Bretbel-Haurwitz,Marsb State-level indices of collectivism-individualism and Bureau.7)=.83(p=.005),p7)=83(p=.005).Thes religiosity were also examined because of suggested results support the existence of a strong positive relation- links between thes variable and prosocial be ship between regional-leve well-being and the preva ntrol for the p survey of individuals conducted from 1993 to 1996,in collected we also c which states were ranked according to respondents' donation data from 2010 only.Although this variable i ieiocRtcftndnitaliandcolectinibehaios inherently more unstable because it relies on a smalle 2011G99 001 Gallup Daily Tracking (Gallup,2012).in which respon- These correlations were again replicated for the nine dents were asked the extent to which religion was an broader geographic regions,()=.83 (p=.006),p(7)= 85 5).which reaffirmed the stro ach sta s.On hip bet biased the observed association between well-being fied as "very religious and altruism,we next estimated a regression model pre altruistic-donation rates from Analysis strategy th Several strategies were used to examine the association following covariates:median houschold income.Gin between altruistic kidney donation and well-being at the coefficient,percentage of non-Hispanic Whites,sex ratio state level.In addition to investigating biva ducational tions phys we acco identify r otential contributions of sociodemogranhi altruism.B=0.58.(40)=236.p=023 (see Table 2 for economic,health,and social influences,including state results).Well-being also remained a significant predictor 0 capita donation rates when 2011 religiosity was the primary units of ana we als d as a cov te, 59,39 .50,P -023 raphic regions of the United States as defined set of covariates collected by the U.s.census Bureau). in1and 20.well-bein again predicted altruism,B=0.48,(40)=2.25.p=.030 (see Table 3 for results). In this well-being Results rem r o sm whe 10. State-level rates of altruistic kidne donation ae.B=04239=216. 037 Tog these findings suggest that the relatic nship between well from0to 76,which translates to per capita lonation rates being and altruism is not a function of basic sociodemo 0.000029(LUa graphic regional variation in religiosity Well-Being Indes d from 61.7(West virginia)to redicts that im (Hawaii;see Fig.1). ments in objective well-being result in subjective changes We first examined the bivariate association between hat promote well-being outcomes, such as altru 2010 state-leve -be 0 nog and er capita donation rate 3,201 gges s tha ective w (48)=52.001 (see Fig.2).To account for possible exa mined how positive changes in obiective measures o high-leverage outliers (in particular,Utah's donation rate well-being,which are known to promote subjective well- was 4.55 SD above the so conducted a ra eing (Di et al. donations <.001.1hs7S ve c 100 201 After controlling for baseline (1999)median income broader geographic regions defined by the U.S.Census we found that increases in median income from 1999 to4 Brethel-Haurwitz, Marsh State-level indices of collectivism-individualism and religiosity were also examined because of suggested links between these variables and prosocial behavior (Henderson et al., 2003; Vandello & Cohen, 1999). The collectivism-individualism index was based on a national survey of individuals conducted from 1993 to 1996, in which states were ranked according to respondents’ endorsement of individualist and collectivist behaviors (Vandello & Cohen, 1999). The religiosity index was based on a 2011 Gallup survey, also a component of Gallup Daily Tracking (Gallup, 2012), in which respon￾dents were asked the extent to which religion was an important aspect of their lives and the frequency with which they attended religious services. On the basis of responses to these questions, each state’s religiosity score was calculated as the percentage of respondents classi￾fied as “very religious.” Analysis strategy Several strategies were used to examine the association between altruistic kidney donation and well-being at the state level. In addition to investigating bivariate correla￾tions between altruistic donations and well-being, we used multiple linear regression and mediation analyses to identify potential contributions of sociodemographic, economic, health, and social influences, including state￾level measures of collectivism and religiosity. The 50 states constituted the primary units of analysis. We also investigated associations at the regional level (i.e., the nine geographic regions of the United States as defined by the U.S. Census Bureau). Results State-level rates of altruistic kidney donation vary widely. Between 1999 and 2010, total donors per state ranged from 0 to 76, which translates to per capita donation rates ranging from 0 (Delaware, Mississippi) to 0.000029 (Utah; see Fig. 1). Scores from the 2010 Gallup-Healthways Well-Being Index ranged from 61.7 (West Virginia) to 71 (Hawaii; see Fig. 1). We first examined the bivariate association between 2010 state-level well-being and per capita donation rates from 1999 to 2010 and found a positive association between well-being and per capita altruistic donation, r(48) = .52, p < .001 (see Fig. 2). To account for possible high-leverage outliers (in particular, Utah’s donation rate was 4.55 SD above the mean), we also conducted a rank￾order correlation, which yielded comparable results, ρ(48) = .70, p < .001. This pattern was also replicated when state-level data were collapsed into the nine broader geographic regions defined by the U.S. Census Bureau, r(7) = .83 (p = .005), ρ(7) = .83 (p = .005). These results support the existence of a strong positive relation￾ship between regional-level well-being and the preva￾lence of extraordinary altruism. To control for the period during which the data were collected, we also conducted analyses using altruistic￾donation data from 2010 only. Although this variable is inherently more unstable because it relies on a smaller sample, the pattern of observed results remained very similar, r(48) = .42 (p = .003), ρ(48) = .45 (p = .001). These correlations were again replicated for the nine broader geographic regions, r(7) = .83 (p = .006), ρ(7) = .83 (p = .005), which reaffirmed the strong positive rela￾tionship between regional well-being and altruism. To control for confounding variables that may have biased the observed association between well-being and altruism, we next estimated a regression model pre￾dicting per capita altruistic-donation rates from 1999 to 2010 from subjective well-being, as measured by Gallup-Healthways, with the additional inclusion of the following covariates: median household income, Gini coefficient, percentage of non-Hispanic Whites, sex ratio, median age, educational attainment, and indices of poor mental and physical health. Even after we accounted for covariates, well-being remained a significant predictor of altruism, β = 0.58, t(40) = 2.36, p = .023 (see Table 2 for results).1 Well-being also remained a significant predictor of per capita donation rates when 2011 religiosity was added as a covariate, β = 0.59, t(39) = 2.36, p = .023. When we performed this regression using the same set of covariates collected in 1999 and 2000, well-being again predicted altruism, β = 0.48, t(40) = 2.25, p = .030 (see Table 3 for results). In this model, well-being remained a significant predictor of altruism when collec￾tivism (collected between 1993 and 1996) was added as a covariate, β = 0.42, t(39) = 2.16, p = .037. Together, these findings suggest that the relationship between well￾being and altruism is not a function of basic sociodemo￾graphic factors, including regional variation in religiosity or collectivism-individualism. The engine model of well-being predicts that improve￾ments in objective well-being result in subjective changes that promote well-being outcomes, such as altruism (Jayawickreme et al., 2012). This model suggests that subjective well-being mediates the relationship between objective well-being and altruism. To test this model, we examined how positive changes in objective measures of well-being, which are known to promote subjective well￾being (Diener et al., 2013), influence altruistic donations. We considered positive changes in both median income and health status across states from 1999 to 2010. After controlling for baseline (1999) median income, we found that increases in median income from 1999 to Downloaded from pss.sagepub.com by Cai Xing on February 11, 2014