- Email: [email protected]
The relationship between subjective social class and aggression: A serial mediation model
Personality and Individual Differences 131 (2018) 174–179
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
The relationship between subjective social class and aggression: A serial mediation model Bing Chena, Youxia Zuoa, Yufang Zhaoa,b, a b
T
⁎
Faculty of Psychology, Southwest University, 2 Tiansheng Road, Beibei, ChongQing 400715, China Center for Studies of Education and Psychology of Minorities in Southwest China, Southwest University, Chongqing, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Subjective social class Perceived social support Sense of control Negative affect Aggression Serial mediation
Although the relationship between the phenomena of social class and aggression is well known, very few studies have been conducted to investigate its underlying mechanisms. This study examines the relationship between subjective social class (SSC) and aggression, proposing a serial mediation model for the relationship. A total of 305 valid participants, ranging in age from 17 to 26, completed questionnaires assessing SSC, perceived social support, sense of control, negative affect, and aggression. Serial mediation analysis revealed a pathway whereby lower SSC was associated with less perceived social support, which was associated with decreased sense of control and increased negative affect, which were then associated with more aggression. Our results shed light on the associations between the above-mentioned variables in the SSC–aggression relationship. We also suggest possible prevention and intervention programs for reducing aggression among individuals with low SSC and suggest courses for future inquiry.
1. Introduction Aggression is a major contemporary social problem and may be seen as a destructive means of modern conflict. It is also a complex phenomenon, typically said to be made up by biological, environmental, psychological, and social factors (Anderson & Bushman, 2002). Social class is considered a powerful social factor, and low social class has been associated with increased levels of aggression, for example, lowincome (McFarlin, Fals-Stewart, Major, & Justice, 2001) and low-education groups (Barefoot et al., 1991) were found to be relatively more aggressive than high-income and high-education groups. People's subjective perceptions of their social class can also predict aggression. One study found that participants perceiving their social class as relatively lower behaved more aggressively relative to those who perceived theirs as relatively higher (Greitemeyer & Sagioglou, 2016). Despite the existing evidence, further empirical work is needed to explore the underlying mechanisms of the relationship between social class and aggression. The current study thus investigated whether perceived social support, sense of control, and negative affect mediate this relationship. Social class, which in academia is often used interchangeably with socioeconomic status (Côté, 2011), is a multifaceted construct comprising an individual's material resources as well as his or her perceived rank within the social hierarchy (Kraus, Piff, & Keltner, 2009). Researchers often assess the construct as objective social class (OSC) or
⁎
subjective social class (SSC); typical measures of the former include income, education, and profession (Goodman et al., 2001), while the latter emphasizes individuals' perceived rank relative to others in society (Kraus et al., 2009). SSC can be seen as a comparative perception of, for example, material and social resources, i.e., individuals' sense of what they have or possess relative to others. Several studies have demonstrated that SSC is positively and moderately correlated with OSC (Adler, Epel, Castellazzo, & Ickovics, 2000; Johnson & Krueger, 2006), and that it is a relatively better predictor of psychological outcome than OSC (Adler et al., 2000; Kraus et al., 2009). In the present study, we thus assess social class subjectively and mainly focus on mechanisms pertaining to the relation between SSC and aggression. According to the cognitive neoassociation model (CNA model; Berkowitz, 1990), aversive events that lead to negative affect can lead to aggression. Given their limited resources and lower social rank, lower class individuals are more likely to experience threats (Nelson, 2009), to be characterized by chronic levels of cynical mistrust and hostility (Gallo & Matthews, 2003), and to feel more socially rejected (Johnson, Richeson, & Finkel, 2011) than higher class individuals. Therefore, they are at risk of suffering chronically from exposure to aversive events, which, in turn, may foster negative affect and hence increase the risk of behaving aggressively. The CNA model thus plausibly explains class differences in aggression. However, the question remains how aversive events influence or bring about the subsequent
Corresponding author at: Faculty of Psychology, Southwest University, 2 Tiansheng Road, Beibei, ChongQing 400715, China. E-mail address: [email protected] (Y. Zhao).
https://doi.org/10.1016/j.paid.2018.04.036 Received 15 September 2017; Received in revised form 18 March 2018; Accepted 23 April 2018 0191-8869/ © 2018 Published by Elsevier Ltd.
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
have the most money, the most education, and the most respected jobs. At the bottom the people who are the worst off – who have the least money, the least education, and the least respected jobs or no jobs”. Then they were instructed to select a number representing their perception of their family's placement on this 10-point social scale, with higher numbers indicating higher perceived social class. This measure is widely used and has demonstrated adequate test-retest reliability (Operario, Adler, & Williams, 2004). We used family information instead of participants' own information as college students typically are not financially independent and their status thus based on their upbringing (Henry, 2009).
experiences of negative affect. In the present study, we investigate the potential roles of perceived social support and sense of control in this process. Social support, thought to mitigate the effects of stressful life events on mental health (Cohen & Wills, 1985), may to some extent depend on social class. Previous studies have demonstrated that male students from lower income countries report a lack of social support (Peltzer, Pengpid, & James, 2016), and that social class positively predicts social support (Wangberg et al., 2008). Hence, relatively lower class individuals are more likely to perceive lower levels of social support. This absence of social support, however, is often furthermore associated with a decreased sense of control (Ell, Mantell, Hamovitch, & Nishimoto, 1989; Ruthig, Haynes, Stupnisky, & Perry, 2009). Sense of control is a concept that refers to peoples' beliefs about the extent to which they are able to shape their own social outcomes (Lachman & Weaver, 1998). It is further seen as a fundamental social need of humans (Williams, 2007), and a perceived lack of control is associated with increased experiences of negative affect. Studies have shown that people, when they perceive a lack of control, feel upset or unhappy (Baumeister, 2005), and that lower levels of control correlate with higher levels of anxiety (Ong, Bergeman, & Bisconti, 2005). Previous studies have thus demonstrated that differences in aggression appear to be contingent on experiences of negative or aggression-related affect (Greitemeyer & Sagioglou, 2016; Smith, Pettigrew, Pippin, & Bialosiewicz, 2012), and the CNA model also suggests that negative affect rather than aversive events should serve as the proximate mechanism for eliciting aggression (Berkowitz, 1990). Hence, we hypothesize that living with low SSC may decrease levels of perceived social support and subsequently sense of control, resulting in more experiences of negative affect; this, in turn, may lead to the final aggression. Guided by the CNA model and previous results, the present study aimed to elucidate potential mechanisms underlying the relationship between SSC and aggression, proposing a serial mediation model for this relationship (see Fig. 1). We expand previous research by providing a comprehensive test of the serial process of how SSC influences aggression through negative affect. We also sought to provide several prevention and intervention methods that may help reduce aggression and maintain social stability.
2.2.2. Perceived social support The Chinese version of Perceived Social Support Scale (Jiang, 1999; Zimet, Dahlem, Zimet, & Farley, 1988) was used to measure participants' perceived social support. It consists 12 items designed to assess three sources of support: significant other (leaders, relatives, and colleagues), family, and friends. The examples of significant other were changed into “teachers, relatives, and classmates” which were more suitable for our sample (Yan & Zheng, 2006). Each subscale consists of four items that are answered on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). In this study, Cronbach's α for the subscales and the whole scale was 0.81, 0.82, 0.88, and 0.91. 2.2.3. Sense of control The Chinese version of Sense of Control Scale (Lachman & Weaver, 1998; Li, 2012) was used to assess participants' sense of control. This scale includes the two subscales: personal mastery and perceived constraints. The items are rated on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). The perceived constraints scale is reverse scored. In this study, Cronbach's α for the subscales and the whole scale was 0.70, 0.79, and 0.80. 2.2.4. Negative affect The Negative Affect Scale was selected from a subscale of the Chinese version Positive and Negative Affect Scale (Qiu, Zheng, & Wang, 2008; Watson, Clark, & Tellegen, 1988) to assess participants' negative affect. This scale consists of nine adjectives describing negative emotions (e.g., anger, nervous), which are rated on a 5-point Likert scale (1 = very slightly or not at all, 5 = extremely) based on how one has felt during the past week. In this study, Cronbach's α for the scale was 0.85.
2. Methods 2.1. Participants and procedure The participants were 305 college students (248 females and 57 males) from a university in southwest China, aged 17–26 (M = 20.59, SD = 1.99). They were recruited via QQ (a famous chat tool in China), and participated voluntarily. Participants first gave written informed consent, after which they completed an anonymous questionnaire in about 10 min in a laboratory room. Subsequent to completion, students were compensated with a small gift (approximately 1.50 USD).
