The Relationship Between Life Satisfaction and Emotional Experience in 21 European Countries

Abstract

In the current study, the relationship between life satisfaction (LS) and the affective components of subjective well-being (SWB) was examined in a sample of 40,487 people across 21 European countries using data from the European Social Survey. After running multilevel confirmatory factor analyses in order to establish the measurement invariance of the constructs across the countries, the individual-level dataset was linked to available country-level aggregate personality traits, cultural values, and human development index (HDI). Results from hierarchical linear modeling (HLM) analysis showed that LS is best predicted by positive and negative affect (PA and NA, respectively), but may also be predicted by the degree of mixed emotions (ME). At the country level, national mean scores of Extraversion and Neuroticism moderated the relationship between LS and ME in different directions, whereas neither of the two personality traits had a significant impact on the relationship of LS to PA and NA. Survival/self-expression and the HDI ranking influenced the LS-PA and LS-ME relationships, whereas individualism/collectivism did not. Our research indicates that in addition to analyzing separate effects of NA and PA, it is also important to consider emotional complexity in SWB research, whereas these analyses need to take into account the moderating effect of cultural aspects, such as survival/self-expression values and countries’ level of development. Our findings also emphasize the importance of employing representative samples, as the age variance of participants can have a profound impact on results.

Keywords

life satisfaction positive emotions negative emotions mixed emotions Extraversion Neuroticism cultural values HDI

Subjective well-being (SWB) or “happiness” is composed of several separate elements (Tay & Diener, 2011). First, life satisfaction is a judgment process that seems to be in the “eye of the beholder.” The judgment of how satisfied people are with their present state of affairs is based on a comparison with a standard that each individual sets for him- or herself (Diener, Emmons, Larsen, & Griffin, 1985). Second, in addition to rational comparison and thought, SWB also has an affective side, which depends on the experience of different emotions, positive as well as negative. Life satisfaction (LS) has been shown to be clearly separable from positive and negative affect across different kinds of reports and methods (Lucas, Diener, & Suh, 1996). For lay individuals, it is evident that positive emotions enhance their satisfaction with life, whereas negative emotions have a detrimental effect on this. However, research has shown that the strength of the relationship between the cognitive and affective components of SWB depends on many different aspects, such as cultural values (Kuppens, Realo, & Diener, 2008), cognitive processes (Seidlitz, Wyer, & Diener, 1997), activated versus deactivated emotions (Yamasaki, Sasaki, Uchida, & Katsuma, 2011), and emotional intelligence (Schutte & Malouff, 2011); therefore, the exact nature of the interconnectedness requires a fuller investigation.

Individual differences in SWB probably do not solely originate in temperamental predispositions (Lykken & Tellegen, 1996; Weiss, Bates, & Luciano, 2008), nor are they explicitly tied to external factors, such as economic welfare (Diener & Biswas-Diener, 2002; Diener, Horowitz, & Emmons, 1985) or employment status (Lucas, Clark, Georgellis, & Diener, 2004). The latest research suggests that the truth is actually somewhere in between, emphasizing influences that derive from certain universal aspects of human nature. Tay and Diener (2011) have proposed that the levels of the different elements of SWB can be explained by the fulfillment of particular basic and psychological universal human needs. Given that society strongly influences people’s basic needs, which are in turn linked to LS (Tay & Diener, 2011), and that several cultural value dimensions moderate the relationships between the components of SWB (see Kuppens et al., 2008; Suh, Diener, Oishi, & Triandis, 1998), it is evident that the national or cultural context has an impact on SWB judgments.

In this study, we focus on the complex relationships subsumed under the umbrella term SWB. Our objective is to unravel the connection between the cognitive and affective components of SWB and to examine the universality of this relationship across different European countries. We aim to elaborate the relationship between LS and emotions more comprehensively than has been done in previous cross-cultural research (Diener, Oishi, & Lucas, 2003; Kang, Shaver, Sue, Min, & Jing, 2003; Kuppens et al., 2008; Pethtel & Chen, 2010; Suh et al., 1998; Wirtz, Chiu, Diener, & Oishi, 2009) by examining the effect of positive and negative emotions separately as well as in interaction; and besides exploring the country-level impact of cultural values and level of human development on the relation between LS and emotional experience, this study also comprises aggregate levels of personality traits.

Nature of the Hedonic Component of SWB

Positive or negative emotions?

For a long time, a large part of psychology has predominantly focused on the negative: on what can go wrong with people, on illness, negativity, deficit, and dysfunction (Larsen, 2000; Seligman & Csikszentmihalyi, 2000). However, in cases of SWB, it is far from truth―one of the hallmarks of research on SWB has been that the absence of negative factors alone is not enough to understand subjectively good life; positive measures need to be incorporated as well, as people require positive stimuli and do not simply avoid misery (Diener, Suh, Lucas, & Smith, 1999).

Positive (PA) and negative affect (NA) are not simply opposite poles on the same continuum (please note that in this study, the terms emotion and affect are used largely in the same meaning). Instead, they are hypothesized to be discriminable factors that, in some cases, appear to be only slightly correlated (Kööts, Realo, & Allik, 2012; Lucas et al., 1996; Watson, Clark, & Tellegen, 1988) but in other occasions may even relate moderately to strongly with each other (Gere & Schimmack, 2011; Russell & Carroll, 1999), depending on the research methodology and treatment of affect terms. Nevertheless, researchers mostly agree that the relationship between NA and PA is never a perfectly inverse one, and that is one of the reasons why they are also unequal in their contribution to LS. In the case of SWB, prior studies have shown that positive emotions predict LS judgments more strongly than negative experiences, and this has been found across a diverse set of nations (Cohn, Fredrickson, Brown, Mikels, & Conway, 2009; Kuppens et al., 2008; Lucas et al., 1996; Suh et al., 1998). Yet there have also been contradictory findings implying that negative affect is more pertinent to LS than positive affect (Pilcher, 1998). It is therefore necessary to take into account the affective experiences of both valences.

Mixed emotions (ME)

Although the linear impact of positive and negative emotions on LS judgments has been documented many times, there is an interesting and important issue that has gained almost no attention thus far. Namely, whether and how does emotional complexity or the nonlinear interaction between PA and NA influence the satisfaction with life? Considering emotional experiences in a complex interplay, rather than as simply separate entities, has several reasons. First, extremely high levels of happiness might not be a desirable goal, because very high levels of pleasant emotions, especially intense or aroused ones, may be mixed in their effects on subjective well-being (Oishi, Diener, & Lucas, 2007). In fact, constantly striving for higher levels of well-being could lead an individual to more risk-seeking and potentially dangerous behaviors such as illicit drug use and an abundance of sexual partners (Diener & Ryan, 2009).

Second, by experience sampling methods it has been shown that the within-person relationship between PA and NA can be either independent or positively or negatively correlated (see Kööts et al., 2012, for a review), which means that an individual can be high or low on PA, on NA, or both. According to Cheng’s (2006) affective endowment-contrast theory, the effect of PA on LS actually depends on the level of NA—the effect being stronger when NA is higher. Therefore, the total affective experience is larger than the sum of positive, negative, and neutral moments (Cheng, 2006). And third, recent data on expressive writing as emotion regulation strategy has literally shown that it is possible to find happiness in negative emotions: When people disclose a highly distressing problem in their lives and try to reappraise the negative emotions positively, their psychological well-being increases (North, Pai, Hixon, & Holahan, 2011). North et al. (2011) argue that accepting one’s negative emotions and then trying to seek out positives might be an optimal strategy for building happiness.

Furthermore, there exist cultural differences in the frequency of the experience of positive and negative feelings and whether these are experienced sequentially or simultaneously. So-called mixed emotions are, for example, more prevalent in East Asian than Western cultures (Schimmack, Oishi, & Diener, 2002), because Asians conceptualize the self in a dualistic manner and are more tolerant of contradiction (see Spencer-Rodgers, Peng, & Wang, 2010). Analyzing the role of mixed emotions in SWB cross-culturally is important, because it helps to explain whether and how emotional complexity relates to LS judgments.

Emotional complexity or the degree of mixed emotions has been calculated using different methods—for example, by simply multiplying NA × PA (Cheng, 2006) or using the negative acceleration model (NAM; Spencer-Rodgers et al., 2010), a more complicated formula. An alternative and a more intelligible approach would be to correlate specifically the individual’s positive and negative emotions. Thus, in this study we assess the individual correlations between reports of PA and NA (frequency of emotional states over the past week), derived from Asendorpf’s (1992) approach, which was originally designed to measure individual stability and change in personality test scores. For the sake of concision, we named this index Mixed Emotions (ME), albeit not studying the simultaneous, momentary co-occurrence of emotions. In short, one of our aims is to clarify whether higher complexity (experiencing both negative and positive emotions over a specific time frame) brings higher LS judgments, as could be expected on the basis of studies cited above.

