would emerge differently depending on social situations in which people are involved. As evidence related to such a situational influence, cultural priming studies for biculturally socialized individuals have shown that cultural knowledge of ideas and beliefs, which are activated by such "primed" linguistic symbols and icons, guides the corresponding psychological tendencies (e.g., Hong et al., 2000).
Unfortunately, the possibility on switching between the two tendencies in Hokkaido people was not tested in Kitayama et al. (2006a). For example, in their first study, because the participants were asked to describe the most emotional episode they had recently experienced, the types of emotional situations experienced were not manipulated. Thus, the result that as with happiness for the North American participants, happiness for the Hokkaido Japanese participants was distinctly more personal might appear because these Hokkaido Japanese participants were likely to incidentally remember emotional situations fostering the orientation toward the idea of independence. In the present research we examined the untested hypothesis that contemporary residents of Hokkaido would be "bi-cultural," in the sense that they show both tendencies toward personal goal pursuit and achievement of interpersonal harmony depending on the specific goals that are emphasized in given social situations, focusing on predictors of happiness.

CULTURAL DIFFERENCES IN PREDICTORS OF HAPPINESS
Past studies have shown that what predicts happiness differs across cultures, corresponding to the ideas of independence and interdependence (e.g., Oishi and Diener, 2001). Happiness is associated with personal achievement in Western cultures where the idea of independence is spread, whereas happiness is associated with social harmony in East Asian cultures where the idea of interdependence is dominant. For instance, Kitayama et al. (2000Kitayama et al. ( , 2006b demonstrated that the general positive emotional experiences (e.g., happiness) reported by North Americans are strongly associated with emotional experiences related to the accomplishment of independence (e.g., self-esteem and pride in self), which are known as disengaging positive emotions. In contrast, they demonstrated that the general positive emotional experiences reported by Japanese are strongly associated with emotional experiences related to an accomplishment of interdependence (e.g., respect and friendly feelings), which are termed engaging positive emotions.
The general patterns of North Americans and Japanese were replicated by Kitayama et al. (2006a). They asked four groups of participants (North Americans, Hokkaido-born Japanese in Hokkaido, mainland-born Japanese in Hokkaido, and mainland Japanese) to briefly describe the most emotional episode they had recently experienced and to indicate how strongly they had experienced a list of emotions presented. The reported strength of generally positive emotions was more strongly predicted by that of disengaging positive emotions than that of engaging positive emotions in North Americans, whereas it was more strongly predicted by that of engaging positive emotions than that of disengaging positive emotions in mainland Japanese. Moreover, the patterns of the two Hokkaido groups fell between those of North American and mainland Japanese groups. Although the reported strength of generally positive emotions was predicted by both disengaging and engaging positive emotions in Hokkaido-born Japanese in Hokkaido, their pattern was similar to that of North Americans, but different from that of mainland Japanese. While the reported strength of generally positive emotions was more strongly predicted by engaging positive emotions than disengaging positive emotions in mainland-born Japanese in Hokkaido, the association between generally positive emotions and disengaging positive emotions was significantly stronger compared to that in mainland Japanese. In addition, no significant difference between the patterns of the two Hokkaido groups was found.
These results suggest that the idea of independence, encouraged by the history of voluntary settlement, underlies predictors of one's happiness in Hokkaido. However, given the difference between North America and Hokkaido in cultural heritage where the settlement occurred, the cultural differences in predictors of happiness might be moderated by the types of behaviors that were required in a given situation.

THE PRESENT RESEARCH
We conducted two studies to test the predictions that the ideas of independence and interdependence coexist in contemporary resident of Hokkaido, and that whether happiness is better predicted by either disengaging positive emotions or engaging positive emotions depends on social situations they experience. In Study 1, we asked participants to report the intensity of experiencing emotions in a preselected set of situations, including personal and interpersonal behaviors. In Study 2, we asked participants to report how intensely they experienced a number of emotions in each of the influence and adjustment situations they remembered. In both studies, we examined the associations between general positive emotions and disengaging/engaging positive emotions for each of the types of situations.