2.2.5. Aggression The Chinese version of Aggression Questionnaire (Bao, 2009; Buss & Perry, 1992) was used to assess participants' aggression. It consists of the four subscales: physical aggression, verbal aggression, anger, and hostility. The items are answered on a 5-point Likert scale (1 = definitely doesn't apply to me, 5 = definitely applies to me). In this study, Cronbach's α for the subscales and the whole scale was 0.77, 0.49, 0.79, 0.70, and 0.87.
2.2. Measures
2.3. Analysis strategy
2.2.1. Subjective social class (SSC) The MacArthur Scale of subjective socioeconomic status (Adler et al., 2000) was used to measure SSC. The measure consists of a drawing of a ladder with 10 rungs representing people with different levels of income, education, and occupational status. Each rung of the ladder was given a number between 1 and 10. Participants were told: “At the top of the ladder are the people who are the best off – those who
According to Anderson and Gerbing's (1988) suggestion, we adopted the two-step strategy for analysis of mediation effects. First, in order to assess the extent to which each of the latent variables was represented by its indicators, the measurement model was confirmed using confirmatory factor analysis (CFA). Second, structural equation modeling (SEM) analysis was performed to measure the fit and path coefficients of the hypothesized structural model. Furthermore, biasFig. 1. Hypothesized model Note: SSC = subjective social class, PSS = perceived social support, SOC = sense of control, NA = negative affect.
175
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
Table 1 Reliability and validity.
1.SSC 2.NA 3.PSS 4.SOC 5.Aggression
CR
AVE
MSV
1
2
3
4
5
– 0.81 0.63 0.60 0.76
– 0.60 0.46 0.45 0.45
– 0.28 0.17 0.36 0.36
– −0.02 0.16 0.23⁎⁎⁎ −0.15⁎
0.77 −0.18 −0.45⁎⁎⁎ 0.53⁎⁎⁎
0.68 0.42⁎⁎⁎ −0.29⁎
0.67 −0.60⁎⁎⁎
0.67
(1) ⁎p < 0.05; ⁎⁎p < 0.01; ⁎⁎⁎p < 0.001 (two-tailed); N = 305. (2) Diagonal elements in boldface are the square root of AVE; Off-diagonal elements are the correlation coefficients.
SRMR = 0.05, and all the indicators loaded significantly on their corresponding factors. Second, reliability, convergent validity, and discriminant validity were tested by the measures of composite reliability (CR), average variance extracted (AVE), maximum shared squared variance (MSV), and inter-construct correlation. Good reliability is indicated by a CR above 0.70 (Hair, Black, Babin, & Anderson, 2010), with values above 0.60 deemed acceptable (Tseng, Dörnyei, & Schmitt, 2006). AVE should be at above 0.50 in order to establish convergent validity, and values above 0.40 are acceptable (Fornell & Larcker, 1981; Huang, Wang, Wu, & Wang, 2013). MSV should be less than AVE, and the inter-construct correlations should be less than the square root of AVE in order to establish the discriminant validity (Hair et al., 2010). Regarding this study, the results in Table 1 shows that all the indices have reached acceptable levels. Third, we employed the Harman one-factor test in CFA (Korsgaard & Roberson, 1995) to assess the common method biases (CMB), if CMB is a serious problem, the single factor model should fit the data as well as a more complex model. The results revealed a deteriorated model fit with χ2/df = 9.56, GFI = 0.75, AGFI = 0.65, CFI = 0.51, TLI = 0.41, RMSEA = 0.17, and SRMR = 0.16. This indicated that CMB was not a major concern in this study.
corrected bootstrap confidence intervals based on 5000 bootstraps (Hayes, 2009) were calculated to test the significance of the mediating effects. The following indices were utilized to evaluate the goodness of fit of the model (see Hooper, Coughlan, & Mullen, 2008; Hu & Bentler, 1999): (a) value of χ2/df of 5 or less; (b) Goodness of fit index (GFI) of 0.90 or more; (c) Adjusted GFI (AGFI) of 0.90 or more; (d) Comparative fit index (CFI) of 0.90 or more; (e) Tucker–Lewis index (TLI) of 0.90 or more; (f) Root-mean-square error of approximation (RMSEA) of 0.08 or less; and (g) Standardized root-mean-square residual (SRMR) of 0.08 or less. 3. Results 3.1. Preliminary analyses To reduce the number of parameters in the SEM analysis and to increase the accuracy of parameter estimates, we parceled the items on each measure (except for SSC, which was a single value). We used the internal-consistency approach (Kishton & Widaman, 1994) to parcel the items on perceived social support, sense of control, and aggression that all had multidimensional item sets. This approach creates different parcels that use the facets as grouping criteria. That is, each subscale or dimension of these scales were parceled into one manifest indicator. We used the factorial algorithm approach (Rogers & Schmitt, 2004) to parcel the items on negative affect that did not have multidimensional item sets. This approach rank orders the manifest indicators using their loadings on the first principal axis factor, and each parcel was sequentially assigned the remaining indicators with the highest and lowest rankings, alternating direction through the parcels, until all indicators were assigned. In order to ensure the unidimensionality of our parceled targets, we conducted several independent CFAs for negative affect, and each of the subscales of perceived social support, sense of control, and aggression before the parceling as recommended (Wu & Wen, 2011). Six items were eliminated as their factor loadings were below 0.40 (Fliege et al., 2005). The output of the CFA also suggested that modification indexes of ten pairs of items were extensively high, indicating a certain degree of overlap. This prompted the elimination of one item in each pair from the subsequent analyses (Ugulu, 2013).
3.3. Structural model testing The structural-modeling results indicated that the hypothesized model fit the data well, χ2/df = 2.47, GFI = 0.94, AGFI = 0.90, CFI = 0.93, TLI = 0.90, RMSEA = 0.07, and SRMR = 0.05. Fig. 2 shows the properties of the research hypothesis, including all the standardized path coefficients and factor loadings. 3.4. Test of mediating effects The bootstrap procedure was used to examine the significance levels of indirect effects for the hypothesized model. Results, shown in Table 2, indicate that the total effect of SSC on aggression was significant, β = −0. 1513, SE = 0.0637, 95% CI [−0.2722, −0.0255]. When the mediators were included in the analysis, this coefficient was no longer statistically significant (direct effect), β = −0.0414, SE = 0.0633, 95% CI [−0.1578, 0.0879]. In line with our hypothesis, the indirect effect of the SSC on aggression via perceived social support, sense of control, and negative affect was significant, β = −0.0098, SE = 0.0085, 95% CI [−0.0469, −0.0008]. Specifically, SSC significantly and positively predicted perceived social support, perceived social support significantly and positively predicted sense of control, sense of control significantly and negatively predicted negative affect, and negative affect significantly and positively predicted aggression (for standardized path coefficients, see Fig. 2). Moreover, the indirect effect of SSC on aggression through perceived social support and sense of control was significant, β = −0.0264, SE = 0.0260, 95% CI [−0.1332, −0.0004]. The indirect effect of SSC on aggression via sense of control alone was also significant, β = −0.0706, SE = 0.0435, 95% CI [−0.1843, −0.0137]. And the indirect effect of SSC on aggression via sense of control and
3.2. Measurement model testing First, we tested the fit of the measurement model, χ2/df = 2.19, GFI = 0.94, AGFI = 0.90, CFI = 0.95, TLI = 0.93, RMSEA = 0.06, and SRMR = 0.05. We found a nonsignificant variance in error in one factor in the estimate (the parceled item “significant other” of perceived social support scale). According to Bagozzi and Yi's (1988) suggestion, nonsignificant error variances usually suggest specification errors, and one solution when it occurs is to drop a measurement corresponding to the nonsignificant error. We thus deleted this item and reran the analysis. The estimates of the overall goodness-of-fit criteria exceeded their acceptable levels, which confirmed that the measurement model exhibited a fairly good fit with the data, χ2/df = 2.47, GFI = 0.94, AGFI = 0.90, CFI = 0.93, TLI = 0.90, RMSEA = 0.07, and 176
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
Fig. 2. The structural model. Note: Family = family support, Friend = friend support, PM = personal mastery, PC = perceived constraints, NA1-NA3 = 3 parcels of negative affect, A = anger, PA = physical aggression, VA = verbal aggression, H = hostility.
negative affect was significant as well, β = −0.0263, SE = 0.0163, 95% CI [−0.0685, −0.0055]. However, neither the indirect effect of SSC on aggression via perceived social support or negative affect alone, nor a chain of these two variables was significant, as witnessed by 95% CIs that included zero (see Table 2).