Country-Level Influences

Twin studies have demonstrated a genetic basis for SWB (Lykken & Tellegen, 1996; Weiss et al., 2008), but considering the vast differences in LS across societies, culture is doubtlessly also an important source of influence. Although factors such as positive emotions, happiness, ambitions, and energy may be desirable all across Europe, certain societies may place different values on these. Judgments of LS may be grounded primarily on intrapsychic experiences in certain cultures, whereas they may be based more on social elements in others (Suh et al., 1998). The purpose of our multilevel study is to disentangle the umbrella term SWB at the level of the individual—to examine the relationships among its elements as well as within its affective component—bearing in mind, at the same time, the relevance of the cultural/national context at the country level.

Personality Dispositions

Aggregate national scores of personality traits might be relevant to a host of social, economic, and health outcomes (McCrae, Terracciano, Terracciano, & 79 Members of the Personality Profiles of Cultures Project, 2005). Research has shown that the distribution of personality traits is organized geographically, probably due to cultural differences and genetic influences. Specifically, European cultures differ systematically from Asian and African cultures, chiefly with respect to Extraversion and Openness to Experience scores, on which Europeans score higher. However, there are also differences within Europe. For example, Southern European cultures tend to score higher on Neuroticism than Northern European cultures (Allik & McCrae, 2004).

When exploring SWB, it is therefore also necessary to take into consideration stable dispositions, in addition to individual and cultural differences in emotional variability. In particular, Extraversion and Neuroticism have been found to be consistent predictors of SWB, having been found cross-culturally and by measuring with different instruments (cf., Schimmack, Radhakrishnan, Oishi, Dzokoto, & Ahadi, 2002). This is understandable since the fundamental core of Extraversion is positive emotions (Lucas, Diener, Grob, Suh, & Shao, 2000) and negative feelings are the glue that binds together neurotic dispositions (Markon, Krueger, & Watson, 2005). The influence of Extraversion and Neuroticism on LS judgments is thought to be mediated by the use of memories of past emotional experiences: The mediator model of Schimmack, Radhakrishnan et al. (2002) asserts that the influence of personality on LS depends on the weight of hedonic balance in LS judgments (Schimmack, Radhakrishnan et al., 2002). Furthermore, these two personality traits are likely to produce stable individual differences in the actual amount of pleasure and displeasure in people’s lives (Schimmack, Oishi et al., 2002). Also the functional or optimal levels of happiness might vary across individuals, depending on their personality dispositions (Oishi et al., 2007).

Bearing in mind the relevance of affective experience in SWB, the relationship between personality traits and LS (Schimmack, Radhakrishnan et al., 2002), and the geography of personality, another research question that requires deeper empirical examination emerges. Namely, do country-level personality traits explain cross-cultural differences in the relationships between the components of SWB? In this study, we expect to find that how the citizens of a country typically see and describe their personality is related to the size of impact of emotions on LS at the individual level—the rationale behind this premise is that it might create a sort of overall affective background to the appraisals. More specifically, we hypothesize that if people within a country or nation tend to see themselves as, for example, highly extraverted (cheerful, gregarious, friendly, etc.), the impact of additional pleasant emotions on LS would be weaker, because of the overall high degree of positive experiences and sociability. We expected to find the effect of aggregate personality traits, although these mean values do not characterize each participant individually (in our study, personality data had a different origin than that of LS and emotional experience), they instead provide a sense of a “typical” personality of a nation.

Cultural values

Culture influences subjective well-being in a direct and an indirect way (Schimmack, Oishi et al., 2002). In this article we will focus on the latter—how culture moderates the relationship between hedonic balance and LS. Data from 46 countries worldwide (Kuppens et al., 2008) clearly show that national culture moderates how strongly positive and negative experiences are related to judgment on LS. The maximization of pleasure and minimization of displeasure are not equally important in all nations but vary along with cultural values (Kuppens et al., 2008). Two value dimensions that have shown a tendency to be related to the influence of emotions on LS are Hofstede’s (2001) individualism/collectivism and Inglehart’s (1997; Inglehart & Welzel, 2005) survival/self-expression value dimensions (for a more comprehensive overview about these cultural values in the study of SWB, please see Kuppens et al., 2008). More specifically, their research showed that people living in individualistic countries have a higher vulnerability to negative emotions in making their LS judgments; and in nations that value self-expression, positive emotions have a stronger influence on LS than in survival-valuing countries, at least in younger people (Kuppens et al., 2008).

According to prior studies, these cultural values are related to wealth of nations (Hofstede, 2001; Inglehart & Baker, 2000), whereas material prosperity has been found to be correlated with SWB across cultures (Diener & Biswas-Diener, 2002). Therefore, in addition to aggregate personality traits and cultural value dimensions, we will also include the most well-known composite indicator of development and progress, the Human Development Index (HDI), as a country-level moderator in our analyses. The HDI, which could be regarded as a more objective measure than those describe above, is constructed on the basis of three important domains: a long and healthy life (life expectancy at birth), knowledge (adult literacy rate and gross enrolment ratio), and standard of living (gross domestic product per capita in purchasing power parities). All are attributed equal importance in the overall index. We expect that a higher HDI index value is related to higher mean LS in a given country. The main purpose of including HDI ranking as a country-level predictor is to parallel it to the more subjective cross-cultural measures as a comparison.

The Present Study

The aim of this research is to gain a systematic understanding of the emotional ingredients of SWB as well as an understanding of whether this subjective experience depends on different cultural contexts. The current study had two specific goals. First and similarly to authors of several previous studies (e.g., Cohn et al., 2009; Kuppens et al., 2008; Schimmack, 2003; Suh et al., 1998), we were interested in the relationship between LS and the hedonic component of SWB. However, the aim of our study was to gain a deeper understanding of the interplay between the emotional states in SWB. That is, besides predicting LS with positive and negative affective experiences, we additionally seek to find out whether emotional complexity—that is, the frequent experience of both negative and positive affect, which intuitively seem opposite, but are often not—plays a role in a person’s LS judgment. Our second goal was to analyze whether the nature of the relationships between the components of SWB varies across different countries or cultural groups. Furthermore, if there indeed is any cross-national difference, as was found by Kuppens et al. (2008) using college student samples, we aim to examine whether the countries’ overall levels of emotion-relevant personality traits, cultural values, or level of human development are related to such variance. In short, we will analyze the moderating influence of personality, cultural values, and human development levels on the within-country relationships between LS and emotional variables.

Our study differs from earlier research in several aspects. First, we analyzed data from nationally representative samples instead of college student samples (e.g., Kuppens et al., 2008; Suh et al., 1998). Such a data set enables us to examine variables across the lifespan and across distinct subsets of populations. Second, we measured LS and emotional experience by a set of indicators that was based on well-established research methodologies (see Huppert et al., n.d.). Third, our scale of positive emotions included the term happy, in contrast to the most widely used scale, the Positive and Negative Affect Schedule (PANAS). SWB measures should include happiness because mean levels of happiness are a better predictor of LS than other types of positive affect (Schimmack, 2003). Therefore, the most important strength of our study is that we use the European Social Survey database, which so far has not been sufficiently exploited in the context of SWB analysis and the relationships between its cognitive and affective components but embodies a good assessment of emotions.

Method

Participants

The analyses are based on the third wave of the European Social Survey (the ESS; www.europeansocialsurvey.org). The ESS is a biennial multicountry survey and the third wave was conducted in 2006. We used data for 21 of the 25 participating countries¹: Austria, Belgium, Bulgaria, Denmark, Estonia, Germany, Finland, France, Ireland, the Netherlands, Norway, Poland, Portugal, the Russian Federation, Slovakia, Slovenia, Spain, Sweden, Switzerland, the United Kingdom, and Ukraine. Demographic data for these countries are shown in Table 1. Sample sizes in each country ranged from 1,400 (in Bulgaria) to 2,916 (in Germany). Across the 21 countries, there were 40,487 participants in the ESS3 2006. The average response rate was approximately 63%, ranging from 46% in France to 73% in Slovakia (retrieved from http://ess.nsd.uib.no/ess/round3/deviations.html). The overall mean age of the participants was 48 (SD = 19) years, and approximately 54% of all individuals were females.