Personal goal pursuit and interpersonal influence are emphasized when independent agency is highly valued (Morling et al., 2002). People are motivated toward behaviors related to personal goal and influence in social environments and situations where the idea of independence prevails. They will feel disengaging positive emotions such as self-esteem and pride by accomplishing a personal achievement and then experience happiness. On the other hand, interpersonal harmony and adjustment are emphasized when interdependence is highly valued (Morling et al., 2002). People are motivated toward behaviors related to adjustment and harmony in social environments and situations where the idea of interdependence is widespread. They will feel engaging positive emotions such as respect and friendly feelings by accomplishing an interpersonal goal including adjustment and harmony and then experience happiness. If happiness of contemporary Hokkaido residents depends on which of the ideas is emphasized in a given social situation, it may be likely to link to disengaging positive emotions in social situations where the idea of independence is highly valued (e.g., personal and influence situations). In contrast, their happiness may be likely to link to engaging positive emotions in social situations where the idea of interdependence is highly valued (e.g., interpersonal and adjustment situations).
As in Kitayama et al. (2006a), we tested groups of North Americans, Hokkaido-born Japanese in Hokkaido, mainlandborn Japanese in Hokkaido, and mainland Japanese. This enabled us to show cultural differences regarding the association between happiness and disengaging/engaging positive emotions that may be moderated by the types of situations. Given the previous findings by Morling et al. (2002), if behaviors related to influence and efficacy are highly valued in North Americans, the association between happiness and disengaging positive emotions may become particularly strong when a personal goal is achieved (e.g., influence situations). In contrast, if behaviors related to adjustment and harmony are highly valued in mainland Japanese, the association between happiness and engaging positive emotions may become particularly strong when an interpersonal goal is achieved (e.g., adjustment situations).
Moreover, Kitayama et al. (2006a) found that mainlandborn Japanese in Hokkaido was not statistically different from Hokkaido-born Japanese in Hokkaido in terms of the degree of an association between happiness and disengaging positive emotions and that of an association between happiness and engaging ones. They interpreted the result based on the possibility of selection bias. That is, a minority of mainland-born Japanese who are highly motivated toward personal goal pursuit might have psychological sympathy to and be attracted to Hokkaido. The possibility suggests that this group might represent contemporary settlers who moved voluntarily to Hokkaido. In contrast, differences on their heritage culture whether they were born and brought up either in Hokkaido or in mainland Japan might influence how solid their independent tendencies are. Thus, the similarity between the two Hokkaido groups might be replicated in the current study. At the same time, however, situational effects on personal goal pursuit might be weaker in mainlandborn Hokkaido Japanese compared to Hokkaido-born Hokkaido Japanese.
In sum, we hypothesized that (1) in Hokkaido-born Japanese in Hokkaido, happiness would be associated with disengaging positive emotions in social situations where the idea of independence is highly valued, whereas happiness would be associated with engaging positive emotions in social situations where the idea of interdependence is highly valued, (2) in North Americans, happiness would be associated with disengaging positive emotions, particularly in social situations where the idea of independence is highly valued, (3) in mainland Japanese, happiness would be associated with engaging positive emotions, particularly in social situations where the idea of interdependence is highly valued, and (4) while the pattern on predictors of happiness in mainland-born Japanese in Hokkaido might be similar to that in Hokkaido-born Japanese in Hokkaido, the situational influences might be weak in mainland-born Japanese in Hokkaido. We tested these hypotheses together. Following Simmons et al. (2012)'s recommendation, we report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures in the present research. females and 311 males) at Kyoto University 2 , and 59 students (33 females and 26 males, M age = 19.97, SD = 1.36) at the University of Michigan participated in the study for partial course credit or monetary compensation. The mainland-born Hokkaido Japanese participants' mean length of stay in Hokkaido was 5.31 months (SD = 11.93). This study was approved by the Ethics Committee of the Center for Experimental Research in Social Sciences at Hokkaido University. All participants filled out the Implicit Social Orientation Questionnaire (ISOQ) (Kitayama and Park, 2007;Kitayama et al., 2009). In the ISOQ, participants are presented with 6 situations related to personal behaviors and 4 situations related to interpersonal behaviors (see Table 1). Participants were asked to remember the last event pertaining to each of the 10 situations. Then, they were to report the extent to which they experienced a series of emotions in each of the situations on a 6-point rating scale (1 = not at all, 6 = very strongly). The emotions used were adopted from previous studies by Kitayama and Park (2007;Kitayama et al., 2009). Emotions were sampled from six theoretically derived categories. Some emotions were general in the sense that they could be associated with either independence or interdependence, both positive (happy, elated, and calm) and negative (unhappy). Others were socially disengaging, resulting from either success (proud and self-esteem) or failure with regard to independence (angry and frustration). Still others were socially engaging, stemming from either success (friendly feelings and close feelings) or failure with regard to interdependence (ashamed and guilty).