which is facilitative of aggression. However, the aversive events that lower class individuals experience may furthermore facilitate lower perceptions of their social class rank, making up what may be considered a vicious circle. Therefore, policies for attenuating the actual class inequality may help to reduce instances of interpersonal hostility. Importantly, our results highlight the significance of sense of control in the relationship between SSC and aggression. We found that the three-mediator pathway (via perceived social support, sense of control, and negative affect), the two-mediator pathways (via perceived social support and sense of control, and via sense of control and negative affect), and the single-mediator pathway (via sense of control) were all statistically significant. The indirect effects via perceived social support and negative affect, and via either alone were not. This suggests that perceived social support, sense of control, and negative affect all act as mediators, influencing the internal mechanism of the SSC-aggression relationship, with sense of control being the more important factor. In addition, our results illustrate that SSC may influence aggression directly through sense of control. When people report that they are of low SSC, they are indicating that they have fewer resources and are of subordinate rank relative to others. These self-perceptions of reduced resources and subordinate rank are often associated with a diminished sense of control (Kraus et al., 2009). As a result of this reduced control, people may behave aggressively, possibly as a means to regain or increase their sense of control (Mueller, 1983). Therefore, increasing the perceived control of lower class individuals, such as recalling a
4. Discussion The main purpose of this study was to examine the possible roles of perceived social support, sense of control, and negative affect as mediators of the relationship between SSC and aggression. Our results suggest that a lack of perceived social support, resulting from lower SSC, decreases sense of control and, in turn, is associated with more negative affect and finally greater aggression. This indirect path was statistically significant, rendering the direct path from SSC to aggression statistically insignificant. This indicates that the relationship between the SSC and aggression is fully, serially mediated by perceived social support, sense of control, and negative affect. The negative relationship between SSC and aggression is consistent with previous studies (Greitemeyer & Sagioglou, 2016; McFarlin et al., 2001). Compared to people with higher SSC, people with lower SSC report more aggression; as outlined above, this finding can be explained by the CNA model (Berkowitz, 1990). Lower class individuals often suffer chronically from aversive-event exposure (Johnson et al., 2011; Nelson, 2009), and they thus tend to experience more negative affect, Table 2 Indirect effects, direct effect, and total effect. Point estimate
Ind1(SSC–PSS–aggression) Ind2(SSC–PSS–SOC–aggression) Ind3(SSC–PSS–NA–aggression) Ind4(SSC–PSS–SOC–NA–aggression) Ind5(SSC–SOC–aggression) Ind6(SSC–SOC–NA–aggression) Ind7(SSC–NA–aggression) Total indirect effect Direct effect Total effect
−0.0072 −0.0264 0.0005 −0.0098 −0.0706 −0.0263 0.0298 −0.1100 −0.0414 −0.1513
SE
BC bootstrap 95% CI
0.0231 0.0260 0.0073 0.0085 0.0435 0.0163 0.0240 0.0549 0.0633 0.0637
Lower
Upper
(two-tailed)
−0.0742 −0.1332 −0.0145 −0.0469 −0.1843 −0.0685 −0.0070 −0.2299 −0.1578 −0.2722
0.0213 −0.0004 0.0147 −0.0008 −0.0137 −0.0055 0.0882 −0.0135 0.0879 −0.0255
0.3990 0.0444 0.9333 0.0300 0.0144 0.0144 0.1099 0.0259 0.5336 0.0211
Note: All the values of point estimates are standardized; BC bootstrap = bias-corrected bootstrap; 5000 bootstrap samples. 177
p
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
Acknowledgement
situation in which they had a great deal of control (Kraus et al., 2009), is likely to reduce their aggression. Future studies should investigate potential beneficial interventions, including at the preventative level. Our results also expand the CNA model by demonstrating how SSC influences aggression through negative affect. That is, people with lower SSC tend to perceive less social support and less sense of control. As previous studies have found that a lack of control may be considered an aversive state or experience (Whitson & Galinsky, 2008), these people may then subsequently experience varying degrees of negative affect (Baumeister, 2005; Ong et al., 2005), which, in turn, is associated with the final aggression (Greitemeyer & Sagioglou, 2016; Smith et al., 2012). The CNA model (Berkowitz, 1990) also implies that aversive experiences can lead to aggressive responding through a multistage process. In the first stage, the aversive event will lead to an undifferentiated negative affective state. This reaction will stimulate fight or flight tendencies, with the fight tendency being related to both aggression and rudimentary feelings of anger. Flight, on the other hand, is escape-related and will give rise to rudimentary feelings of fear. In the second stage, higher-order cognitive processing comes into play. As people think about what happened and consider the possible consequences, the rudimentary emotions of anger or fear are differentiated into more elaborate emotional states, like annoyance, irritation, anger, etc. This implies that the aggression is not inevitable and may be either promoted or suppressed by the emotional experience elicited by the aversive event (Krahé, 2001). Similarly to this viewpoint, Sun, Yu, Luo, and Yang (2011) found that children characterized by low levels of aggression scored higher on the measures of emotion regulation than those characterized by high levels of aggression. Thus, enhancing emotion-management skills of individuals who perceive their social class as low may decrease their potential aggressive tendencies to some extent. Of note, the positive association between perceived social support and sense of control in the present study is consistent with previous studies (Ell et al., 1989; Ruthig et al., 2009). Social support may be said to act as a psychosocial resource that can buffer individuals against stress and depression and help people control what happens to them in their daily life (Cohen & Wills, 1985; Ruthig et al., 2009). Previous research has found that the experience of sufficient social support positively develops peoples' sense of personal control (Ross & Broh, 2000). Therefore, efforts to strengthen social support networks of the lower class individuals may also play a potential preventative role. Some limitations of this study should be addressed. First, the crosssectional nature of our study limits the extent to which causal relationships among our variables may be pointed out. Future studies should test the SSC-aggression links using longitudinal or experimental designs. Second, the data we collected are based on students' self-reporting, which can be a potential threat to the internal validity of our findings. Future studies should collect data from multiple sources to minimize this potential effect. Third, the present sample was predominately female, which may limit the generalizability of our findings. Future studies should recruit larger samples and compare the indirect effects by gender. Finally, the strategies we used to delete the items of the constructs before the parceling process were relatively strict; hence, caution is recommended in interpreting the results of our analyses. Despite its limitations, our study sheds considerably light on the relationship between SSC and aggression. We expand on previous research and the CNA model by revealing indirect pathways comprising perceived social support, sense of control, and negative affect, from low SSC to aggression. Based on our results, we have furthermore provided a number of prevention as well as intervention suggestions aimed at treating aggression in individuals with low SSC. Future research should investigate their relative efficiencies and also try to replicate our findings in larger and more diverse samples.