Table 1.

Sample Characteristics for 21 Countries in ESS Round 3

Country	N	Mean Age (SD)	% Women
Austria	2,405	45 (18)	54
Belgium	1,798	46 (19)	53
Bulgaria	1,400	50 (18)	61
Denmark	1,505	50 (18)	51
Estonia	1,517	47 (19)	56
Finland	1,896	49 (19)	52
France	1,986	48 (18)	53
Germany	2,916	48 (18)	51
Ireland	1,800	46 (18)	52
The Netherlands	1,889	49 (18)	54
Norway	1,750	46 (18)	49
Poland	1,721	44 (19)	53
Portugal	2,222	51 (19)	61
Russia	2,437	46 (19)	60
Slovakia	1,766	43 (18)	51
Slovenia	1,476	47 (19)	55
Spain	1,876	46 (19)	52
Sweden	1,927	47 (19)	51
Switzerland	1,804	50 (18)	55
United Kingdom	2,394	50 (19)	55
Ukraine	2,002	49 (19)	61
All Groups	40,487	48 (19)	54

The thoroughgoing ESS sampling procedures rely on random (probability) samples, with estimates based on the entire eligible residential populations aged 15 and over. Aiming for full coverage of the population, minimizing nonresponse, and considering design effects are prerequisites for the comparability of unbiased estimates, or at least prerequisites for bias minimization (see ESS homepage: Project Specification, Sampling Guide Round 3).

Materials and Procedure

The ESS “source questionnaire” was designed in English and then translated into other national languages. In countries in which any minority language is used as a first language by 5% or more of the population, the questionnaire was translated into that language too. Methodical translation procedures were used to ensure that each of these nationally organized translation efforts was informed by best practices and that the different language versions of the source questionnaire were functionally equivalent.

The ESS is an hour-long face-to-face interview made up of a core module (repeated from previous rounds) and two rotating modules (each devoted to a substantive topic or theme). In Round 3, one of these rotating modules also covered personal and social well-being.²

Person-Level Variables

Life Satisfaction (LS)

The ESS3 questionnaire contains eight questions that examine either satisfaction with life directly or satisfaction with different aspects of living. In order to assess LS specifically, we chose two of the theoretically most relevant items related to general satisfaction with life (other items were related to satisfaction with the country’s economics, government, and level of democracy, satisfaction with job or standard of living). One of the questions we used in our study is found in the core module (from Section B) and the other from one of the rotating modules (Section E). The items were as follows: “All things considered, how satisfied are you with your life as a whole nowadays?” (Question B24; ESS Round 3 2006) and “How satisfied are you with how your life has turned out so far?” (Question E31).³ (Please note that the origin of the adapted items in this paragraph is described in Huppert et al., n.d.) Questions were answered on an 11-point rating scale (0 = extremely dissatisfied/unhappy and 10 = extremely satisfied/happy). We calculated the average score of the two items in order to obtain an indicator of LS. Internal consistency of this two-item index was acceptable in each of the countries under study. Table 2 shows that the Cronbach’s alphas for this scale ranged from .60 to .79, whereas the median alpha over all the countries was .72.

Table 2.

Descriptive Statistics of Life Satisfaction and Affective Experience Scales

	Life Satisfaction		PA		NA		ME
Country	M (SD)	α	M (SD)	α	M (SD)	α	M (SD)*
Austria	7.39 (1.81)	.69	2.82 (0.65)	.73	1.52 (0.53)	.74	−0.43 (2.27)
Belgium	7.30 (1.61)	.71	2.78 (0.64)	.73	1.56 (0.55)	.78	−0.56 (2.57)
Bulgaria	4.83 (2.31)	.72	2.44 (0.70)	.83	1.81 (0.64)	.80	−1.28 (3.73)
Denmark	8.21 (1.39)	.70	2.91 (0.62)	.73	1.26 (0.40)	.74	−0.36 (1.73)
Estonia	6.48 (1.92)	.79	2.70 (0.63)	.77	1.65 (0.51)	.75	−0.41 (2.46)
Finland	7.89 (1.28)	.73	2.71 (0.60)	.78	1.30 (0.40)	.70	−0.11 (1.71)
France	6.61 (1.85)	.60	2.77 (0.66)	.75	1.62 (0.60)	.78	−0.75 (2.94)
Germany	6.82 (1.84)	.70	2.74 (0.59)	.72	1.39 (0.46)	.74	−0.18 (1.86)
Ireland	7.42 (1.75)	.73	2.85 (0.65)	.75	1.45 (0.47)	.72	−0.33 (1.82)
The Netherlands	7.32 (1.38)	.66	2.84 (0.58)	.74	1.52 (0.48)	.73	−0.25 (1.90)
Norway	7.66 (1.46)	.76	2.87 (0.56)	.71	1.26 (0.38)	.72	−0.13 (1.44)
Poland	6.86 (2.03)	.79	2.61 (0.69)	.76	1.64 (0.64)	.81	−1.01 (3.45)
Portugal	5.84 (1.70)	.65	2.56 (0.72)	.83	1.78 (0.59)	.77	−0.99 (3.12)
Russia	5.34 (2.27)	.76	2.56 (0.61)	.81	1.88 (0.62)	.81	−1.05 (3.20)
Slovakia	6.24 (1.90)	.73	2.73 (0.66)	.78	1.72 (0.56)	.77	−0.60 (2.42)
Slovenia	7.02 (1.78)	.71	2.84 (0.55)	.73	1.50 (0.51)	.78	−0.23 (1.97)
Spain	7.42 (1.52)	.77	2.67 (0.63)	.76	1.48 (0.53)	.80	−0.43 (2.48)
Sweden	7.88 (1.44)	.72	2.92 (0.59)	.73	1.48 (0.45)	.73	−0.21 (1.73)
Switzerland	7.74 (1.45)	.73	2.75 (0.66)	.79	1.37 (0.47)	.76	−0.45 (2.28)
United Kingdom	7.19 (1.79)	.76	2.69 (0.64)	.77	1.50 (0.53)	.76	−0.48 (2.44)
Ukraine	4.77 (2.30)	.71	2.64 (0.67)	.77	1.97 (0.70)	.78	−1.38 (3.98)
All	6.86 (2.01)	.77	2.73 (0.64)	.76	1.55 (0.56)	.79	−0.53 (2.54)

Note. PA = Positive Affect; NA = Negative Affect; ME = Mixed Emotions, i.e. individual-level correlations between positive and negative affect; *PA and NA were standardized across all countries.

Positive (PA) and Negative (NA) affective experience

In the ESS3 interview, participants were asked to assess the frequency of a total of 15 feelings and emotions they had experienced during the past week (e.g., “How much of the time during the past week you were happy?”). They rated the items on a 4-point scale (1 = none or almost none of the time and 4 = all or almost all of the time). Six items describe positive feelings, however two of these that represented cognitive and physical states (E19 “absorbed in what you were doing,” and E22 “felt really rested when you woke up in the morning”) were omitted from this study, because we were specifically interested in emotional experience. Therefore, PE was measured by four items: “happy” (Question E11), “enjoyed life” (Question E13), “had a lot of energy” (Question E16),⁴ and “felt calm and peaceful” (Question E20).⁵ (Please note that the origin of the adapted items given in this paragraph is described in Huppert et al., n.d.) Within-country internal consistency reliabilities (Cronbach’s alphas) ranged from .71 to .83 across the countries, with a median of .76 across all 21 countries.

In the original interview, the frequency of negative feelings was assessed by nine questions. We decided to leave out five items that are naturally related to negative mood but do not describe classical negative emotional experience (for example, E15 “could not get going” or E21 “felt bored”). This is in line with earlier research that has shown the underlying structure of the NE items in ESS3 (adapted mostly from CES-D; Radloff, 1977) to be considered as either one-dimensional or two-dimensional; the latter approach includes factors that can be described as measuring depressed affect and somatic complaints (Van de Velde, Bracke, Levecque, & Meuleman, 2010). The frequency of experiencing negative emotions was therefore measured by four items in this study: “depressed” (Question E8), “lonely” (Question E12), “sad” (Question E14), and “anxious” (Question E17).⁶ The internal consistency reliabilities (Cronbach’s alphas) of this scale ranged from .70 to .81, with a median of .76 across all countries.

