In recent years, the psychology of social class has become an emerging and promising area of research. Membership in a particular social class [or socioeconomic status (SES)], a multifaceted social category defined by material wealth, occupation, participation in educational and social institutions, as well as by subjective perceptions of one’s social class rank in society (Oakes and Rossi 2003; Piff et al. 2010), influences the processes and psychological states and social life of individuals. For example, social class has been linked to the type of school one attends (Domhoff 1998), academic performance (Bradley and Corwyn 2002), the social activities in which one participates (Bourdieu 1985), health status or mood (Adler et al. 1994), subjective well-being (Diener et al. 1993; Howell and Howell 2008), sense of personal control (Kraus et al. 2009; Lachman and Weaver 1998) and assumed identities (Stephens et al. 2007).

The social class to which we belong shapes the social contexts in which we live and grow, exposing us to shared material and social conditions that, in turn, require and promote specific behavioral repertoires (Kraus and Stephens 2012; Stephens et al. 2007). Over time, these patterns of behavior become norms and expectations that determine our identity and the manner in which we relate to others (Kraus et al. 2012). Therefore, in this research we examine the relationship between social class and trust in others. Specifically, we focus on three types of trust (generalized trust, interpersonal trust, and depersonalized ingroup trust).

1 Trust

There are different understandings and definitions of trust. Economists and sociologists conceptualize it as a phenomenon that operates within and between institutions and something that individuals place in relationships with said institutions, whereas social psychologists understand it as the expectations that are held regarding the counterpart with whom one interacts (Worchel 1979). Hence, trust is linked to the idea of vulnerability (Bigley and Pearce 1998). For example, Rousseau et al. (1998) define trust as a “psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another” (p. 395). If there is no vulnerability, that is, some risk of being deceived or of suffering negative consequences in relationships with others, then trust is not necessary. In this sense, Foddy et al. (2009) conceptualize trust as the expectation of receiving positive treatment from others in situations of uncertainty and risk.

Various types of trust are studied in psychology. Generalized trust is a general belief in human benevolence (Yamagishi and Yamagishi 1994) through which one can trust most people. This trust promotes approaching others to form social relationships and abandoning the comfort zone created by secure interpersonal relationships (Yamagishi et al. 1998), and it has been linked to a higher tendency to cooperate with strangers (Yamagishi 1986).

Interpersonal trust (also called “assurance”, “relationism” or “particularistic trust”) is based on the creation and maintenance of committed relationships with specific people, such as friends or family members, that is, the known people with whom an interpersonal connection is shared (Uleman et al. 2000; Yamagishi 1998; Yamagishi and Yamagishi 1994). With known people, their behavior is predictable, leading to the sense of security according to which one will not be betrayed (Hayashi et al. 1999).

Frequently, we must decide whether to trust people who we do not know personally but who are not total strangers when we share membership in a particular social group or category with them. Depersonalized ingroup trust alludes to the fact that this common categorical membership heuristically serves to infer that social interaction or exchange with these people will be beneficial (Yuki et al. 2005). Although the possible implications of depersonalized ingroup trust are barely known, this type of trust could be adaptive for members of a group in situations characterized by uncertainty or by lack of personal control (Fritsche et al. in press; Kenworthy and Jones 2009).

2 The Consequences of Trust

Trusting others has numerous positive consequences. For example, at the individual and interpersonal levels, people who are more willing to trust other individuals tend to be less unhappy or maladjusted (Rotter 1980), have better health status (Stafford et al. 2004), and live longer (Barefoot et al. 1998). Additionally, people with greater trust in others have a more positive view of democratic institutions, participate more in political activities and civic organizations, are more tolerant of minorities (Uslaner 2002), and are more inclined toward social mobilization (Glaeser 2016).

In the organizational field, trust is “a central component of organizational success” (Shaw 1997, p. 1) and a “vital component of effective working relationships” (Colquitt et al. 2007, p. 918). More specifically, trust facilitates cooperative behavior, reduces harmful conflict, decreases the costs of interactions, and promotes effective responses to crisis situations (Rousseau et al. 1998).

Additionally, on a social level, societies with higher levels of trust tend to be better governed and are safer, more egalitarian, and wealthier (Delhey and Newton 2005).

3 The Relationship Between Social Class and the Different Types of Trust

Social class and trust comprise two related phenomena. Previous studies have revealed a positive relationship between social class and generalized trust, such that people from higher social classes trust more in human benevolence (e.g. Alesina and La Ferrara 2002; Elgar 2010; Gheorghiu et al. 2009; Hamamura 2012; Pew Research 2007; Whiteley 1999).