This work was supported by the National Natural Science Foundation of China [grant number 31371055]; the Major Project for Key Research Institutes of Humanities and Social Science by the Ministry of Education [grant number 16JJD190007]. No competing financial interests existed. References Adler, N. E., Epel, E. S., Castellazzo, G., & Ickovics, J. R. (2000). Relationship of subjective and objective social status with psychological and physiological functioning: Preliminary data in healthy, White women. Health Psychology, 19(6), 586–592. http://dx.doi.org/10.1037/0278-6133.19.6.586. Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53(1), 27–51. http://dx.doi.org/10.1146/annurev.psych.53.100901. 135231. Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. http://dx.doi.org/10.1037/0033-2909.103.3.411. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–94. http://dx.doi.org/10.1007/ BF02723327. Bao, N. (2009). The relationship between self-worth and aggressive behavior in undergraduate of internet game addiction (Master's thesis)Jilin, China: Northeast Normal University. Barefoot, J. C., Peterson, B. L., Dahlstrom, W. G., Siegler, I. C., Anderson, N. B., & Williams, R. B., Jr. (1991). Hostility patterns and health implications: Correlates of cook-medley hostility scale scores in a national survey. Health Psychology, 10(1), 18–24. http://dx.doi.org/10.1037/0278-6133.10.1.18. Baumeister, R. F. (2005). The cultural animal: Human nature, meaning, and social life. New York: Oxford University Press. Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive-neoassociationistic analysis. American Psychologist, 45(4), 494–503. http://dx. doi.org/10.1037/0003-066X.45.4.494. Buss, A. H., & Perry, M. (1992). The aggression questionnaire. Journal of Personality and Social Psychology, 63(3), 452–459. http://dx.doi.org/10.1037/0022-3514.63.3.452. Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98(2), 310–357. http://dx.doi.org/10.1037/0033-2909.98.2. 310. Côté, S. (2011). How social class shapes thoughts and actions in organizations. Research in Organizational Behavior, 31, 43–71. http://dx.doi.org/10.1016/j.riob.2011.09.004. Ell, K. O., Mantell, J. E., Hamovitch, M. B., & Nishimoto, R. H. (1989). Social support, sense of control, and coping among patients with breast, lung, or colorectal cancer. Journal of Psychosocial Oncology, 7(3), 63–89. http://dx.doi.org/10.1300/ J077v07n03_05. Fliege, H., Becker, J., Walter, O. B., Bjorner, J. B., Klapp, B. F., & Rose, M. (2005). Development of a computer-adaptive test for depression (D-CAT). Quality of Life Research, 14(10), 2277–2291. http://dx.doi.org/10.1007/s11136-005-6651-9. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. http://dx.doi.org/10.2307/3151312. Gallo, L. C., & Matthews, K. A. (2003). Understanding the association between socioeconomic status and physical health: Do negative emotions play a role? Psychological Bulletin, 129(1), 10–51. http://dx.doi.org/10.1037/0033-2909.129.1.10. Goodman, E., Adler, N. E., Kawachi, I., Frazier, A. L., Huang, B., & Colditz, G. A. (2001). Adolescents' perceptions of social status: Development and evaluation of a new indicator. Pediatrics, 108(2), e31. http://dx.doi.org/10.1542/peds.108.2.e31. Greitemeyer, T., & Sagioglou, C. (2016). Subjective socioeconomic status causes aggression: A test of the theory of social deprivation. Journal of Personality and Social Psychology, 111(2), 178–194. http://dx.doi.org/10.1037/pspi0000058. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis (7th Edition). Upper Saddle River, NJ: Prentice Hall. Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408–420. http://dx.doi.org/10. 1080/03637750903310360. Henry, P. (2009). Low-status compensation: A theory for understanding the role of status in cultures of honor. Journal of Personality and Social Psychology, 97(3), 451–466. http://dx.doi.org/10.1037/a0015476. Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural equation modeling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53–60. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. http://dx.doi.org/10.1080/ 10705519909540118. Huang, C. C., Wang, Y. M., Wu, T. W., & Wang, P. A. (2013). An empirical analysis of the antecedents and performance consequences of using the moodle platform. International Journal of Information and Education Technology, 3(2), 217–221. http:// dx.doi.org/10.7763/IJIET.2013.V3.267. Jiang, Q. (1999). Perceived social support scale. Rating scales for mental health (pp. 131– 133). (Enlarged edition). Beijing: Chinese Mental Health Journal. Johnson, S. E., Richeson, J. A., & Finkel, E. J. (2011). Middle class and marginal? Socioeconomic status, stigma, and self-regulation at an elite university. Journal of
178
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
multidimensional composites: A comparison of four empirical parceling algorithms. Multivariate Behavioral Research, 39(3), 379–412. http://dx.doi.org/10.1207/ S15327906MBR3903_1. Ross, C. E., & Broh, B. A. (2000). The roles of self-esteem and the sense of personal control in the academic achievement process. Sociology of Education, 73(4), 270–284. http:// dx.doi.org/10.2307/2673234. Ruthig, J. C., Haynes, T. L., Stupnisky, R. H., & Perry, R. P. (2009). Perceived academic control: Mediating the effects of optimism and social support on college students' psychological health. Social Psychology of Education, 12(2), 233–249. http://dx.doi. org/10.1007/s11218-008-9079-6. Smith, H. J., Pettigrew, T. F., Pippin, G. M., & Bialosiewicz, S. (2012). Relative deprivation: A theoretical and meta-analytic review. Personality and Social Psychology Review, 16(3), 203–232. http://dx.doi.org/10.1177/1088868311430825. Sun, Y., Yu, Y., Luo, Y., & Yang, Y. (2011). Relationship between child aggressive behavior and emotion regulation. Chinese Journal of School Health, 32(8), 912–913. http://dx.doi.org/10.16835/j.cnki.1000-9817.2011.08.007. Tseng, W. T., Dörnyei, Z., & Schmitt, N. (2006). A new approach to assessing strategic learning: The case of self-regulation in vocabulary acquisition. Applied Linguistics, 27(1), 78–102. http://dx.doi.org/10.1093/applin/ami046. Ugulu, I. (2013). Confirmatory factor analysis for testing validity and reliability of traditional knowledge scale to measure university students' attitudes. Educational Research Review, 8(16), 1399–1408. http://dx.doi.org/10.5897/ERR2013.1537. Wangberg, S. C., Andreassen, H. K., Prokosch, H. U., Santana, S. M. V., Sørensen, T., & Chronaki, C. E. (2008). Relations between internet use, socio-economic status (SES), social support and subjective health. Health Promotion International, 23(1), 70–77. http://dx.doi.org/10.1093/heapro/dam039. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. http://dx.doi.org/10.1037/0022-3514.54.6. 1063. Whitson, J. A., & Galinsky, A. D. (2008). Lacking control increases illusory patter perception. Science, 322(5898), 115–117. http://dx.doi.org/10.1126/science.1159845. Williams, K. D. (2007). Ostracism. Annual Review of Psychology, 58(1), 425–452. http:// dx.doi.org/10.1146/annurev.psych.58.110405.085641. Wu, Y., & Wen, Z. (2011). Item parceling strategies in structural equation modeling. Advances in Psychological Science, 19(12), 1859–1867. http://dx.doi.org/10.3724/SP. J.1042.2011.01859. Yan, B., & Zheng, X. (2006). Researches into relations among social-support, self-esteem and subjective well-being of college students. Psychological Development and Education, 22(3), 53–59. http://dx.doi.org/10.16187/j.cnki.issn1001-4918.2006.03. 011. Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The multidimensional scale of perceived social support. Journal of Personality Assessment, 52(1), 30–41. http://dx.doi.org/10.1207/s15327752jpa5201_2.
Personality and Social Psychology, 100(5), 838–852. http://dx.doi.org/10.1037/ a0021956. Johnson, W., & Krueger, R. F. (2006). How money buys happiness: Genetic and environmental processes linking finances and life satisfaction. Journal of Personality and Social Psychology, 90(4), 680–691. http://dx.doi.org/10.1037/0022-3514.90.4.680. Kishton, J. M., & Widaman, K. F. (1994). Unidimensional versus domain representative parceling of questionnaire items: An empirical example. Educational and Psychological Measurement, 54(3), 757–765. http://dx.doi.org/10.1177/0013164494054003022. Korsgaard, M. A., & Roberson, L. (1995). Procedural justice in performance evaluation: The role of instrumental and non-instrumental voice in performance appraisal discussions. Journal of Management, 21(4), 657–669. http://dx.doi.org/10.1177/ 014920639502100404. Krahé, B. (2001). The social psychology of aggression. New York: Psychology Press. Kraus, M. W., Piff, P. K., & Keltner, D. (2009). Social class, sense of control, and social explanation. Journal of Personality and Social Psychology, 97(6), 992–1004. http://dx. doi.org/10.1037/a0016357. Lachman, M. E., & Weaver, S. L. (1998). The sense of control as a moderator of social class differences in health and well-being. Journal of Personality and Social Psychology, 74(3), 763–773. http://dx.doi.org/10.1037/0022-3514.74.3.763. Li, J. (2012). The study of attribution tendency in the rich-poor gap for different social classes (Unpublished doctoral dissertation)Hubei, China: Central China Normal University. McFarlin, S. K., Fals-Stewart, W., Major, D. A., & Justice, E. M. (2001). Alcohol use and workplace aggression: An examination of perpetration and victimization. Journal of Substance Abuse, 13(3), 303–321. http://dx.doi.org/10.1016/S0899-3289(01) 00080-3. Mueller, C. W. (1983). Environmental stressors and aggressive behavior. Aggression: Theoretical and empirical reviews, vol. 2: Issues in research (pp. 51–76). New York: Academic Press. Nelson, T. (2009). Handbook of threat, stereotype, and discrimination. New York: Psychology Press. Ong, A. D., Bergeman, C. S., & Bisconti, T. L. (2005). Unique effects of daily perceived control on anxiety symptomatology during conjugal bereavement. Personality and Individual Differences, 38(5), 1057–1067. http://dx.doi.org/10.1016/j.paid.2004.07. 004. Operario, D., Adler, N. E., & Williams, D. R. (2004). Subjective social status: Reliability and predictive utility for global health. Psychology & Health, 19(2), 237–246. http:// dx.doi.org/10.1080/08870440310001638098. Peltzer, K., Pengpid, S., & James, C. (2016). The globalization of whitening: Prevalence of skin lighteners (or bleachers) use and its social correlates among university students in 26 countries. International Journal of Dermatology, 55(2), 165–172. http://dx.doi. org/10.1111/ijd.12860. Qiu, L., Zheng, X., & Wang, Y. (2008). Revision of the positive affect and negative affect scale. Chinese Journal of Applied Psychology, 14(3), 249–254. Rogers, W. M., & Schmitt, N. (2004). Parameter recovery and model fit using
179
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
The relationship between subjective social class and aggression: A serial mediation model Bing Chena, Youxia Zuoa, Yufang Zhaoa,b, a b
T
⁎
Faculty of Psychology, Southwest University, 2 Tiansheng Road, Beibei, ChongQing 400715, China Center for Studies of Education and Psychology of Minorities in Southwest China, Southwest University, Chongqing, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Subjective social class Perceived social support Sense of control Negative affect Aggression Serial mediation
Although the relationship between the phenomena of social class and aggression is well known, very few studies have been conducted to investigate its underlying mechanisms. This study examines the relationship between subjective social class (SSC) and aggression, proposing a serial mediation model for the relationship. A total of 305 valid participants, ranging in age from 17 to 26, completed questionnaires assessing SSC, perceived social support, sense of control, negative affect, and aggression. Serial mediation analysis revealed a pathway whereby lower SSC was associated with less perceived social support, which was associated with decreased sense of control and increased negative affect, which were then associated with more aggression. Our results shed light on the associations between the above-mentioned variables in the SSC–aggression relationship. We also suggest possible prevention and intervention programs for reducing aggression among individuals with low SSC and suggest courses for future inquiry.