Mixed Emotions (ME)

We used the scales of PA and NA to calculate another index for affective experience: namely, the index of ME. Asendorpf (1992) proposed a formula to measure individual stability over time. We used this formula to obtain individual contributions to the overall correlation between PA and NA. The formula was adjusted as follows:

ME = 1 - {(z_{PA} - z_{NA})}^{2} / 2

Here, z_PA and z_NA are the z-transformed scores for Positive and Negative Affect. The z scores were calculated across all countries; therefore, the population mean of ME is identical to the overall Pearson product moment correlation between the two emotion scales. The individual stability of a person is identical to the intra-individual variance of that person between the two z-transformed assessments subtracted from 1. This individual correlation practically measures an individual’s tolerance for experiencing both positive and negative emotions over a period of time. A high negative correlation indicates low tolerance (experiencing positive emotions excludes experiencing negative emotions and vice versa) and a positive correlation indicates higher tolerance for mixed emotions.

Country-Level Variables

Individualism/collectivism

We used in our study Hofstede’s (2001) index of individualism/collectivism. The index refers to the degree to which individuals are integrated into groups and ranges from 0 (most collectivistic) to 100 (most individualistic). Out of the 21 countries, the United Kingdom is the most individualistic (with an index of 89) and Portugal together with Slovenia are the most collectivistic (both have indices of 27).

Survival/self-expression

Scores of the survival/self-expression value dimension were based on the nation-level data from Waves 4 (in 2000) and 5 (in 2005) of the World Values Survey (WVS). The actual scores were taken from Inglehart and Welzel (2005). Higher scores reflect higher levels of self-expression. In our data, Sweden has the highest self-expression values (2.35), and survival is most valued in Russia (−1.42).

HDI ranking

The rankings were received from the Human Development Report 2007/2008 (n.d.), which reflects the time period of ESS Round 3 (2006). A higher number indicates lower ranking and lower human development; the smaller the number, the more developed a given country is. Most of the countries under study have high human development, except for Ukraine, which ranks into the category of medium human development (76th). Norway has the highest ranking of the HDI across the 21 ESS countries (2nd).

Personality Dispositions

We analyzed personality traits at the country level to find out whether the nation’s aggregate levels of Neuroticism and Extraversion influence the relationship between emotional experience and LS. According to the descriptive Five Factor Model of personality, Neuroticism is measured by six facets: anxiety, hostility, depression, self-consciousness, impulsiveness, and vulnerability to stress. Extraversion consists of facets such as warmth, gregariousness, assertiveness, activity, excitement seeking, and positive emotion (Costa & McCrae, 1992).

For Austria, Belgium, Denmark, Estonia, France, Germany, the Netherlands, Norway, Portugal, Russia, Spain, Sweden, and Switzerland, the NEO Personality Inventory (NEO PI-R; Costa & McCrae, 1992) composite factor self-reported personality T-scores were found in McCrae (2002). For Poland and Slovenia, the NEO PI-R observer ratings (composite factor personality T-scores) were taken from McCrae et al. (2005). The mean values of the NEO Five-Factor Inventory (NEO-FFI; Costa & McCrae, 1992), a short form of NEO PI-R, were used for Slovakia (Ruisel & Halama, 2007), the United Kingdom, and Finland (De Moor et al., 2012). The means were converted to T-scores using the U.S. norms (Costa & McCrae, 1992). No Big Five personality data were available for Ukraine, which is why we applied Russian personality data for this country instead. For historical, political, linguistic, and religion-related reasons (Sushko, 2008), we inferred that Ukraine’s aggregate Neuroticism and Extraversion would be similar to those of its eastern neighbor. We found no FFM personality data for Ireland, therefore the NEO PI-R observer ratings for Northern Ireland were applied (McCrae et al., 2005). We were unable to find any FFM personality data for Bulgaria. Table 3 shows that the aggregate T-scores of Extraversion range from 43.9 (the Netherlands) to 55.6 (Ireland), and the T-scores of Neuroticism range from 46.3 (Sweden) to 57.1 (Spain).

Table 3.

Country-Level Aggregate Variables for ESS Round 3 Countries: Personality Traits as T-scores, National Score of Cultural Values, and HDI Ranking

	Neuroticism	Extraversion	IDV	Survival–SE	HDI Ranking
Austria^a	52.9	48.4	55	1.43	15
Belgium^a	53.0	47.7	75	1.13	17
Bulgaria	—	—	30	−1.01	53
Denmark^a	46.5	52.8	74	1.87	14
Estonia^a	49.7	49.9	60	−1.19	44
Finland^c	47.1	47.5	63	1.12	11
France^a	55.4	47.3	71	1.13	10
Germany^a	52.8	47.3	67	0.74	22
Ireland^b	50.1	55.6	70	1.18	5
The Netherlands^a	48.6	43.9	80	1.39	9
Norway^a	47.4	53.6	69	2.17	2
Poland^b	50.7	49.2	60	−0.14	37
Portugal^a	55.5	46.3	27	0.49	29
Russia^a	53.7	45.1	39	−1.42	67
Slovakia^d	53.2	52.9	52	−0.43	42
Slovenia^b	50.6	49.5	27	0.36	27
Spain^a	57.1	48.3	51	0.54	13
Sweden^a	46.3	50.6	71	2.35	6
Switzerland^a	53.2	48.5	68	1.90	7
United Kingdom^c	47.4	48.8	89	1.68	16
Ukraine	53.7	45.1	—	-0.83	76

Note. Means were converted to T-scores using the U.S. norms (Costa & McCrae, 1992). For the Ukraine and Ireland the means of Russia and Northern Ireland were used, respectively. IDV = index of individualism/collectivism (Hofstede, 2001; Exhibits A5.1, A5.2, and A5.3); Survival–SE = Inglehart’s index of survival/self-expression (World Values Survey; Waves 4 and 5; Inglehart & Welzel, 2005); HDI = human development index (from the Human Development Report 2007/2008).

NEO PI-R self-report aggregate T-scores from McCrae (2002).

NEO PI-R observer ratings from McCrae, Terracciano et al. (2005).

NEO-FFI mean personality data from De Moor et al. (2012).

NEO-FFI mean personality data from Ruisel and Halama (2007).

Results

Measurement Invariance (MI)

Before comparing country-level scores, it is necessary to address whether the instrument measures the same construct across nations (Van de Vijver & Leung, 1997)

Self-report data in psychological research may yield unreliable results due to measurement biases, especially when data from several cultural groups are compared with each other (Milfont & Fischer, 2010). In order to make meaningful comparisons between countries, it is important to establish whether a certain degree of equivalence is met across the datasets (for a detailed approach to MI, see Byrne, Shavelson, & Muthén, 1989; Meredith, 1993; Steenkamp & Baumgartner, 1998; Vandenberg & Lance, 2000).

It is considered appropriate to analyze individual- and culture-level data simultaneously via multilevel models (cf., Cheung, Leung, & Au, 2006). Therefore, we decided to run a multilevel confirmatory factor analysis (MCFA), which allows investigation of the stability of a proposed factor model across countries by way of a refined analysis of the effects of measurement error at different levels (Cheung et al., 2006; Van de Vijver & Leung, 2000). The multilevel factor model is calculated for one population, with observations at two levels of aggregation (Heck & Thomas, 2000). Therefore, MCFA decomposes the total sample covariance matrix into pooled within- and between-group covariance matrices and uses these two matrices in the analyses of the factor structure at each level.

In order to conduct MCFA, we used the statistical software package for structural equation modeling, LISREL 8.80 for Windows (Jöreskog & Sörbom, 1999), which allows the fitting of latent variable models to two-level hierarchical multivariate data sets by using Full Information Maximum Likelihood (FIML) estimation. Common and necessary models that test relationships between variables measured and latent constructs are configural invariance (requiring factorial structures to be the same across levels), metric invariance (requiring factor loadings to be the same across levels), and scalar invariance (requiring item intercepts to be the same across levels) models. Optionally (and if theoretically meaningful), additional tests, such as the model of factor mean invariance, can be run (requiring latent means to be the same across levels) (Milfont & Fischer, 2010). In our study, the fit of data to the above mentioned invariance models was estimated by a goodness of fit index, the Root Mean Squared Error of Approximation (RMSEA; Steiger, 1990). Different cutoff points have been suggested: for example, a RMSEA value of 0.05 indicating good fit and 0.08 indicating reasonable or mediocre fit, whereas models with RMSEA over 0.1 should be rejected (MacCallum, Browne, & Sugawara, 1996).