RESULTS AND DISCUSSION
The average intensity for each participant was computed for general positive emotions, disengaging positive emotions, and engaging positive emotions for each situation. To see situational influences in each culture, the mean standardized regression Table 1 | Ten situations involved in the ISOQ (Study 1).

Personal
Situation 1: When you thought about your appearance Situation 2: When you read a novel or book Situation 3: When you watched TV or listened to music Situation 4: When you got ill or injured Situation 5: When you were caught in a traffic jam Situation 6: When you were overloaded with work

Interpersonal
Situation 1: When you had positive interaction with friends Situation 2: When you had good interaction with a family member Situation 3: When something good happened to a family member of yours Situation 4: When you had a problem with a family member coefficients predicting happiness for disengaging and engaging positive emotions for each situation in each culture were computed. They are shown in Table 2. We then compared the degrees of the two standardized regression coefficients for each situation in each culture following a procedure explained in Steiger (1980). As shown in Table 2, regression coefficients for engaging positive emotion were significantly larger in mainland Japanese than those for disengaging positive emotion regardless of the situations, except for one situation (situation 1, personal behavior). In contrast, no significant difference was observed between the degrees of the two regression coefficients for North Americans, except for one situation. In the exceptional situation (situation 4, interpersonal behavior), the regression coefficient for disengaging positive emotion was significantly larger than that for engaging positive emotion. Notably, regression coefficients for engaging positive emotion in the two Hokkaido groups were significantly larger than those for disengaging positive emotion in some of the interpersonal situations. This tendency was similar to that in mainland Japanese. Unlike for mainland Japanese, however, for the two Hokkaido groups there was no difference in the degrees of the two regression coefficients for disengaging positive emotions and engaging positive emotions in the personal situations, except for one situation for mainland-born Japanese in Hokkaido (situation 5). Rather, this tendency was similar to that in North Americans 3 .
For each situation we also performed a multiple regression analysis including the culture and disengaging positive emotion and the culture and engaging positive emotion interactions to see cultural differences in the associations between happiness and disengaging positive emotions and between happiness and engaging positive emotions. The results are summarized in Table 3. Given that regression coefficients for engaging positive emotion were larger in mainland Japanese than those for disengaging positive emotion regardless of the situations, whereas no significant difference was observed between the degrees of the two regression coefficients for North Americans, it would be expected that the association between happiness and engaging positive emotions was greater in mainland Japanese than in North Americans. The results fit with the expectation in more than half of the situations (3 out of the 6 personal situations and 3 out of the 4 interpersonal situations including marginally significant interaction effects). Moreover, the association between happiness and disengaging 3 The average intensity for each participant was also computed for general negative emotions, disengaging negative emotions, and engaging negative emotions across situations. The average intensity for a generally negative emotion was regressed with the average intensity for disengaging negative emotion and that for engaging negative emotion for each situation in each culture. Overall, regardless of situations, the size of standardized regression was larger than 0 except for a few cases (for disengaging negative emotion, North Americans: 0.18 < βs < 0.76, Hokkaido-born Japanese in Hokkaido: −0.01 < βs < 0.60, mainland-born Japanese in Hokkaido: −0.04 < βs < 0.60, mainland Japanese: 0.36 < βs < 0.58; for engaging negative emotion, North Americans: −0.02 < βs < 0.64, Hokkaidoborn Japanese in Hokkaido: −0.03 < βs < 0.86, mainland-born Japanese in Hokkaido: −0.06 < βs < 0.65, mainland Japanese: 0.12 < βs < 0.37). A 2 (situation) × 4 (culture) × 2 (emotion) ANOVA on the mean standardized regression coefficients did not show any significant effects, p > 0.10.   positive emotions tended to be greater in North Americans than in mainland Japanese particularly in the interpersonal situations.