To the best of our knowledge, there are no studies that have directly analyzed the relationship between social class and interpersonal trust and depersonalized ingroup trust. Nonetheless, there is indirect evidence that allows us to infer a negative relationship between social class and both types of trust. On one hand, Fiske et al. (2012) found a negative relationship between social class and relationism (a concept similar to interpersonal trust), such that the higher the social class is, the lower the scores in relationism. In this regard, it has also been found that people from low social classes are more relationally interdependent (Stephens et al. 2007) and adaptively respond to threats from their environment by building social support networks (Stellar et al. 2012). On the other hand, having to confront higher levels of uncertainty than people in higher social classes, the social contexts of lower classes expose people belonging to these classes to everyday situations that hinder the development of their personal capacity to influence, choose or control (Lachman and Weaver 1998). A psychological mechanism that can mitigate or reduce personal uncertainty is the assimilation of the self to a certain ingroup (Hogg 2007a).

In short, according to the discussion above, we expect that social class is related positively to generalized trust and negatively to interpersonal trust and depersonalized ingroup trust. Indeed, the first type of relationship has received previous empirical support, whereas the other two relationships would be novel. Because the differences between people in their levels of trust also vary depending on other variables (e.g. gender, age, and political ideology; Buchan et al. 2008; Rudolph and Evans 2005; Uslaner 2002), we have also analyzed whether the proposed relationships between social class and trust are independent of these variables.

4 Method

4.1 Participants

A total of 899 people (480 women and 409 men; 10 missing) with a mean age of 38.19 (SD = 14.13; range from 18 to 86) drawn from the general population and residing in a city in southern Spain participated in this study. Table 1 presents the sociodemographic information of the participants, which is in line with the characteristics of the Spanish general population (see INE 2016a; OECD 2009).

Table 1 Frequencies related to participants’ monthly income level and educational level
Full size table

4.2 Procedure

Two previously trained researchers requested the cooperation of the participants, informing them of the estimated duration (approximately 15 min), in addition to compliance with strict confidentiality and anonymity criteria in the submitted responses. The sample was obtained through a convenience sampling procedure in a city with approximately 236.000 inhabitants, sparsely diverse ethnically (the foreign population accounts for only around 6% of the total population), and located in southern Spain. Although it is one of the Spanish cities most negatively affected by the current context of economic crisis (the unemployment rate in 2012 was 35%, percentage slightly lower to the regional average (37%), but higher than the national average, which was 25%) (INE 2013), the participants of this study mirror, in general terms, the sociodemographic characteristics of the Spanish population, as indicated above with regard to the data included in Table 1.

Basically, the participants agreed to voluntarily complete a questionnaire following one of the next three procedures: (1) Participants were asked to take part in this study while they were waiting at the bus station, as well as in other public spaces. These participants did not receive any reward for their participation; (2) College students participated in this study in exchange for receiving credit courses; and (3) College students requested their parents to fill in the above-mentioned questionnaire (this group was different from the previous one in the sense that when their parents participated in the study, the students themselves did not do so). The college students who achieved their parents’ participation were also rewarded with credit courses. Finally, after having completed the questionnaire, all the participants were fully debriefed and thanked.

4.3 Instruments

Participants completed a questionnaire that included the following measures:

  • Sociodemographic Data Information on age, gender, level of family income and education was collected. The answer format for these last two variables is shown in Table 1.

  • Political Ideology Participants rated their political ideology on a scale ranging from 0 (left wing) to 10 (right wing). The mean for this measure was 4.19 (SD = 2.40).

  • Objective Social Class In line with previous studies (e.g. Piff et al. 2010), objective social class was operationalized based on income level and completed formal education. Monthly family income was coded into six categories, listed in Table 1. A number between 1 and 6 was assigned to each category, with higher numbers indicating greater income (M = 2.63; SD = 1.17). Educational level, according to the responses submitted to the categories shown in Table 1, was scored from 1 to 6 (higher numbers indicated greater educational level; M = 4.07; SD = 1.71). Scores on income and educational level were standardized and summed to obtain a general standardized measure of objective social class or socioeconomic status (SES).

  • Subjective Social Class Participants completed the MacArthur Scale of subjective SES (e.g. Ostrove et al. 2000). It consists of a drawing of a 10-rung ladder representing people with different levels of income, education, and occupation status in society. A number between 1 and 10 was assigned to each rung of the ladder, with higher numbers indicating higher placement on the social ladder. Mean and standard deviation for this measure are included in Table 2.