1. Introduction Aggression is a major contemporary social problem and may be seen as a destructive means of modern conflict. It is also a complex phenomenon, typically said to be made up by biological, environmental, psychological, and social factors (Anderson & Bushman, 2002). Social class is considered a powerful social factor, and low social class has been associated with increased levels of aggression, for example, lowincome (McFarlin, Fals-Stewart, Major, & Justice, 2001) and low-education groups (Barefoot et al., 1991) were found to be relatively more aggressive than high-income and high-education groups. People's subjective perceptions of their social class can also predict aggression. One study found that participants perceiving their social class as relatively lower behaved more aggressively relative to those who perceived theirs as relatively higher (Greitemeyer & Sagioglou, 2016). Despite the existing evidence, further empirical work is needed to explore the underlying mechanisms of the relationship between social class and aggression. The current study thus investigated whether perceived social support, sense of control, and negative affect mediate this relationship. Social class, which in academia is often used interchangeably with socioeconomic status (Côté, 2011), is a multifaceted construct comprising an individual's material resources as well as his or her perceived rank within the social hierarchy (Kraus, Piff, & Keltner, 2009). Researchers often assess the construct as objective social class (OSC) or
⁎
subjective social class (SSC); typical measures of the former include income, education, and profession (Goodman et al., 2001), while the latter emphasizes individuals' perceived rank relative to others in society (Kraus et al., 2009). SSC can be seen as a comparative perception of, for example, material and social resources, i.e., individuals' sense of what they have or possess relative to others. Several studies have demonstrated that SSC is positively and moderately correlated with OSC (Adler, Epel, Castellazzo, & Ickovics, 2000; Johnson & Krueger, 2006), and that it is a relatively better predictor of psychological outcome than OSC (Adler et al., 2000; Kraus et al., 2009). In the present study, we thus assess social class subjectively and mainly focus on mechanisms pertaining to the relation between SSC and aggression. According to the cognitive neoassociation model (CNA model; Berkowitz, 1990), aversive events that lead to negative affect can lead to aggression. Given their limited resources and lower social rank, lower class individuals are more likely to experience threats (Nelson, 2009), to be characterized by chronic levels of cynical mistrust and hostility (Gallo & Matthews, 2003), and to feel more socially rejected (Johnson, Richeson, & Finkel, 2011) than higher class individuals. Therefore, they are at risk of suffering chronically from exposure to aversive events, which, in turn, may foster negative affect and hence increase the risk of behaving aggressively. The CNA model thus plausibly explains class differences in aggression. However, the question remains how aversive events influence or bring about the subsequent
Corresponding author at: Faculty of Psychology, Southwest University, 2 Tiansheng Road, Beibei, ChongQing 400715, China. E-mail address: [email protected] (Y. Zhao).
https://doi.org/10.1016/j.paid.2018.04.036 Received 15 September 2017; Received in revised form 18 March 2018; Accepted 23 April 2018 0191-8869/ © 2018 Published by Elsevier Ltd.
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
have the most money, the most education, and the most respected jobs. At the bottom the people who are the worst off – who have the least money, the least education, and the least respected jobs or no jobs”. Then they were instructed to select a number representing their perception of their family's placement on this 10-point social scale, with higher numbers indicating higher perceived social class. This measure is widely used and has demonstrated adequate test-retest reliability (Operario, Adler, & Williams, 2004). We used family information instead of participants' own information as college students typically are not financially independent and their status thus based on their upbringing (Henry, 2009).
experiences of negative affect. In the present study, we investigate the potential roles of perceived social support and sense of control in this process. Social support, thought to mitigate the effects of stressful life events on mental health (Cohen & Wills, 1985), may to some extent depend on social class. Previous studies have demonstrated that male students from lower income countries report a lack of social support (Peltzer, Pengpid, & James, 2016), and that social class positively predicts social support (Wangberg et al., 2008). Hence, relatively lower class individuals are more likely to perceive lower levels of social support. This absence of social support, however, is often furthermore associated with a decreased sense of control (Ell, Mantell, Hamovitch, & Nishimoto, 1989; Ruthig, Haynes, Stupnisky, & Perry, 2009). Sense of control is a concept that refers to peoples' beliefs about the extent to which they are able to shape their own social outcomes (Lachman & Weaver, 1998). It is further seen as a fundamental social need of humans (Williams, 2007), and a perceived lack of control is associated with increased experiences of negative affect. Studies have shown that people, when they perceive a lack of control, feel upset or unhappy (Baumeister, 2005), and that lower levels of control correlate with higher levels of anxiety (Ong, Bergeman, & Bisconti, 2005). Previous studies have thus demonstrated that differences in aggression appear to be contingent on experiences of negative or aggression-related affect (Greitemeyer & Sagioglou, 2016; Smith, Pettigrew, Pippin, & Bialosiewicz, 2012), and the CNA model also suggests that negative affect rather than aversive events should serve as the proximate mechanism for eliciting aggression (Berkowitz, 1990). Hence, we hypothesize that living with low SSC may decrease levels of perceived social support and subsequently sense of control, resulting in more experiences of negative affect; this, in turn, may lead to the final aggression. Guided by the CNA model and previous results, the present study aimed to elucidate potential mechanisms underlying the relationship between SSC and aggression, proposing a serial mediation model for this relationship (see Fig. 1). We expand previous research by providing a comprehensive test of the serial process of how SSC influences aggression through negative affect. We also sought to provide several prevention and intervention methods that may help reduce aggression and maintain social stability.
2.2.2. Perceived social support The Chinese version of Perceived Social Support Scale (Jiang, 1999; Zimet, Dahlem, Zimet, & Farley, 1988) was used to measure participants' perceived social support. It consists 12 items designed to assess three sources of support: significant other (leaders, relatives, and colleagues), family, and friends. The examples of significant other were changed into “teachers, relatives, and classmates” which were more suitable for our sample (Yan & Zheng, 2006). Each subscale consists of four items that are answered on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). In this study, Cronbach's α for the subscales and the whole scale was 0.81, 0.82, 0.88, and 0.91. 2.2.3. Sense of control The Chinese version of Sense of Control Scale (Lachman & Weaver, 1998; Li, 2012) was used to assess participants' sense of control. This scale includes the two subscales: personal mastery and perceived constraints. The items are rated on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). The perceived constraints scale is reverse scored. In this study, Cronbach's α for the subscales and the whole scale was 0.70, 0.79, and 0.80. 2.2.4. Negative affect The Negative Affect Scale was selected from a subscale of the Chinese version Positive and Negative Affect Scale (Qiu, Zheng, & Wang, 2008; Watson, Clark, & Tellegen, 1988) to assess participants' negative affect. This scale consists of nine adjectives describing negative emotions (e.g., anger, nervous), which are rated on a 5-point Likert scale (1 = very slightly or not at all, 5 = extremely) based on how one has felt during the past week. In this study, Cronbach's α for the scale was 0.85.
2. Methods 2.1. Participants and procedure The participants were 305 college students (248 females and 57 males) from a university in southwest China, aged 17–26 (M = 20.59, SD = 1.99). They were recruited via QQ (a famous chat tool in China), and participated voluntarily. Participants first gave written informed consent, after which they completed an anonymous questionnaire in about 10 min in a laboratory room. Subsequent to completion, students were compensated with a small gift (approximately 1.50 USD).