The patterns of the culture and disengaging positive emotion and the culture and engaging positive emotion interactions were equivocal when the two Hokkaido groups were compared with either North Americans or mainland Japanese. This may suggest that the patterns on the predictors of happiness in the two Hokkaido groups fell between those of North Americans and mainland Japanese. As indirect evidence, when Hokkaido Japanese were compared with North Americans, the culture and engaging positive emotion interaction became relatively negligible in the personal situations, whereas it was still significant in some of the interpersonal situations. In addition, the culture and disengaging positive emotion and the culture and engaging positive emotion interaction patterns were also unclear when the two Hokkaido groups were compared. Although cultural differences in the patterns on the predictors of happiness were equivocal, the results indicate that in Hokkaido-born Japanese in Hokkaido, a tendency that people feel happier as they experience more disengaging positive emotions is relatively strong in a situation where a personal goal is likely to be activated, whereas a tendency that people feel happier as they experience more engaging positive emotions becomes clear in a situation where an interpersonal goal is likely to be activated. Moreover, the pattern in mainland-born Japanese in Hokkaido was similar to that in Hokkaido-born Japanese in Hokkaido. The similarity between the two Hokkaido groups was thus replicated. Furthermore, happiness was more predicted by engaging positive emotions than disengaging positive emotions for mainland Japanese, whereas no difference between the two regression coefficients was found in North Americans. Interestingly, these patterns appeared regardless of the types of behaviors. This may suggest that the association between happiness and disengaging positive emotions in North Americans and the association between happiness and engaging positive emotions in mainland Japanese are relatively stable across situations and behaviors 4 .

STUDY 2
Study 1 provided initial evidence on situational influences in the predictors of happiness in Hokkaido. The purpose of Study 2 was to replicate this by testing the same four groups and using a different set of social behavior aspects. Although overall the results of Study 1 fitted with our expectations, some of them were unexpected. For example, the effect of disengaging positive emotions did not exceed that of engaging positive emotions in the personal situations either for North Americans or for Hokkaido-born 4 Although associations between disengaging/engaging positive emotions and happiness across individuals were analyzed in the main text, it is unclear whether similar patterns of association can be observed within each individual. To determine it, following a procedure explained in Kitayama et al. (2006b) and Kitayama et al. (2009), for each individual, we initially computed standardized regression coefficients predicting happiness for disengaging and engaging positive emotions over all the six behaviors of personal situations and all the four behaviors of interpersonal situations, respectively. The standardized regression coefficients were then averaged over the participants in the two situations in the four cultural groups. Kitayama et al. (2006b) mentioned that the two steps "were simultaneously estimated with an algorithm provided by the hierarchical linear models (HLM) analysis" (p. 895). One hundred and eighty-six participants (7 North Americans, 42 Hokkaido-born Japanese in Hokkaido, 35 mainland-born Japanese in Hokkaido, and 102 mainland Japanese) were excluded in the following analysis because regression coefficients predicting happiness could not be computed due to missing values and a variable having no variance. A 2 (situation) × 4 (culture) × 2 (emotion) ANOVA on the mean standardized regression coefficients showed a main effect of emotion. Overall, one's happiness was associated more with engaging positive emotions (M = 0.67) than with disengaging positive emotions (M = 0.16), F (1, 492) = 65.86, p < 0.0001. However, the effect was not qualified by either situation or culture, Fs < 1.5, ps > 0.20. The mean standardized regression coefficients in the two situations in the four cultural groups were as follows: For disengaging positive emotions in personal situations, disengaging positive emotions in interpersonal situations, engaging positive emotions in personal situations, and engaging positive emotions in interpersonal situations = 0. 28,0.27,0.54,and 0.64 in North Americans,0.09,0.07,0.74,0.20,0.22,0.59,and 0.67 in mainland-born Japanese in Hokkaido, and 0.14, 0.14, 0.56, and 0.81 in mainland Japanese. Japanese in Hokkaido. In addition, the expected association with happiness in North Americans and mainland Japanese was found not only in culturally matched situations (i.e., personal behaviors for North Americans and interpersonal behaviors for mainland Japanese), but also in culturally unmatched situations (i.e., interpersonal behaviors for North Americans and personal behaviors for mainland Japanese). To test our hypotheses again and further examine these findings, we focused on more culturally specific values in the second study: influence and adjustment.
