Although it has long been established that SVO predicts resource allocation behaviour, previous studies have tended simply to accept this relation without examining the proximal causes of why prosocials and proselfs differ in the way they allocate resources between self and other. The studies reported in this thesis shed some light on why there is this difference: Prosocials and proselfs differ in their allocation behaviour because they anticipate that they will feel different kinds of emotion as a result of allocating resources in a way that is fair or in a way that favours the self over the other person. The results of the current studies show that prosocials anticipate more cooperative emotions (e.g., pride about being fair and regret and guilt about being unfair) and less competitive emotions (e.g., pride about being unfair and regret and guilt about being fair), and this pattern of anticipated emotion is linked to a more equal allocation of resources between themselves and the recipients. Proselfs, on the other hand, anticipate more competitive emotions (e.g., pride about being unfair and regret and guilt about being fair) and less cooperative emotions (e.g., pride about being fair and regret and guilt about being unfair), and this pattern of anticipated emotion is linked to a more unequal allocation of resources between them and the recipient. In all reported studies there was a consistent indirect effect of SVO on allocation behaviour via anticipated emotions and this (at least in part) answers the research question about what goes on in the minds of prosocials and proselfs when making these resource allocation decisions. Below I discuss the theoretical implications of these findings.
As noted above, research using SVO measures has shown that individuals who differ in SVO scores have different preferences for certain outcomes (Balliet et al., 2009; Murphy et al., 2011; Van Lange, Bekkers, et al., 2007). For example, prosocials tend to be cooperative people who prefer an equal outcome between themselves and other people.
Proself individuals, on the other hand, are competitive/individualistic individuals who prefer
unequal outcomes favouring themselves. My newly developed ICE measure is a reliable and valid measure that captures an important psychological construct underpinning these preferences for different outcomes. By showing that scores on the ICE measure are significantly associated with SVO scores, and that ICE scores are also predictive of allocation behaviour in economic games, this research extends what we know about SVO. Put simply, the current research suggests that prosocials and proselfs appraise settings in which resources are to be divided between themselves and others differently, and that these differential appraisals lead them to anticipate different emotional experiences if they were to divide resources more or less equally.
Between them, the studies reported in this thesis can be seen as establishing a causal chain linking stable individual differences in preferences for outcomes (SVO) with differences in resource allocation behaviour via differential anticipated emotions. The three key constructs that were investigated in this thesis are SVO, anticipated cooperative and competitive emotions and resource allocation behaviour. In the two studies reported in Chapter 3, I measured each of these constructs in a correlational design and demonstrated that SVO influenced allocation behaviour through the proposed mediator, anticipated cooperative and competitive emotions. This is a measurement-of-mediation design, drawing on a statistical analysis procedure originally introduced by Baron and Kenny (1986). It tests whether an independent or predictor variable influences a dependent or outcome variable through a mediating variable. However, Spencer, Zanna and Fong (2005) argued that the Baron and Kenny method, despite its widespread use, is often misapplied. Spencer et al.
argued that in the measurement-of-mediation design, the relations between the predictor and outcome variable mediated by a third variable are often based on correlations. Like any correlational evidence, this leaves open the possibility that the causal relations assumed by the mediation model could be reversed or could be due to shared variance with an external,
unmeasured variable. This point applies in particular to the relation between the mediating variable, which is often intended to capture a psychological process, and the outcome variable, because the mediator is typically measured and not manipulated. To establish that the psychological process captured by the mediator genuinely has a causal impact on the outcome variable, the mediator should be manipulated (Spencer et al., 2005). In combination with evidence from the measurement-of-mediation design, this provides stronger support for the proposed mediation model.
Consistent with the argument of Spencer et al. (2005), in the studies reported in Chapter 5 I manipulated the proposed mediator (anticipated emotions). The findings of these two studies suggest that anticipated emotions indeed play an important role in decision making. When participants were instructed to down-regulate their emotions when contemplating resource allocation decisions, this led to differences in their allocation behaviour. The evidence from this manipulation-of-mediation approach therefore complements the evidence from the measurement-of-mediation approach used in the previous studies and helps to establish that there is a causal chain between the three constructs, such that SVO has its influence on allocation behaviour through its impact on anticipated emotions rather than the other way around.
Previous research has established that SVO is a stable preference (Messick &
McClintock, 1968; Murphy et al., 2011). The current studies are consistent with this past work in showing that SVO is a stable individual difference. Measuring SVO using the slider measure at two different time-points showed substantial test-retest reliability. However, the present research shows that down-regulating or up-regulating anticipated emotions can override the influence of an individual’s SVO on his or her allocation behaviour. Thus, regardless of an individual’s SVO, manipulating anticipated emotions has an impact on allocation behaviour. This further strengthens the evidence for my main research prediction
by showing that anticipated emotions play an important role in resource allocation behaviour by overriding the influence of the supposedly stable trait of SVO.
The current research also provides evidence that SDO is predictive of allocation behaviour, with persons who score highly in SDO being more likely to make unequal resource allocation decisions favouring themselves. These findings are consistent with social dominance theory, which proposes that group-based social hierarchies are formed and maintained through processes of discrimination and prejudice (Pratto et al., 1994). Also consistent with social dominance theory is the finding that the SDO scores of members of the majority and minority groups in Malaysia differed, reflecting the social hierarchical status these groups have in Malaysia. Chinese participants, who are a minority group in that country, scored lower in SDO than did Malay participants, who form the majority group in Malaysia. Although I did not find a significant interaction between the majority/minority group membership of allocators and SDO scores on allocation behaviour, the fact that SDO scores differed significantly between Malays and Chinese and were also predictive of allocation behaviour was consistent with this reasoning. Taken together, the current studies provide some suggestive evidence that belonging to a majority or minority social group plays a role in resource allocation behaviour, but the precise nature of this relation is still an open question. Perhaps future research could pursue the question of why majority and minority members in Malaysia differ in their SDO scores but this does not reflect in their allocation behaviour.
Although there is evidence that some emotions signals, such as laughter and screams, are recognised similarly across different cultures, there are other specific emotions that have different signals in different cultures (Sauter, Eisner, Ekman, & Scott, 2010). Furthermore, Russell (1994) proposed that although the valence and arousal dimensions of emotion are universally recognised, specific emotions are more culturally variable. Evidence of the
cultural variability of emotions raises the possibility that there might also be cultural variability in the influence of anticipated emotions. This issue was addressed in the current studies by recruiting participants from two different cultural backgrounds (the UK and Malaysia). There were similar patterns of results across these two cultures when investigating whether anticipated cooperative and competitive emotions mediated the relation between SVO and allocation behaviour. In the two different populations studied, the national languages differed, with the UK samples (mainly students of British nationality) speaking English and the Malaysian samples speaking Malay. This meant that the definitions of each anticipated emotions (in the context of resource allocation decisions) had to be translated carefully from English to Malay for the Malaysian participants. Despite the evidence that emotions have different connotations in different cultures and languages (Sauter et al., 2010;
Wierzbicka, 1994), the current studies show a robust pattern of mediation of the relation between SVO and allocation behaviour by anticipated emotions. This suggests that the anticipated emotions that are assessed by the ICE measure have similar meanings and have a similar influence across the two cultures I investigated. These findings are consistent with research that found universal recognition of pride (Tracy & Robins, 2008), guilt (Furukawa, Tangney, & Higashibara, 2012) and regret (Gilovich, Wang, Regan, & Nishina, 2003).