We conducted two separate lines of MCFA analyses. First, we analyzed the multilevel relationships of the two satisfaction variables observed (satisfied with life and satisfied with how life has turned out so far) with the latent construct of LS. For the two-item scale of LS, a good fit for the scalar invariance (RMSEA = 0.01; χ² = 14.42, df = 7) and an acceptable fit for the factor mean invariance (RMSEA = 0.062; χ² = 1440.54, df = 19) models was established.

Next, we conducted MCFA for affective experience variables. More specifically, we examined the relationships of the four positive emotion variables observed (happy, enjoyed life, a lot of energy, and peaceful) with the latent PA and the four negative emotion variables (depressed, lonely, sad, and anxious) with the latent construct of NA. Here, we also found support for the scalar invariance model; RMSEA was 0.034 (χ² = 827.73, df = 35), indicating a good fit. A good fit for the factor mean invariance model was found as well (RMSEA = 0.046; χ² = 2343.67, df = 55). To sum up, we found that the item intercepts and factor means of LS, NA, and PA are invariant at the levels of within and between countries. Thus, the cross-level equivalence of the constructs being studied is supported, and we may make comparisons across the 21 ESS3 countries.

Life Satisfaction and Emotion Variables Across Countries

The mean scores of LS and the different components of emotional experience for each country under study are presented in Table 2. Ratings of LS ranged from 4.77 (in Ukraine) to 8.21 (in Denmark); mean LS across countries was 6.86 (SD = 2.01). The mean score for PA was 2.73 (SD = 0.64); mean ratings ranged from 2.44 (in Bulgaria) to 2.92 (in Switzerland). Mean NA scores ranged from 1.26 (in Norway and Denmark) to 1.97 (in the Ukraine); the mean for the negative emotion scale was 1.55 (SD = 0.56). The highest mean score of ME was in Ukraine (M = −1.38, SD = 3.98) and the smallest in Finland (M = −0.11, SD = 1.71)

There were statistically significant differences in the variances of all these variables across countries. The strongest differences between the 21 European countries are in the variance of LS (F = 550.84, df = 20, p < .001). The next largest cultural variances are for negative emotions (F = 251.01, df = 20, p < .001). Positive emotions have less variance across countries (F = 64.62, df = 20, p < .001); almost all countries report quite high average positive emotions. However, the smallest, yet statistically significant, variation across countries was found for the index of mixed emotions (F = 38.28, df = 20, p < .001).

At the level of individuals, the three variables of SWB were all significantly related to each other. The overall within-country correlation between LS and PA was r = .49 (p < .001); the relationship between LS and NA was about the same in size (r = −.51, p < .001); and the correlation between LS and ME was r = .31 (p < .001). The correlation between NA and PA across all individuals was r = −.55 (p < .001).

Multilevel Analysis

To assess whether the relationship between LS and emotional experience is different across European countries and if country-level personality traits moderate these differences, we conducted a series of multilevel linear modeling analyses. Using the hierarchical linear modeling technique (HLM 6.02; Raudenbush & Bryk, 2002), we modeled LS simultaneously at two levels—individuals (Level 1) nested within countries (Level 2). First, we aimed to find out how much of the overall variance in LS judgments is due to country-level differences. Therefore, we ran the simplest (unconditional) model without any predictors or moderators, which allows the calculation of the intraclass correlation coefficient (ICC; see Raudenbush & Bryk, 2002, p. 24, for the formula). The ICC measures the proportion of variance in an outcome that is between Level 2 units. Results indicated that approximately 15% of variance in LS is found between countries, meaning that LS judgments vary more substantively between individuals than the countries in which they live. Next, in order to find out if emotional experience explains some of the within-country variance of LS, we added three predictors to Level 1 (individual-level) of the model, all of them group-mean centered: (1) positive affect (PA), (2) negative affect (NA), and (3) mixed emotions (ME):

LS = β_{0} + β_{1} (PA) + β_{2} (NA) + β_{3} (ME) r_{ij}

In the Level 1 model, LS is the amount of LS for person i in the country. The intercept, β₀, is a random coefficient representing the mean LS for the country (across the individuals within it) when its Level 1 predictors are at the country’s mean level. A Level 1 coefficient β₁ (PA) is the within-country effect of positive affect on LS; β₂ (NA) is the effect of negative at the individual level; and β₃ (ME) is the effect of mixed emotions on LS. The error term, r, represents the unique effect associated with a specific person.

We then assessed if (and to what extent) these two variables helped to explain the large within-country differences in LS. For this purpose, we calculated an index of the proportion reduction in variance (see Bryk & Raudenbush, 2002, p. 79, for the formula). We found that adding PA, NA, and ME as predictors of LS explained about 30.6% of the remaining variance at the individual level. Table 4 shows that LS is predicted by within-country PA (β = 0.97, p < .001), NA (β = −0.75, p < .001), as well as ME (β = 0.05, p < .001).

Table 4.

Life Satisfaction Predicted by Positive and Negative Affect and Mixed Emotions, Moderated by Aggregate Personality Traits, Cultural Values, and HDI Ranking (Results From Hierarchical Linear Modeling)

	Fixed		Random
	Coefficient	T-Ratio	SD	χ²
For LS	7.107	62.53**	0.71	8531.57**
PA; slope (L1)	0.971	36.14**	0.14	126.96**
NA; slope (L1)	−0.746	−27.38**	0.11	64.82**
ME; slope (L1)	0.165	12.72**	0.02	62.55**
For LS			0.61	4,447.37**
Neuroticism (L2 moderator)	−0.104	−3.25*
Extraversion (L2 moderator)	0.045	1.46 (ns)
IDV (L2 moderator)	0.006	0.87 (ns)
Survival–SE (L2 moderator)	0.387	3.61**
HDI Rank (L2 moderator)	−0.037	108.76**
For PA slope			0.20	257.73**
Neuroticism (L2 moderator)	0.004	0.41 (ns)
Extraversion (L2 moderator)	−0.015	−1.19 (ns)
IDV (L2 moderator)	0.004	1.80 (p = .08)
Survival–SE (L2 moderator)	−0.154	−3.77**
HDI Rank (L2 moderator)	0.009	4.91**
For NA slope			0.15	107.73**
Neuroticism (L2 moderator)	0.004	0.50 (ns)
Extraversion (L2 moderator)	−0.005	−0.50 (ns)
IDV (L2 moderator)	−0.002	−1.16 (ns)
Survival–SE (L2 moderator)	0.005	0.15 (ns)
HDI Rank (L2 moderator)	0.000	0.05 (ns)
For ME slope			0.06	252.56**
Neuroticism (L2 moderator)	0.007	2.45*
Extraversion (L2 moderator)	−0.009	−2.96*
IDV (L2 moderator)	0.001	0.67 (ns)
Survival–SE (L2 moderator)	−0.046	−3.22*
HDI Rank (L2 moderator)	0.003	5.91**

Note. LS = life satisfaction; PA = positive affect; NA = negative affect; ME = mixed emotions (i.e., individual-level correlations between positive and negative affect); ns = non-significant; IDV = individualism/collectivism (Hofstede, 2001; Exhibits A5.1, A5.2, and A5.3); Survival–SE = survival/self-expression values (World Values Survey; Waves 4 and 5; Inglehart & Welzel, 2005); Level 1 (L1) predictors were added group-centered; Level 2 (L2) moderators were added grand-mean-centered.

p < .05. **p < .001.

In the subsequent analyses we examined the influence of country-level variables on the individual-level relationship between LS and emotion variables to find out if aggregate personality dispositions, cultural value dimensions, or the HDI ranking help to explain the cross-national differences in the connections within SWB. Therefore, we added the country-level aggregate Extraversion and Neuroticism scores, individualism/collectivism, survival/self-expression values, and the HDI ranking of each country to Level 2 of the model (while Level 1 of the model remained unchanged). The between-country-level moderators were added as grand-mean centered and separately in consecutive models. The Level 2 models were as follows (only the intercept function is shown below, but moderators were similarly added to both of the slope functions as well):

β_{0} = γ_{00} + γ_{01} (aggregate personality / cultural values / HDI) + u_{0}

Here, γ₀₀ represents the grand mean of PA / NA / ME. The coefficient β₀ was modeled as a random effect; the random error term, u₀, represents the unique effect of personality traits (or cultural values or the HDI ranking) on the affective variables.