In Study 2, participants were asked to think of situations in which they influenced or adjusted themselves to the surrounding people and to describe five situations that they had recently experienced. Given the results of Study 1, we expected that in Hokkaido-born Hokkaido Japanese, the association between happiness and disengaging positive emotions would emerge significantly in influence situations as in North Americans, whereas the association between happiness and engaging positive emotions would emerge significantly in adjustment situations as in mainland Japanese. Moreover, these tendencies in North Americans and mainland Japanese might be enhanced in culturally matched situations, compared to culturally unmatched situations. Furthermore, the pattern of mainland-born Hokkaido Japanese might not differ from that of Hokkaido-born Hokkaido Japanese.

METHOD
Thirty-seven Hokkaido-born Japanese students (14 females and 23 males, M age = 19.46, SD = 1.13) and 38 mainland-born Japanese students (10 females and 28 males, M age = 19.47, SD = 0.99) at Hokkaido University, 54 mainland-born Japanese students at Kyoto University (25 females and 29 males, M age = 20.03, SD = 2.21), and 47 North American students (36 females and 11 males, M age = 20.46, SD = 1.52) at the University of Michigan participated in the study for partial course credit or monetary compensation 5 . The mainland-born Hokkaido Japanese participants' mean length of stay in Hokkaido was 11.72 months (SD = 9.85). This study was approved by the Ethics Committee of the Center for Experimental Research in Social Sciences at Hokkaido University. About half of the participants in each of the cultures were asked to think of situations in which they had influenced the surrounding people and to describe five situations that they had recently experienced (influence condition). The remaining participants were asked to think of situations in which they had adjusted 5 The regions where mainland-born Japanese in Hokkaido and mainland Japanese were born and brought up were as follows: Tohoku (6), Kanto (19), Chubu (6), Kansai (5), Chugoku (1), and Kyushu (1) for mainlandborn Hokkaido Japanese, Tohoku (1), Kanto (3), Chubu (14), Kansai (32), Chugoku (1), Shikoku (1), and Kyushu (2)  ) out of 54 mainland Japanese were originally from metropolitan areas. The proportion of participants who were originally from metropolitan areas was not significantly different between the two groups, chi-square (1) = 0.08, p > 0.70. Moreover, the proportion of mainland-born Hokkaido Japanese participants who were originally from metropolitan areas in Study 2 was not significantly different from that in Study 1 (14 out of 68), chi-square (1) = 2.38, p > 0.10. themselves to the surrounding people and to describe five situations that they had recently experienced (adjustment condition). All the participants were then asked to indicate how strongly they experienced a series of emotions in each situation (1 = not at all, 6 = very strongly). The emotions used were identical to those in Study 1 6 .

RESULTS AND DISCUSSION
Because unlike in Study 1, participants asked to describe five different situations in each condition, the average intensity ratings of the five situations were computed for each participant for general positive emotions, disengaging positive emotions, and engaging positive emotions. Internal consistencies of each of the three emotion types calculated for each condition in each of the four cultural groups were reasonable (Cronbach's alphas for general positive emotions, disengaging positive emotions, and engaging positive emotions in the influence condition = 0.72, 0.80, and 0.80 in North Americans, 0.73, 0.68, and 0.87 in Hokkaido-born Japanese in Hokkaido, 0.60, 0.86, and 0.82 in mainland-born Japanese in Hokkaido, and 0.68, 0.84, and 0.91 in mainland Japanese; Cronbach's alphas for general positive emotions, disengaging positive emotions, and engaging positive emotions in the adjustment condition = 0.86, 0.85, and 0.76 in North Americans, 0.73, 0.68, and 0.87 in Hokkaido-born Japanese in Hokkaido, 0.72, 0.76, and 0.84 in mainland-born Japanese in Hokkaido, and 0.75, 0.71, and 0.90 in mainland Japanese). Although disengaging positive emotion and engaging positive emotion were highly correlated except for adjustment situations in Hokkaidoborn Japanese in Hokkaido and mainland Japanese, they were still separate (0.01 < rs < 0.78). In each culture, the average intensity of the generally positive emotion over the five situations was regressed with the corresponding average intensity of the disengaging positive emotion and the corresponding average intensity of the engaging positive emotion in each condition. The standardized regression coefficients for disengagement and engagement in the influence and adjustment conditions are plotted in Figures 1, 2 respectively.