2.2.5. Aggression The Chinese version of Aggression Questionnaire (Bao, 2009; Buss & Perry, 1992) was used to assess participants' aggression. It consists of the four subscales: physical aggression, verbal aggression, anger, and hostility. The items are answered on a 5-point Likert scale (1 = definitely doesn't apply to me, 5 = definitely applies to me). In this study, Cronbach's α for the subscales and the whole scale was 0.77, 0.49, 0.79, 0.70, and 0.87.
2.2. Measures
2.3. Analysis strategy
2.2.1. Subjective social class (SSC) The MacArthur Scale of subjective socioeconomic status (Adler et al., 2000) was used to measure SSC. The measure consists of a drawing of a ladder with 10 rungs representing people with different levels of income, education, and occupational status. Each rung of the ladder was given a number between 1 and 10. Participants were told: “At the top of the ladder are the people who are the best off – those who
According to Anderson and Gerbing's (1988) suggestion, we adopted the two-step strategy for analysis of mediation effects. First, in order to assess the extent to which each of the latent variables was represented by its indicators, the measurement model was confirmed using confirmatory factor analysis (CFA). Second, structural equation modeling (SEM) analysis was performed to measure the fit and path coefficients of the hypothesized structural model. Furthermore, biasFig. 1. Hypothesized model Note: SSC = subjective social class, PSS = perceived social support, SOC = sense of control, NA = negative affect.
175
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
Table 1 Reliability and validity.
1.SSC 2.NA 3.PSS 4.SOC 5.Aggression
CR
AVE
MSV
1
2
3
4
5
– 0.81 0.63 0.60 0.76
– 0.60 0.46 0.45 0.45
– 0.28 0.17 0.36 0.36
– −0.02 0.16 0.23⁎⁎⁎ −0.15⁎
0.77 −0.18 −0.45⁎⁎⁎ 0.53⁎⁎⁎
0.68 0.42⁎⁎⁎ −0.29⁎
0.67 −0.60⁎⁎⁎
0.67
(1) ⁎p < 0.05; ⁎⁎p < 0.01; ⁎⁎⁎p < 0.001 (two-tailed); N = 305. (2) Diagonal elements in boldface are the square root of AVE; Off-diagonal elements are the correlation coefficients.
SRMR = 0.05, and all the indicators loaded significantly on their corresponding factors. Second, reliability, convergent validity, and discriminant validity were tested by the measures of composite reliability (CR), average variance extracted (AVE), maximum shared squared variance (MSV), and inter-construct correlation. Good reliability is indicated by a CR above 0.70 (Hair, Black, Babin, & Anderson, 2010), with values above 0.60 deemed acceptable (Tseng, Dörnyei, & Schmitt, 2006). AVE should be at above 0.50 in order to establish convergent validity, and values above 0.40 are acceptable (Fornell & Larcker, 1981; Huang, Wang, Wu, & Wang, 2013). MSV should be less than AVE, and the inter-construct correlations should be less than the square root of AVE in order to establish the discriminant validity (Hair et al., 2010). Regarding this study, the results in Table 1 shows that all the indices have reached acceptable levels. Third, we employed the Harman one-factor test in CFA (Korsgaard & Roberson, 1995) to assess the common method biases (CMB), if CMB is a serious problem, the single factor model should fit the data as well as a more complex model. The results revealed a deteriorated model fit with χ2/df = 9.56, GFI = 0.75, AGFI = 0.65, CFI = 0.51, TLI = 0.41, RMSEA = 0.17, and SRMR = 0.16. This indicated that CMB was not a major concern in this study.
corrected bootstrap confidence intervals based on 5000 bootstraps (Hayes, 2009) were calculated to test the significance of the mediating effects. The following indices were utilized to evaluate the goodness of fit of the model (see Hooper, Coughlan, & Mullen, 2008; Hu & Bentler, 1999): (a) value of χ2/df of 5 or less; (b) Goodness of fit index (GFI) of 0.90 or more; (c) Adjusted GFI (AGFI) of 0.90 or more; (d) Comparative fit index (CFI) of 0.90 or more; (e) Tucker–Lewis index (TLI) of 0.90 or more; (f) Root-mean-square error of approximation (RMSEA) of 0.08 or less; and (g) Standardized root-mean-square residual (SRMR) of 0.08 or less. 3. Results 3.1. Preliminary analyses To reduce the number of parameters in the SEM analysis and to increase the accuracy of parameter estimates, we parceled the items on each measure (except for SSC, which was a single value). We used the internal-consistency approach (Kishton & Widaman, 1994) to parcel the items on perceived social support, sense of control, and aggression that all had multidimensional item sets. This approach creates different parcels that use the facets as grouping criteria. That is, each subscale or dimension of these scales were parceled into one manifest indicator. We used the factorial algorithm approach (Rogers & Schmitt, 2004) to parcel the items on negative affect that did not have multidimensional item sets. This approach rank orders the manifest indicators using their loadings on the first principal axis factor, and each parcel was sequentially assigned the remaining indicators with the highest and lowest rankings, alternating direction through the parcels, until all indicators were assigned. In order to ensure the unidimensionality of our parceled targets, we conducted several independent CFAs for negative affect, and each of the subscales of perceived social support, sense of control, and aggression before the parceling as recommended (Wu & Wen, 2011). Six items were eliminated as their factor loadings were below 0.40 (Fliege et al., 2005). The output of the CFA also suggested that modification indexes of ten pairs of items were extensively high, indicating a certain degree of overlap. This prompted the elimination of one item in each pair from the subsequent analyses (Ugulu, 2013).
3.3. Structural model testing The structural-modeling results indicated that the hypothesized model fit the data well, χ2/df = 2.47, GFI = 0.94, AGFI = 0.90, CFI = 0.93, TLI = 0.90, RMSEA = 0.07, and SRMR = 0.05. Fig. 2 shows the properties of the research hypothesis, including all the standardized path coefficients and factor loadings. 3.4. Test of mediating effects The bootstrap procedure was used to examine the significance levels of indirect effects for the hypothesized model. Results, shown in Table 2, indicate that the total effect of SSC on aggression was significant, β = −0. 1513, SE = 0.0637, 95% CI [−0.2722, −0.0255]. When the mediators were included in the analysis, this coefficient was no longer statistically significant (direct effect), β = −0.0414, SE = 0.0633, 95% CI [−0.1578, 0.0879]. In line with our hypothesis, the indirect effect of the SSC on aggression via perceived social support, sense of control, and negative affect was significant, β = −0.0098, SE = 0.0085, 95% CI [−0.0469, −0.0008]. Specifically, SSC significantly and positively predicted perceived social support, perceived social support significantly and positively predicted sense of control, sense of control significantly and negatively predicted negative affect, and negative affect significantly and positively predicted aggression (for standardized path coefficients, see Fig. 2). Moreover, the indirect effect of SSC on aggression through perceived social support and sense of control was significant, β = −0.0264, SE = 0.0260, 95% CI [−0.1332, −0.0004]. The indirect effect of SSC on aggression via sense of control alone was also significant, β = −0.0706, SE = 0.0435, 95% CI [−0.1843, −0.0137]. And the indirect effect of SSC on aggression via sense of control and
3.2. Measurement model testing First, we tested the fit of the measurement model, χ2/df = 2.19, GFI = 0.94, AGFI = 0.90, CFI = 0.95, TLI = 0.93, RMSEA = 0.06, and SRMR = 0.05. We found a nonsignificant variance in error in one factor in the estimate (the parceled item “significant other” of perceived social support scale). According to Bagozzi and Yi's (1988) suggestion, nonsignificant error variances usually suggest specification errors, and one solution when it occurs is to drop a measurement corresponding to the nonsignificant error. We thus deleted this item and reran the analysis. The estimates of the overall goodness-of-fit criteria exceeded their acceptable levels, which confirmed that the measurement model exhibited a fairly good fit with the data, χ2/df = 2.47, GFI = 0.94, AGFI = 0.90, CFI = 0.93, TLI = 0.90, RMSEA = 0.07, and 176
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
Fig. 2. The structural model. Note: Family = family support, Friend = friend support, PM = personal mastery, PC = perceived constraints, NA1-NA3 = 3 parcels of negative affect, A = anger, PA = physical aggression, VA = verbal aggression, H = hostility.
negative affect was significant as well, β = −0.0263, SE = 0.0163, 95% CI [−0.0685, −0.0055]. However, neither the indirect effect of SSC on aggression via perceived social support or negative affect alone, nor a chain of these two variables was significant, as witnessed by 95% CIs that included zero (see Table 2).