Our results indicate that the relationship between LS and NA is not moderated by any of the Level 2 indicators (see Table 4). However, The LS-PA relationship is negatively moderated by the cultural value of survival/self-expression (γ = −.154, p < .001) and positively by the HDI ranking number (γ = .009, p < .001), meaning that the relationship between LS and PA is weaker in countries with higher self-expression values and human development level.⁷

The moderation by individualism is insignificant, but there is a tendency toward positive influence (γ = .004, p = .08). Aggregate personality traits did not moderate the connection between LS and PA.

We found that personality moderated significantly, albeit weakly, only the relationship between LS and ME. More specifically, higher aggregate Neuroticism had a positive effect on the relationship between LS and ME (γ = .007, p < .05); the moderating influence of Extraversion was negative and similar in size (γ = −.009, p < .05). Furthermore, it was found that the LS-ME relationship was moderated by the survival/self-expression value dimension (γ = ™.046, p < .05) and the HDI ranking (γ = .003, p < .001). Therefore, the positive relationship between LS and mixed emotions is weaker in countries with higher self-expression values, aggregate Extraversion and human development level, but stronger in countries with higher aggregate Neuroticism. Only individualism/collectivism had no significant influence.

Discussion

In line with our expectations, we found that cognitive judgments of LS are in some part predicted by affective experience—not only by unipolar positive and negative emotion measures but also by the interplay between these two affects. Our results confirm those of many previous studies showing that people who report more positive emotions evaluate their satisfaction with life more highly, whereas experiencing more negative affect predicts lower LS. The finding that PA significantly influences LS judgments is consistent with existing knowledge about positive emotions: they are known to have a beneficial effect on various life outcomes, promoting engagement with others, cooperation among individuals and groups, and positive intimate relationships (cf., Caprara & Steca, 2006). According to the broaden-and-build theory, positive emotions promote valued outcomes such as health, wealth, and longevity because they help build the resources to get there (Fredrickson, 2001). It is possible that happy people become more satisfied with life not simply because they feel better but because they develop resources for living well (Cohn et al., 2009).

Earlier research has demonstrated that LS has actually a stronger connection to positive emotions than negative ones (Kuppens et al., 2008; Lucas et al., 1996). However, in our study this effect was not explicitly confirmed, because the weights of PA and NA were about equal (although there was a slight tendency toward it). It means that for enhancing LS the deficiency of negative emotions is practically as important as frequent pleasant feelings. One possible explanation to this discrepant finding could be that the NA items in ESS3 focus on depression (most items were adapted from the Centre for Epidemiologic Studies Depression Scale, CES-D; Radloff, 1977), whereas depression was, in a previous study, found to be a stronger predictor of LS than other negative affects, such as anger or anxiety (Schimmack, Oishi, Furr, & Funder, 2004).

Next, we went deeper than previous studies in examining the relationship between positive and negative emotions in predicting LS. We were interested in the impact of their non-linear interaction to LS judgments, which is why we correlated the positive and negative emotions for each person individually, resulting in an index of mixed emotions. This index measures individuals’ tolerance for experiencing both positive and negative emotions within a specific timeframe (though not at the same moment in time, as would be the case for experience sampling methodology). Our results indicated that the more an individual is inclined to experiencing both types of emotion within a time period, the higher his or her LS. This is concordant with the affective endowment-contrast theory (Cheng, 2006), which asserts that negative emotions make the effect of positive emotions on LS stronger. We agree with Oishi et al. (2007), who have argued that there is more to psychological well-being than just high levels of happiness. Even though the influence of ME was smaller than in the case of PA and NA analyzed separately, it nevertheless demonstrated that, unlike the stoic belief that passions are detrimental to happiness, it is an intense and multifaceted emotional life that makes people satisfied with their lives. It is important to recognize the benefits of high subjective well-being for individuals and societies, but nevertheless it is a mistake to think that constant euphoria is a desired outcome (Diener & Ryan, 2009).

In this study, we found some variance in mean LS judgments across the 21 countries. This implied that the society where a person lives does impact well-being. The highest LS scores were found for Denmark, Finland, Switzerland, Sweden, and Norway; according to the HDI (2007/ 2008 report covering the period up to 2006), all five of these countries rank in the Top 14 in the world. This means that these countries have very high life expectancy, level of education, and standard of living. It makes sense that people in the most “livable” countries are also the most satisfied with their lives; however, while well-being may result from a structurally sound society, high levels of subjective well-being can contribute toward a more stable, productive, and effectively functioning society as well (Diener & Ryan, 2009). In our research, we also found cross-national variability in how strongly emotional experience impacts LS. Our results indicate that one source of cross-cultural influence on that relationship is aggregate country-level personality dispositions.

The Influence of Personality and Cultural Values

Our findings partly confirm previous studies that have described Extraversion and Neuroticism as dispositions that are systematically related to LS judgments (cf., Schimmack et al., 2004); we found that country-level Neuroticism has a negative effect and Extraversion has a direct positive effect on individual-level satisfaction with life (albeit the influence of Extraversion remained statistically insignificant). However, when the average levels of these personality traits were examined as moderators of the LS-PA and LS-NA relationships, personality had no significant impact in these models. Still, in the present study personality traits showed a slight moderating influence in the context of mixed emotions. More specifically, the more people within a country tend to describe themselves as neurotic and introverted, the stronger the influence of ME on LS judgments. This is partly consistent with an earlier study (Kööts et al., 2012) where momentary happiness and sadness were more likely to co-occur in people who scored higher in self-reported Neuroticism (especially in the facets of Depression and Impulsivity). The reason behind this tendency may be that when people within a nation typically perceive themselves as having few positive emotions (low Extraversion) and being emotionally instable (high Neuroticism), then experiencing mixed emotions harmonizes with the average levels of self-reported emotional traits in their country, and the effect of ME on LS is therefore enhanced.

As personality differences between cultures alone cannot provide a complete explanation for the different levels of LS (Realo & Dobewall, 2011), we supplemented these with prominent cultural values. Contrary to some prior studies (Kuppens et al., 2008; Suh et al., 1998), individualism had no statistically significant direct influence on LS; likewise it did not moderate the relationship between the cognitive and affective components of SWB. This can probably be explained by the fact that most European countries are relatively individualistic, compared with many other world regions, and therefore, the variance is too small to show any significant or reliable effect. Another unexpected finding was that in countries that value survival over self-expression (and also have a lower HDI ranking), LS was slightly more strongly related to positive affect and mixed emotions. This result seems to be directly the opposite of what Kuppens et al. (2008) had found in their study with college students from 46 countries. However, we believe that the latter (i.e., composition of samples) is actually the key to this discrepancy, because each country in the ESS has representative probability samples, covering the residents from age 15 up to the very old. Thus, when we included only younger participants in the analyses, this negative moderating effect indeed disappeared and the results were similar to Kuppens et al. (2008). Hence, even people in similar cultural contexts, but from different age groups, may not use exactly the same information in forming their life satisfaction judgments, and it is essential to examine as many representative samples as possible.

Limitations and Conclusions

First, it is disputable whether Europe (and the 21 countries we analyzed) is heterogeneous enough to make meaningful cultural comparisons. Europe is the second-smallest continent by surface area, yet it is the third most populous. However, an important advantage of our research is the ESS database, which ensures that each of the countries is comprehensively represented and includes a broad age spectrum of individuals (for a discussion about the relationship between age and LS, please see Realo & Dobewall, 2011). Second, subjective emotional experience and LS evaluations are both only accessible via self-report, which is why they are subject to potential personal perceptual bias. It would be useful to validate these data with non-self-report methods, such as observer reports, facial measures, physiological measures, and emotion-sensitive tasks (Diener & Ryan, 2009). Cross-cultural self-report measures are also vulnerable to various measurement biases, because the scales we use might have slightly variable meanings across different countries. Despite the encouraging results from the MCFA, the possibility nevertheless remains that the items did not comprise all relevant aspects in each of the countries. Third, the use of recall-based measures of affect and LS can also be problematic, as these could be prone to systematic distortions (Wirtz et al., 2009); retrospective judgments are argued to be constructions drawn on the spot on the basis of currently available information and circumstances and thus are sensitive to changes in the context of inquiry (Alexandrova, 2005). The fourth limitation of this study concerns the number of groups in multilevel analyses. The number of participating countries was in the current study quite small as in many other cross-cultural projects; ideally the sample should consist of 40 to 60 groups. Statistically, small group sample sizes may hamper estimations for the between-level parameters (Meuleman & Billiet, 2009). Therefore, caution is needed in interpreting the country-level moderators reported here. Our fifth concern is that we cannot make causal inferences from cross-sectional data; we cannot be completely certain that it is indeed emotions that influence life satisfaction and not vice versa. Additionally we have to take into account the fact that the country-level variables in our research (cultural values and aggregated personality traits) were not collected using the same (representative) samples as the ESS survey.