In the influence condition, for North Americans happiness was significantly predicted by disengaging positive emotion, β = 0.49, t (21) = 2.50, p < 0.05, whereas the association between happiness and engaging positive emotion was relatively weak, β = 0.37, t (21) = 1.89, p < 0.10. This same tendency was found in the two Hokkaido Japanese groups. For Hokkaido-born Japanese in 6 Tsai et al. (2007) suggests that influence goals are associated with higharousal positive states (HAP), whereas adjustment goals are associated with low-arousal positive states (LAP). Among a series of emotions used in the current study, "elated" and "calm" are defined as HAP and LAP, respectively. Preliminary analysis showed a marginally significant interaction between type of situation (2: influence and adjustment) and type of positive state (elated and calm) on the average intensity reported by participants, F (1, 168) = 3.25, p = 0.073. Participants in the influence situation reported elated (M = 3.01, SD = 1.08) more strongly than calm (M = 2.72, SD = 0.90), t (168) = 2.71, p < 0.001, whereas the tendency disappeared in the adjustment situation (elated: M = 2.52, SD = 0.93, calm: M = 2.52, SD = 0.94). The affect-bysituation interaction was not qualified by a significant 3-way interaction with cultural group, F < 1. The situational differences on the intensity of experiencing the emotions suggest that the experimental manipulation was effective across cultures.
Finally, a multiple regression analysis including the culture and disengaging positive emotion and the culture and engaging positive emotion interactions was performed in each condition to see cultural differences in the associations of disengaging and engaging positive emotions with happiness. As shown in Table 4, neither the culture and disengaging positive emotion interaction nor the culture and engaging positive emotion interaction was significant due to relatively small sample size, except for a marginally significant interaction between culture and disengaging positive emotion when mainland Japanese and Hokkaido-born Hokkaido Japanese were compared in the influence condition. This effect indicates that the association between happiness and disengaging positive emotion for the Hokkaido-born Japanese in Hokkaido was somewhat greater than for the mainland Japanese.
Although cultural differences in the patterns on predictors of happiness were equivocal as in Study 1, Hokkaido-born Japanese in Hokkaido felt happier as they experienced more disengaging positive emotions when they influenced the surrounding people. The pattern was similar to that in North Americans. In contrast, they felt happier as they experienced more engaging positive emotions when they adjusted themselves to the surrounding people. The pattern was similar to that in mainland Japanese. In Hokkaido-born Japanese in Hokkaido, thus, the patterns of the predictors of happiness differed depending on whether they experienced either influence or adjustment situations.