which is facilitative of aggression. However, the aversive events that lower class individuals experience may furthermore facilitate lower perceptions of their social class rank, making up what may be considered a vicious circle. Therefore, policies for attenuating the actual class inequality may help to reduce instances of interpersonal hostility. Importantly, our results highlight the significance of sense of control in the relationship between SSC and aggression. We found that the three-mediator pathway (via perceived social support, sense of control, and negative affect), the two-mediator pathways (via perceived social support and sense of control, and via sense of control and negative affect), and the single-mediator pathway (via sense of control) were all statistically significant. The indirect effects via perceived social support and negative affect, and via either alone were not. This suggests that perceived social support, sense of control, and negative affect all act as mediators, influencing the internal mechanism of the SSC-aggression relationship, with sense of control being the more important factor. In addition, our results illustrate that SSC may influence aggression directly through sense of control. When people report that they are of low SSC, they are indicating that they have fewer resources and are of subordinate rank relative to others. These self-perceptions of reduced resources and subordinate rank are often associated with a diminished sense of control (Kraus et al., 2009). As a result of this reduced control, people may behave aggressively, possibly as a means to regain or increase their sense of control (Mueller, 1983). Therefore, increasing the perceived control of lower class individuals, such as recalling a
4. Discussion The main purpose of this study was to examine the possible roles of perceived social support, sense of control, and negative affect as mediators of the relationship between SSC and aggression. Our results suggest that a lack of perceived social support, resulting from lower SSC, decreases sense of control and, in turn, is associated with more negative affect and finally greater aggression. This indirect path was statistically significant, rendering the direct path from SSC to aggression statistically insignificant. This indicates that the relationship between the SSC and aggression is fully, serially mediated by perceived social support, sense of control, and negative affect. The negative relationship between SSC and aggression is consistent with previous studies (Greitemeyer & Sagioglou, 2016; McFarlin et al., 2001). Compared to people with higher SSC, people with lower SSC report more aggression; as outlined above, this finding can be explained by the CNA model (Berkowitz, 1990). Lower class individuals often suffer chronically from aversive-event exposure (Johnson et al., 2011; Nelson, 2009), and they thus tend to experience more negative affect, Table 2 Indirect effects, direct effect, and total effect. Point estimate
Ind1(SSC–PSS–aggression) Ind2(SSC–PSS–SOC–aggression) Ind3(SSC–PSS–NA–aggression) Ind4(SSC–PSS–SOC–NA–aggression) Ind5(SSC–SOC–aggression) Ind6(SSC–SOC–NA–aggression) Ind7(SSC–NA–aggression) Total indirect effect Direct effect Total effect
−0.0072 −0.0264 0.0005 −0.0098 −0.0706 −0.0263 0.0298 −0.1100 −0.0414 −0.1513
SE
BC bootstrap 95% CI
0.0231 0.0260 0.0073 0.0085 0.0435 0.0163 0.0240 0.0549 0.0633 0.0637
Lower
Upper
(two-tailed)
−0.0742 −0.1332 −0.0145 −0.0469 −0.1843 −0.0685 −0.0070 −0.2299 −0.1578 −0.2722
0.0213 −0.0004 0.0147 −0.0008 −0.0137 −0.0055 0.0882 −0.0135 0.0879 −0.0255
0.3990 0.0444 0.9333 0.0300 0.0144 0.0144 0.1099 0.0259 0.5336 0.0211
Note: All the values of point estimates are standardized; BC bootstrap = bias-corrected bootstrap; 5000 bootstrap samples. 177
p
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
Acknowledgement
situation in which they had a great deal of control (Kraus et al., 2009), is likely to reduce their aggression. Future studies should investigate potential beneficial interventions, including at the preventative level. Our results also expand the CNA model by demonstrating how SSC influences aggression through negative affect. That is, people with lower SSC tend to perceive less social support and less sense of control. As previous studies have found that a lack of control may be considered an aversive state or experience (Whitson & Galinsky, 2008), these people may then subsequently experience varying degrees of negative affect (Baumeister, 2005; Ong et al., 2005), which, in turn, is associated with the final aggression (Greitemeyer & Sagioglou, 2016; Smith et al., 2012). The CNA model (Berkowitz, 1990) also implies that aversive experiences can lead to aggressive responding through a multistage process. In the first stage, the aversive event will lead to an undifferentiated negative affective state. This reaction will stimulate fight or flight tendencies, with the fight tendency being related to both aggression and rudimentary feelings of anger. Flight, on the other hand, is escape-related and will give rise to rudimentary feelings of fear. In the second stage, higher-order cognitive processing comes into play. As people think about what happened and consider the possible consequences, the rudimentary emotions of anger or fear are differentiated into more elaborate emotional states, like annoyance, irritation, anger, etc. This implies that the aggression is not inevitable and may be either promoted or suppressed by the emotional experience elicited by the aversive event (Krahé, 2001). Similarly to this viewpoint, Sun, Yu, Luo, and Yang (2011) found that children characterized by low levels of aggression scored higher on the measures of emotion regulation than those characterized by high levels of aggression. Thus, enhancing emotion-management skills of individuals who perceive their social class as low may decrease their potential aggressive tendencies to some extent. Of note, the positive association between perceived social support and sense of control in the present study is consistent with previous studies (Ell et al., 1989; Ruthig et al., 2009). Social support may be said to act as a psychosocial resource that can buffer individuals against stress and depression and help people control what happens to them in their daily life (Cohen & Wills, 1985; Ruthig et al., 2009). Previous research has found that the experience of sufficient social support positively develops peoples' sense of personal control (Ross & Broh, 2000). Therefore, efforts to strengthen social support networks of the lower class individuals may also play a potential preventative role. Some limitations of this study should be addressed. First, the crosssectional nature of our study limits the extent to which causal relationships among our variables may be pointed out. Future studies should test the SSC-aggression links using longitudinal or experimental designs. Second, the data we collected are based on students' self-reporting, which can be a potential threat to the internal validity of our findings. Future studies should collect data from multiple sources to minimize this potential effect. Third, the present sample was predominately female, which may limit the generalizability of our findings. Future studies should recruit larger samples and compare the indirect effects by gender. Finally, the strategies we used to delete the items of the constructs before the parceling process were relatively strict; hence, caution is recommended in interpreting the results of our analyses. Despite its limitations, our study sheds considerably light on the relationship between SSC and aggression. We expand on previous research and the CNA model by revealing indirect pathways comprising perceived social support, sense of control, and negative affect, from low SSC to aggression. Based on our results, we have furthermore provided a number of prevention as well as intervention suggestions aimed at treating aggression in individuals with low SSC. Future research should investigate their relative efficiencies and also try to replicate our findings in larger and more diverse samples.