    Table 2 Descriptive statistics and correlations between measures of (objective and subjective) social class, generalized trust, interpersonal trust, and depersonalized ingroup trust
    Full size table

  • Generalized Trust Adopting the scale by Yamagishi and Yamagishi (1994) on generalized trust, a version in Spanish developed by us was used. It consists of six items (e.g. “Most people are basically honest”), with a Likert-type answer format of five options ranging from 1 (totally disagree) to 5 (totally agree). The Cronbach’s alpha coefficient in this study is .74.

  • Interpersonal Trust To measure this construct, a measure recently developed that attempts to capture trust in close relationships (family members and friends) was used (Moya et al. 2011). It consists of five items (e.g. “I only trust people who I know personally”), with a Likert-type answer format with five options ranging from 1 (totally disagree) to 5 (totally agree). The psychometric data reflect a Cronbach’s alpha coefficient of .71.

  • Depersonalized Ingroup Trust A version in Spanish of a measure composed of three items that directly evaluate depersonalized ingroup trust was used (Kenworthy and Jones 2009), with social class being applied as a reference category (e.g. “I trust all members of my social class background”), in addition to a further item developed by us (“For me, everyone who belongs to my social class is trustworthy”). The answer format is a Likert-type scale with five options ranging from 1 (totally disagree) to 5 (totally agree). The Cronbach’s alpha coefficient obtained in this study is .81.

5 Data analytic strategies

To analyze whether social class was related to generalized trust, interpersonal trust, and depersonalized ingroup trust, an analysis of the Pearson bivariate correlations was conducted (see Table 2). Additionally, a series of analyses on hierarchical regressions was conducted to examine the effect of social class, regardless of sociodemographic variables such as gender, age, and political ideology, in the three types of trust examined (see Table 3). The statistical package employed to perform these analyses was SPSS 19.

Table 3 Summary of hierarchical regression analysis for variables predicting generalized trust, interpersonal trust, and depersonalized ingroup trust
Full size table

6 The Relationship Between Social Class and the Different Types of Trust

As seen in Table 2, objective and subjective social class were positively and significantly related to each other. However, only objective social class was positively and significantly associated with generalized trust, and negatively and significantly associated with interpersonal trust and depersonalized ingroup trust. Finally, the three types of trust evaluated were significantly related to each other.

7 The Effect of Social Class and Other Sociodemographic Variables on the Different Types of Trust

To verify whether membership in a particular social class predicted the different types of trust and whether this effect was independent of the participants’ gender, age, and political orientation, three hierarchical regression analyses were conducted (one for each type of trust as criterion variables). In the first step, the participant’s gender (0 = female; 1 = male), age, and political ideology were introduced. In the second step, objective social class and subjective social class were introduced.

7.1 Generalized Trust

Table 3 shows that, of the sociodemographic and ideological variables, only age was positively related to generalized trust (β = .175, p < .001). When objective and subjective social class were introduced, only objective social class significantly predicted generalized trust (β = .149, p < .001), and the model significantly increased (2.1%) the explained variance, F(2, 849) = 9.19, p < .001. Therefore, the higher the participant’s (objective) social class is, the greater his/her score on generalized trust.

7.2 Interpersonal Trust

As shown in Table 3, none of the variables introduced in the first step significantly predicted this type of trust. In the second step, after adding objective and subjective social class, the model explained an additional 0.8% of the variance in interpersonal trust, F(2, 851) = 3.42, p = .033. As in the previous case, only objective social class significantly predicted interpersonal trust (β = −.077, p = .049). Thus, the lower the participant’s (objective) social class is, the greater his/her score on interpersonal trust.

7.3 Depersonalized Ingroup Trust

Finally, none of the sociodemographic variables introduced in the first step significantly predicted depersonalized ingroup trust (see Table 3). When objective and subjective social class were introduced in the second step, the explained variance increased by 1.2%, F(2, 849) = 5.18, p = .006. As in the other two types of trust, only objective social class significantly predicted depersonalized ingroup trust (β = −.111, p = .005). Therefore, the lower the participant’s (objective) social class is, the greater his/her score on depersonalized ingroup trust.

8 Discussion and Conclusions

This study analyzes the existing relationship between social class and three different types of trust: generalized, interpersonal, and depersonalized ingroup (social class). The obtained results, in line with the hypothesis, show that (objective) social class is related significantly and positively to generalized trust and negatively to interpersonal trust and depersonalized ingroup trust. The results also show that (objective) social class is related to trust, regardless of the participants’ gender, age, and political ideology.