As expected, for North Americans happiness was significantly predicted by disengaging positive emotions, whereas for mainland Japanese happiness was significantly predicted by engaging positive emotions. Moreover, as in Study 1, these tendencies emerged regardless of the types of situations. This may suggest that influence and efficacy are so powerful as cultural values in North America that they penetrate even adjustment situations, whereas adjustment and harmony as cultural values are so powerful in mainland Japanese culture that they penetrate even influence situations. 7 For exploratory purposes, we also performed a similar regression analysis on general negative emotions. In the three Japanese groups, except for one case (mainland-born Japanese in the influence situation), unhappiness was significantly predicted by disengaging negative emotion, 0.48 < βs < 0.79, ps < 0.01, whereas it was not predicted by engaging negative emotion, −0.20 < βs < 0.28, p > 0.10, regardless of the types of situations. The tendency was also reported in Kitayama et al. (2006a). In two cases (mainland-born Japanese in the influence situation and North Americans in the adjustment situation), neither disengaging negative emotion (βs = 0.27 and 0.35, ps > 0.10) and nor engaging negative emotion (βs = 0.30 and 0.34, ps > 0.10) predicted unhappiness. In the remaining one (North Americans in the influence situation), unhappiness was significantly predicted by engaging negative emotion, β = 0.78, t (21) = 5.34, p < 0.0001, whereas it was not predicted by disengaging negative emotion, β = 0.07, t (21) = 0.49. Furthermore, as in Study 1, the pattern of mainland-born Japanese in Hokkaido was similar to that of Hokkaido-born Japanese in Hokkaido. This is consistent with Kitayama et al. (2006a) showing that mainland-born Hokkaido Japanese were statistically no different from Hokkaido-born Hokkaido Japanese regarding the extent to which disengaging and engaging positive emotions predicted happiness.

GENERAL DISCUSSION
The present research focused on Hokkaido, a northern island of Japan that underwent a history of voluntary settlement, and tested the hypothesis that interdependent culture of mainland Japan and independent culture associated with frontier coexist there. Examining the predictors of happiness in North Americans, mainland Japanese, and Hokkaido-born and mainland-born Japanese in Hokkaido, two studies demonstrated that in Hokkaido-born Japanese in Hokkaido, happiness was likely to be associated with disengaging positive emotions encouraged by achieving a personal goal through personal and influence behaviors. The pattern was more similar to that in North Americans than that in mainland Japanese. In contrast, in Hokkaido-born Japanese in Hokkaido, happiness was more predicted by engaging positive emotions encouraged by achieving an interpersonal goal through interpersonal and adjustment behaviors. The pattern was more similar to that in mainland Japanese www.frontiersin.org November 2014 | Volume 5 | Article 1311 | 9 fostered normative beliefs valuing independence (e.g., autonomy and self-reliance); nevertheless, it did not accompany an intrinsic motivation to achieve independence. Given this, behaviors of Hokkaido-born Japanese in Hokkaido might be regulated by one's belief that it may be socially beneficial to continue with the norm valuing independence. The ethos of independence might be sustained to some extent accordingly as long as such a belief exists.
Finally, there are some limitations to the current study, which should be taken into account. First, we conducted the current study for only undergraduates in Japanese and American universities. It would be necessary to test participants in a nationwide sample to confirm the validity of the current findings and generalize the effect of voluntary settlement fostering the ethos of independence. Second, although we prepared multiple situations in both studies, the number is relatively small. Although a preliminary analysis was attempted (see Footnote 4), this prevented us from examining potential differences between tendencies toward independence and interdependence and the moderation effect of types of situations within each participant. Thus, the present research could not address clearly either whether individuals in Hokkaido would be happier when they experienced disengaging positive emotions in personal behaviors and influence situations or whether they would be happier when they experienced engaging positive emotions in interpersonal behaviors and adjustment situations. This warrants further study including more situations. Third, past studies have proposed that the ethos of independence is associated with several socio-environmental factors including social class (e.g., Snibbe and Markus, 2005;Kraus et al., 2009), power (e.g., Miyamoto and Wilken, 2010), residential mobility (e.g., Oishi, 2010), daily economic activities (e.g., Uskul et al., 2008), and urbanism (e.g., Kashima et al., 2004;Yamagishi et al., 2012). Although we tried to make potential influences of these factors minimum by testing relatively homogeneous groups of undergraduates, it is still unclear whether the association between voluntary settlement and the ethos of independence is found independently of these alternative factors. Further examination is needed to determine the extent to which these socio-environmental factors including voluntary settlement foster the ethos of independence.
To conclude, the present research provided initial evidence that the consequences of voluntary settlement differ across cultures valuing independence vs. interdependence. Further investigation is needed to overcome the limitations of the current study and elaborate the voluntary settlement hypothesis, which would further our understanding of the mechanisms of cultural change and persistence. We believe, however, that the current evidence clearly highlights the psychological processes that are historically path dependent in lands of frontiers and, thus, deserves particular attention.