Regardless of all these limitations, our study makes a relevant contribution to the field of subjective well-being. In summary, the cognitive component of subjective well-being is influenced by positive and negative emotional experiences. We found that it is reasonable to analyze the effect of positive and negative emotions both separately and in interaction (as the degree of mixed emotions) on LS; the latter enabled us to take into account the relationships between all the components of SWB simultaneously. The most important result of our research is that the LS-PA and LS-ME relationships are enhanced in countries of lower HDI rank and higher survival values. The country-level aggregate personality traits only influenced the strength of the relationship between ME and LS. Our findings also emphasize the importance of employing representative samples in cross-cultural research as the age variance of participants can have a profound impact on results.

Footnotes

Acknowledgements

We acknowledge Ronald Fischer for his helpful suggestions on measurement invariance in cross-cultural research; we also thank Peter Halama for kindly providing the Slovakian NEO-FFI personality data and Delaney Michael Skerrett for his helpful comments on earlier drafts of this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a grant from the Estonian Ministry of Science and Education (SF0180029s08) and by the NORFACE research program on Migration in Europe: Social, Economic, Cultural, and Policy Dynamics (MMJRI10003).

Notes

References

Alexandrova

(2005). Subjective well-being and Kahneman’s ‘objective happiness.’ Journal of Happiness Studies, 6(3), 301-324.

Allik

McCrae

R. R.

(2004). Toward a geography of personality traits: Patterns of profiles across 36 cultures. Journal of Cross-Cultural Psychology, 35(1), 13-28.

Asendorpf

J. B.

(1992). Beyond stability: Predicting inter-individual differences in intra-individual change. European Journal of Personality, 6, 103-117.

Billari

F. C.

Goisis

Liefbroer

A. C.

Settersten

R. A.

Aassve

Hagestad

Speder

(2011). Social age deadlines for the childbearing of women and men. Human Reproduction, 26(3), 616-622.

Billari

F. C.

Liefbroer

A. C.

(2010). Towards a new pattern of transition to adulthood? Advances in Life Course Research, 15(2-3), 49-75.

Brim

O. G.

Ryff

C. D.

Kessler

R. C.

(Eds.). (2004). How healthy are we? A national study of wellbeing at midlife. Chicago: University of Chicago Press.

Börsch-Supan

Brugiavini

Jürges

Kapteyn

Mackenbach

Siegrist

Weber

(2008). Health, ageing and retirement in Europe (2004-2007). Starting the longitudinal dimension. Mannheim, Germany: Mannheim Research Institute for the Economics of Aging (MEA).

Byrne

B. M.

Shavelson

R. J.

Muthén

(1989). Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychological Bulletin, 105, 456-466.

Caprara

G. V.

Steca

(2006). The contribution of self-regulatory efficacy beliefs in managing affect and family relationships to positive thinking and hedonic balance. Journal of Social and Clinical Psychology, 25(6), 603-627.

10.

Cheng

S.- T.

(2006). Negative emotions make positive emotions more salient in well-being appraisal. Personality and Individual Differences, 40(3), 557-567.

11.

Cheung

M. W.- L.

Leung

(2006). Evaluating multilevel models in cross-cultural research: An illustration with social axioms. Journal of Cross-Cultural Psychology, 37(5), 522-541.

12.

Clark

A. E.

Senik

(2011). Is happiness different from flourishing? Cross-country evidence from the ESS. Revue D’Economie Politique, 121(1), 17-34.

13.

Cohn

M. A.

Fredrickson

B. L.

Brown

S. L.

Mikels

J. A.

Conway

A. M.

(2009). Happiness unpacked: Positive emotions increase life satisfaction by building resilience. Emotion, 9(3), 361-368.

14.

Costa

P. T.

McCrae

R. R.

(1992). NEO PI-R professional manual. Odessa, FL: Psychological Assessment Resources.

15.

Cuxart

Riba

(2009). Improving on the basis of experience: Implementation of the third wave of the ESS in Spain. Revista Espanola de Investigaciones Sociologicas, 125, 147-168.

16.

De Moor

M. H. M.

Costa

P. T.

Terracciano

Krueger

R. F.

De Geus

E. J. C.

Toshiko

. . . Boomsma

D. I.

(2012). Meta-analysis of genome-wide association studies for personality. Molecular Psychiatry, 17(3), 337-349.

17.

Diener

Biswas-Diener

(2002). Will money increase subjective well-being? A literature review and guide to needed research. Social Indicators Research, 57, 119-169.

18.

Diener

Emmons

R. A.

Larsen

R. J.

Griffin

(1985). The Satisfaction with Life Scale. Journal of Personality Assessment, 49, 71-75.

19.

Diener

Horowitz

Emmons

R. A.

(1985). Happiness of the very wealthy. Journal of Personality Assessment, 49, 71-75.

20.

Diener

Oishi

Lucas

R. E.

(2003). Personality, culture, and subjective well-being: Emotional and cognitive evaluations of life. Annual Review of Psychology, 54, 403-425.

21.

Diener

Ryan

(2009). Subjective well-being: A general overview. South African Journal of Psychology, 39(4), 391-406.

22.

Diener

Suh

E. M.

Lucas

R. E.

Smith

H. L.

(1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125(2), 276-302.

23.

Ervasti

Venetoklis

(2010). Unemployment and subjective well-being: An empirical test of deprivation theory, incentive paradigm and financial strain approach. Acta Sociologica, 53(2), 119-139.

24.

ESS Round 3: European Social Survey Round 3 Data. (n.d.). Data file edition 3.2. Norwegian Social Science Data Services, Norway – Data Archive and distributor of ESS data. Retrieved from http://www.europeansocialsurvey.org

25.

Fredrickson

B. L.

(2001). The role of positive emotions in positive psychology. American Psychologist, 56(3), 218-226.

26.

Gere

Schimmack

(2011). A multi-occasion multi-rater model of affective dispositions and affective well-being. Journal of Happiness Studies, 12, 931-945.

27.

Heck

R. H.

Thomas

S. L.

(2000). An introduction to multilevel modelling techniques. Mahwah, NJ: Lawrence Erlbaum Associates, Inc., Publishers.

28.

Hofstede

(2001). Culture’s consequences: Comparing values, behaviors, institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage.

29.

Hooghe

Vanhoutte

(2011). Subjective well-being and social capital in Belgian communities: The impact of community characteristics on subjective well-being indicators in Belgium. Social Indicators Research, 100(1), 17-36.

30.

Human Development Report 2007/2008 Data. (n.d.). Table 1: Human development index. Retrieved from http://hdr.undp.org/en/reports/global/hdr2007-8/

31.

Huppert

F. A.

Marks

Clark

Siegrist

Stutzer

Vittersø

(n. d.). Personal and Social Well-Being Module for the European Social Survey, Round 3 (NC8 Paper 3 b). Retrieved from http://www.cambridgewellbeing.org/Files/Well-being-Module_Jun06.pdf

32.

Huppert

F. A.

Marks

Clark

Siegrist

Stutzer

Vittersø

Wahrendorf

(2009). Measuring well-being across Europe: Description of the ESS Well-being Module and preliminary findings. Social Indicators Research, 91, 301-315.

33.

Inglehart

(1997). Modernization and postmodernization: Cultural, economic, and political change in 43 societies. Princeton, NJ: Princeton University Press.

34.

Inglehart

Baker

W. E.

(2000). Modernization, cultural change, and the persistence of traditional values. American Sociological Review, 65, 19-55.

35.

Inglehart

Welzel

(2005). Modernization, cultural change and democracy: The human development sequence. New York: Cambridge University Press.

36.

Jöreskog

K. G.

Sörbom

(1999). LISREL 8: Structural equation modelling with the SIMPLIS command language. Lincolnwood, IL: Scientific Software International.

37.

Kaminska

McCutcheon

A. L.

Billiet

(2010). Satisficing among reluctant respondents in a cross-national context. Public Opinion Quarterly, 74(5), 956-984.

38.

Kang

S. M.

Shaver

P. R.

Sue

Min

K. H.

Jing

H. B.

(2003). Culture-specific patterns in the prediction of life satisfaction: Roles of emotion, relationship quality, and self-esteem. Personality and Social Psychology Bulletin, 29(12), 1596-1608.

39.