Lower levels of generalized trust found in lower social classes confirm the results obtained in previous studies (e.g. Alesina and La Ferrara 2002; Gheorghiu et al. 2009; Hamamura 2012; Pew Research 2007; Whiteley 1999). Possessing limited material resources, people from lower social classes, compared to their counterparts in higher social classes, move in a social environment of greater vulnerability and, therefore, an environment that is also more uncontrollable. In this context, the costs that stem from an unfavorable social interaction with a stranger or an unknown person could be very harmful. Additionally, when people of lower social classes face adversity, they do not have the same network of material and financial security that, by contrast, people from higher social classes have (Lamont 2000). Under such conditions, generalized trust could pose an additional risk because people from more disadvantaged social classes could most likely be affected more negatively by a potential betrayal or unfavorable treatment in a social exchange. Thus, the lower levels of generalized trust found in more disadvantaged social classes could serve as a protective mechanism (Fiske et al. 2012; Simmel 1950).

Moreover, higher levels of interpersonal trust among lower social classes were found. This is greater trust in the known and specific people with whom a history of interpersonal relationships is shared (fundamentally, friends and family members). Focusing on trust in interpersonal committed relationships, people from more disadvantaged social classes most likely seek to ensure the protection of their limited material resources –essential for satisfying their basic everyday needs—and, therefore, reducing the risk and uncertainty that define their social context. Consistent with this idea, it should be noted that interpersonal trust involves the ability to predict and, thus, a greater sense of security that derives from the knowledge that one possesses about these people (Fiske et al. 2012).

In parallel, at an intragroup level, greater levels of depersonalized ingroup (social class) trust were also found in people from lower social classes compared to those from high social classes. Although there are no previous studies that link this type of trust with the fact of belonging to a certain social class, depersonalized ingroup trust could be increased, ultimately, in response to the aversive effects that come from personal states of uncertainty (see Hogg 2007a). Thus, it may be that the personal feelings of uncertainty caused by a lack of financial resources, according to uncertainty-identity theory (Hogg 2007b, 2012, 2015), are overcome through self-categorization as a member of a certain ingroup (social class) and the subsequent unfolding of different processes of depersonalization, such as ingroup trust (Hogg, 2007a).

In short, people from more disadvantaged social classes could have adaptively learned over time to confront the unfavorable material and social conditions that characterize their social context, generally distrusting unknown people and trusting specific known people and those with whom they share a certain social category (social class). These patterns of trust, although having positive consequences (because trust facilitates interpersonal relationships), can have less desirable effects. Generally distrusting unknown people and preferentially focusing trust on those with whom close interpersonal relationships are maintained or who belong to the same social class involve not leaving a comfort zone. Although providing the ability to predict and, therefore, relief, certainty, and security, not leaving a comfort zone impedes us from knowing people who do not belong to the inner circle, which hinders the development of personal initiatives, risk taking and, thus, access to new social and economic opportunities (Yamagishi et al. 1998). Generalized trust facilitates this process of growth (Hayashi and Yamagishi 1998). In this regard, as suggested by Van Lange (2015), a healthy dose of generalized trust promotes obtaining optimal results in one’s social life.

Regarding the study’s limitations, it is worth mentioning the fact that certain variables that are likely to affect trust, such as religion (Tan and Vogel 2008), race (Smith 2010) or exposure to ethnic diversity (Dinesen and Sønderskov 2015), have not been evaluated. However, it should also be pointed out here that Spain, despite some social changes occurred during the last decades, continues to be a very homogeneous country on these matters.Footnote 1

In conclusion, this study is framed within the complex interplay that occurs between constructs of the macrosocial order (social class) and individual psychological processes (trust). In our study, only the objective measure of social class was significantly related to the three types of trust analyzed. Although it has sometimes been found that subjective social class rank is a more potent predictor of psychological outcomes in comparison with different objective components of social class (see Kraus et al. 2009), the subjective conceptualization of social class also has certain limitations. For instance, individuals may place themselves in the social hierarchy according to their social aspirations or tend to identify themselves as middle class though they are not (because they are in a lower or higher social class) (Hout 2008), which could statistically constrain the discriminatory power of the subjective assessments of social class membership. This might have happened in our study, given the participants’ mean scores in the 10-point subjective social class scale (M = 6.05; SD = 1.45). Besides that, the results exposed above are to some extent consistent with the results obtained by Elgar (2010), who found that the educational attainment –one of the most important objective indicators of social class- predicted generalized trust better than subjective social class across thirty-three countries. To sum up, the present study, which is exploratory in nature, requires further research that replicates its results in other cultural contexts and delineates the possible mediating, moderating, and explanatory mechanisms of the relationship found between social class and the different types of trust examined (in addition to research that explores the behavioral consequences of these types of trust).