This work was supported by the National Natural Science Foundation of China [grant number 31371055]; the Major Project for Key Research Institutes of Humanities and Social Science by the Ministry of Education [grant number 16JJD190007]. No competing financial interests existed. References Adler, N. E., Epel, E. S., Castellazzo, G., & Ickovics, J. R. (2000). Relationship of subjective and objective social status with psychological and physiological functioning: Preliminary data in healthy, White women. Health Psychology, 19(6), 586–592. http://dx.doi.org/10.1037/0278-6133.19.6.586. Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53(1), 27–51. http://dx.doi.org/10.1146/annurev.psych.53.100901. 135231. Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. http://dx.doi.org/10.1037/0033-2909.103.3.411. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–94. http://dx.doi.org/10.1007/ BF02723327. Bao, N. (2009). The relationship between self-worth and aggressive behavior in undergraduate of internet game addiction (Master's thesis)Jilin, China: Northeast Normal University. Barefoot, J. C., Peterson, B. L., Dahlstrom, W. G., Siegler, I. C., Anderson, N. B., & Williams, R. B., Jr. (1991). Hostility patterns and health implications: Correlates of cook-medley hostility scale scores in a national survey. Health Psychology, 10(1), 18–24. http://dx.doi.org/10.1037/0278-6133.10.1.18. Baumeister, R. F. (2005). The cultural animal: Human nature, meaning, and social life. New York: Oxford University Press. Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive-neoassociationistic analysis. American Psychologist, 45(4), 494–503. http://dx. doi.org/10.1037/0003-066X.45.4.494. Buss, A. H., & Perry, M. (1992). The aggression questionnaire. Journal of Personality and Social Psychology, 63(3), 452–459. http://dx.doi.org/10.1037/0022-3514.63.3.452. Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98(2), 310–357. http://dx.doi.org/10.1037/0033-2909.98.2. 310. Côté, S. (2011). How social class shapes thoughts and actions in organizations. Research in Organizational Behavior, 31, 43–71. http://dx.doi.org/10.1016/j.riob.2011.09.004. Ell, K. O., Mantell, J. E., Hamovitch, M. B., & Nishimoto, R. H. (1989). Social support, sense of control, and coping among patients with breast, lung, or colorectal cancer. Journal of Psychosocial Oncology, 7(3), 63–89. http://dx.doi.org/10.1300/ J077v07n03_05. Fliege, H., Becker, J., Walter, O. B., Bjorner, J. B., Klapp, B. F., & Rose, M. (2005). Development of a computer-adaptive test for depression (D-CAT). Quality of Life Research, 14(10), 2277–2291. http://dx.doi.org/10.1007/s11136-005-6651-9. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. http://dx.doi.org/10.2307/3151312. Gallo, L. C., & Matthews, K. A. (2003). Understanding the association between socioeconomic status and physical health: Do negative emotions play a role? Psychological Bulletin, 129(1), 10–51. http://dx.doi.org/10.1037/0033-2909.129.1.10. Goodman, E., Adler, N. E., Kawachi, I., Frazier, A. L., Huang, B., & Colditz, G. A. (2001). Adolescents' perceptions of social status: Development and evaluation of a new indicator. Pediatrics, 108(2), e31. http://dx.doi.org/10.1542/peds.108.2.e31. Greitemeyer, T., & Sagioglou, C. (2016). Subjective socioeconomic status causes aggression: A test of the theory of social deprivation. Journal of Personality and Social Psychology, 111(2), 178–194. http://dx.doi.org/10.1037/pspi0000058. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis (7th Edition). Upper Saddle River, NJ: Prentice Hall. Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408–420. http://dx.doi.org/10. 1080/03637750903310360. Henry, P. (2009). Low-status compensation: A theory for understanding the role of status in cultures of honor. Journal of Personality and Social Psychology, 97(3), 451–466. http://dx.doi.org/10.1037/a0015476. Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural equation modeling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53–60. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. http://dx.doi.org/10.1080/ 10705519909540118. Huang, C. C., Wang, Y. M., Wu, T. W., & Wang, P. A. (2013). An empirical analysis of the antecedents and performance consequences of using the moodle platform. International Journal of Information and Education Technology, 3(2), 217–221. http:// dx.doi.org/10.7763/IJIET.2013.V3.267. Jiang, Q. (1999). Perceived social support scale. Rating scales for mental health (pp. 131– 133). (Enlarged edition). Beijing: Chinese Mental Health Journal. Johnson, S. E., Richeson, J. A., & Finkel, E. J. (2011). Middle class and marginal? Socioeconomic status, stigma, and self-regulation at an elite university. Journal of
178
Personality and Individual Differences 131 (2018) 174–179
B. Chen et al.
multidimensional composites: A comparison of four empirical parceling algorithms. Multivariate Behavioral Research, 39(3), 379–412. http://dx.doi.org/10.1207/ S15327906MBR3903_1. Ross, C. E., & Broh, B. A. (2000). The roles of self-esteem and the sense of personal control in the academic achievement process. Sociology of Education, 73(4), 270–284. http:// dx.doi.org/10.2307/2673234. Ruthig, J. C., Haynes, T. L., Stupnisky, R. H., & Perry, R. P. (2009). Perceived academic control: Mediating the effects of optimism and social support on college students' psychological health. Social Psychology of Education, 12(2), 233–249. http://dx.doi. org/10.1007/s11218-008-9079-6. Smith, H. J., Pettigrew, T. F., Pippin, G. M., & Bialosiewicz, S. (2012). Relative deprivation: A theoretical and meta-analytic review. Personality and Social Psychology Review, 16(3), 203–232. http://dx.doi.org/10.1177/1088868311430825. Sun, Y., Yu, Y., Luo, Y., & Yang, Y. (2011). Relationship between child aggressive behavior and emotion regulation. Chinese Journal of School Health, 32(8), 912–913. http://dx.doi.org/10.16835/j.cnki.1000-9817.2011.08.007. Tseng, W. T., Dörnyei, Z., & Schmitt, N. (2006). A new approach to assessing strategic learning: The case of self-regulation in vocabulary acquisition. Applied Linguistics, 27(1), 78–102. http://dx.doi.org/10.1093/applin/ami046. Ugulu, I. (2013). Confirmatory factor analysis for testing validity and reliability of traditional knowledge scale to measure university students' attitudes. Educational Research Review, 8(16), 1399–1408. http://dx.doi.org/10.5897/ERR2013.1537. Wangberg, S. C., Andreassen, H. K., Prokosch, H. U., Santana, S. M. V., Sørensen, T., & Chronaki, C. E. (2008). Relations between internet use, socio-economic status (SES), social support and subjective health. Health Promotion International, 23(1), 70–77. http://dx.doi.org/10.1093/heapro/dam039. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. http://dx.doi.org/10.1037/0022-3514.54.6. 1063. Whitson, J. A., & Galinsky, A. D. (2008). Lacking control increases illusory patter perception. Science, 322(5898), 115–117. http://dx.doi.org/10.1126/science.1159845. Williams, K. D. (2007). Ostracism. Annual Review of Psychology, 58(1), 425–452. http:// dx.doi.org/10.1146/annurev.psych.58.110405.085641. Wu, Y., & Wen, Z. (2011). Item parceling strategies in structural equation modeling. Advances in Psychological Science, 19(12), 1859–1867. http://dx.doi.org/10.3724/SP. J.1042.2011.01859. Yan, B., & Zheng, X. (2006). Researches into relations among social-support, self-esteem and subjective well-being of college students. Psychological Development and Education, 22(3), 53–59. http://dx.doi.org/10.16187/j.cnki.issn1001-4918.2006.03. 011. Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The multidimensional scale of perceived social support. Journal of Personality Assessment, 52(1), 30–41. http://dx.doi.org/10.1207/s15327752jpa5201_2.
Personality and Social Psychology, 100(5), 838–852. http://dx.doi.org/10.1037/ a0021956. Johnson, W., & Krueger, R. F. (2006). How money buys happiness: Genetic and environmental processes linking finances and life satisfaction. Journal of Personality and Social Psychology, 90(4), 680–691. http://dx.doi.org/10.1037/0022-3514.90.4.680. Kishton, J. M., & Widaman, K. F. (1994). Unidimensional versus domain representative parceling of questionnaire items: An empirical example. Educational and Psychological Measurement, 54(3), 757–765. http://dx.doi.org/10.1177/0013164494054003022. Korsgaard, M. A., & Roberson, L. (1995). Procedural justice in performance evaluation: The role of instrumental and non-instrumental voice in performance appraisal discussions. Journal of Management, 21(4), 657–669. http://dx.doi.org/10.1177/ 014920639502100404. Krahé, B. (2001). The social psychology of aggression. New York: Psychology Press. Kraus, M. W., Piff, P. K., & Keltner, D. (2009). Social class, sense of control, and social explanation. Journal of Personality and Social Psychology, 97(6), 992–1004. http://dx. doi.org/10.1037/a0016357. Lachman, M. E., & Weaver, S. L. (1998). The sense of control as a moderator of social class differences in health and well-being. Journal of Personality and Social Psychology, 74(3), 763–773. http://dx.doi.org/10.1037/0022-3514.74.3.763. Li, J. (2012). The study of attribution tendency in the rich-poor gap for different social classes (Unpublished doctoral dissertation)Hubei, China: Central China Normal University. McFarlin, S. K., Fals-Stewart, W., Major, D. A., & Justice, E. M. (2001). Alcohol use and workplace aggression: An examination of perpetration and victimization. Journal of Substance Abuse, 13(3), 303–321. http://dx.doi.org/10.1016/S0899-3289(01) 00080-3. Mueller, C. W. (1983). Environmental stressors and aggressive behavior. Aggression: Theoretical and empirical reviews, vol. 2: Issues in research (pp. 51–76). New York: Academic Press. Nelson, T. (2009). Handbook of threat, stereotype, and discrimination. New York: Psychology Press. Ong, A. D., Bergeman, C. S., & Bisconti, T. L. (2005). Unique effects of daily perceived control on anxiety symptomatology during conjugal bereavement. Personality and Individual Differences, 38(5), 1057–1067. http://dx.doi.org/10.1016/j.paid.2004.07. 004. Operario, D., Adler, N. E., & Williams, D. R. (2004). Subjective social status: Reliability and predictive utility for global health. Psychology & Health, 19(2), 237–246. http:// dx.doi.org/10.1080/08870440310001638098. Peltzer, K., Pengpid, S., & James, C. (2016). The globalization of whitening: Prevalence of skin lighteners (or bleachers) use and its social correlates among university students in 26 countries. International Journal of Dermatology, 55(2), 165–172. http://dx.doi. org/10.1111/ijd.12860. Qiu, L., Zheng, X., & Wang, Y. (2008). Revision of the positive affect and negative affect scale. Chinese Journal of Applied Psychology, 14(3), 249–254. Rogers, W. M., & Schmitt, N. (2004). Parameter recovery and model fit using
179