Kasearu

Kutsar

(2011). Patterns behind unmarried cohabitation trends in Europe. European Societies, 13(2), 307-325.

40.

Kööts

Realo

Allik

(2012). Relationship between linguistic antonyms in momentary and retrospective ratings of happiness and sadness. Journal of Individual Differences, 33(1), 43-53.

41.

Kuppens

Realo

Diener

(2008). The role of positive and negative emotions in life satisfaction judgement across nations. Journal of Personality and Social Psychology, 95(1), 66-75.

42.

Larsen

R. J.

(2000). Maintaining hedonic balance: Reply to commentaries. Psychological Inquiry, 11(3), 218-225.

43.

Lefter

Cornelia

Vasilache

(2010). Statistical analysis of the evolution of values in human resource perspectives. Economic Computation and Economic Cybernetics Studies and Research, 44(3), 5-21.

44.

Lucas

R. E.

Clark

A. E.

Georgellis

Diener

(2004). Unemployment alters the set point for life satisfaction. Psychological Science, 15(1), 8-13.

45.

Lucas

R. E.

Diener

Grob

Suh

E. M.

Shao

(2000). Cross-cultural evidence for the fundamental features of Extraversion. Journal of Personality and Social Psychology, 79(3), 452-468.

46.

Lucas

R. E.

Diener

Suh

(1996). Discriminant validity of well-being measures. Journal of Personality and Social Psychology, 71(3), 616-628.

47.

Lykken

Tellegen

(1996). Happiness is a stochastic phenomenon. Psychological Science, 7, 186-189.

48.

MacCallum

R. C.

Browne

M. W.

Sugawara

H. M.

(1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130-149.

49.

Markon

K. E.

Krueger

R. F.

Watson

(2005). Delineating the structure of normal and abnormal personality: An integrative hierarchical approach. Journal of Personality and Social Psychology, 88(1), 139-157.

50.

McCrae

R. R.

(2002). NEO-PI-R data from 36 cultures: Further intercultural comparisons. In McCrae

R. R.

Allik

(Eds.), The five-factor model of personality across cultures (pp. 105-125). New York: Kluwer Academic/Plenum Publishers.

51.

McCrae

R. R.

Terracciano

, & 79 Members of the Personality Profiles of Cultures Project. (2005). Personality profiles of cultures: Aggregate personality traits. Journal of Personality and Social Psychology, 89(3), 407-425.

52.

Meredith

(1993). Measurement invariance, factor analysis, and factorial invariance. Pyschometrika, 58, 525-543.

53.

Meuleman

Billiet

(2009). A Monte Caro sample size study: How many countries are needed for accurate multilevel SEM? Survey Research Methods, 3(1), 45-58.

54.

Milfont

T. L.

Fischer

(2010). Testing measurement invariance across groups: Applications in cross-cultural research. International Journal of Psychological Research, 3(1), 111-121.

55.

North

R. J.

Pai

A. V.

Hixon

J. G.

Holahan

C. J.

(2011). Finding happiness in negative emotions: An experimental test of a novel expressive writing paradigm. Journal of Positive Psychology, 6(3), 192-203.

56.

Oishi

Diener

Lucas

R. E.

(2007). The optimum level of well-being: Can people be too happy? Perspectives of Psychological Science, 2(4), 346-360.

57.

Pethtel

Chen

(2010). Cross-cultural aging in cognitive and affective components of subjective well-being. Psychology and Aging, 25(3), 725-729.

58.

Pilcher

J. J.

(1998). Affective and daily event predictors of life satisfaction in college students. Social Indicators Research, 43, 291-306.

59.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401.

60.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage Publications.

61.

Realo

Dobewall

(2011). Does life satisfaction change with age? A comparison of Estonia, Finland, Latvia and Sweden. Journal of Research in Personality, 45(3), 297-308.

62.

Ruisel

Halama

(2007). NEO five factor personality inventory (according to NEO Five Factor Inventory of P. T. Costa and R. R. McCrae). Prague: Testcentrum–Hogrefe.

63.

Russell

J. A.

Carroll

J. M.

(1999). On the bipolarity of positive and negative affect. Psychological Bulletin, 125, 3-30.

64.

Schimmack

(2003). Affect measurement in experience sampling research. Journal of Happiness Studies, 4, 79-106.

65.

Schimmack

Oishi

Diener

(2002). Cultural influences on the relation between pleasant emotions and unpleasant emotions: Asian dialectic philosophies or individualism-collectivism? Cognition and Emotion, 16, 705-719.

66.

Schimmack

Oishi

Furr

Funder

D. C.

(2004). Personality and life satisfaction: A facet-level analysis. Personality and Social Psychology Bulletin, 30, 1062-1075.

67.

Schimmack

Radhakrishnan

Oishi

Dzokoto

Ahadi

(2002). Culture, personality, and subjective well-being: Integrating process models of life satisfaction. Journal of Personality and Social Psychology, 82(4), 582-593.

68.

Schutte

N. S.

Malouff

J. M.

(2011). Emotional intelligence mediates the relationship between mindfulness and subjective well-being. Personality and Individual Differences, 50(7), 1116-1119.

69.

Seidlitz

Wyer

R. S.

Diener

(1997). Cognitive correlates of subjective well-being: The processing of valenced life events by happy and unhappy persons. Journal of Research in Personality, 31(2), 240-256.

70.

Seligman

M. E. P.

Csikszentmihalyi

(2000). Positive psychology: An introduction. American Psychologist, 55, 5-14.

71.

Soons

J. P. M.

Kalmijn

(2009). Is marriage more than cohabitation? Well-being differences in 30 European countries. Journal of Marriage and Family, 71(5), 1141-1157.

72.

Spencer-Rodgers

Peng

Wang

(2010). Dialecticism and the co-occurrence of positive and negative emotions across cultures. Journal of Cross-Cultural Psychology, 41(1), 109-115.

73.

Steenkamp

J. B. E. M.

Baumgartner

(1998). Assessing measurement invariance in cross-national consumer research. Journal of Consumer Research, 25(1), 78-90.

74.

Steiger

J. H.

(1990). Structural model evaluation and modification: An interval estimation approach. Multivariate Behavioral Research, 25, 173-180.

75.

Suh

Diener

Oishi

Triandis

H. C.

(1998). The shifting basis of life satisfaction judgments across cultures: Emotions versus norms. Journal of Personality and Social Psychology, 74(2), 482-493.

76.

Sushko

(2008). The impact of Russia on governance structures in Ukraine. DIE Research Project “The Impact of Russia, India and China on Governance Structures in their Regional Environment (RICGOV).” Bonn, Germany: Deutsches Institut für Entwicklungspolitik. Retrieved from http://www.die-gdi.de/CMS-Homepage/openwebcms3.nsf/%28ynDK_contentByKey%29/ANES-7NQBKV/$FILE/DP%2024.2008.pdf

77.

Tay

Diener

(2011). Needs and subjective well-being around the world. Journal of Personality and Social Psychology, 101(2), 354-365.

78.

Vandenberg

R. J.

Lance

C. E.

(2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4-70.

79.

Van de Velde

Bracke

Levecque

Meuleman

(2010). Gender differences in depression in 25 European countries after eliminating measurement bias in the CES-D 8. Social Science Research, 39(3), 396-404.

80.

Van de Vijver

F. J. R.

Leung

(1997). Methods and data analysis for cross-cultural research. Thousand Oaks, CA: Sage.

81.

Van de Vijver

F. J. R.

Leung

(2000). Methodological issues in psychological research on culture. Journal of Cross-Cultural Psychology, 31, 33-51.

82.

Von dem Knesebeck

Pattyn

Bracke

(2011). Education and depressive symptoms in 22 European countries. International Journal of Public Health, 56(1), 107-110.

83.

Watson

Clark

L. A.

Tellegen

(1988). Development and validation of brief measures of positive and negative affect. The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070.

84.

Weiss

Bates

T. C.

Luciano

(2008). Happiness is a personal(ity) thing: The genetics of personality and well-being in a representative sample. Psychological Science, 19(3), 205-210.

85.

Wirtz

Chiu

C.- Y.

Diener

Oishi

(2009). What constitutes a good life? Cultural differences in the role of positive and negative affect in subjective well-being. Journal of Personality, 77(4), 1167-1196.

86.

Yamasaki

Sasaki

Uchida

Katsuma

(2011). Effects of positive and negative affect and emotional suppression on short-term life satisfaction. Psychology, Health & Medicine, 16(3), 313-322.