Updated: Nov 22, 2020
Rationalists assume that humans only value their own outcome given a social decision-making scenario. Then, do humans not care about the consequences of their actions on others? It turns out, they do care. But, it is a qualified affirmation.
Hu et al. (2017), in their study titled ‘Social value orientation modulates the processing of outcome evaluation involving others’, have conducted an Event-Related Potential (ERP) study to measure outcome evaluation, for both self and other, under conditions of social decision-making. They devised a simple gambling task where the participant (also called the decision-maker) had to choose one out of two cards. Every choice led to an outcome, one for the decision-maker him/herself and one for the recipient, the other (a confederate). For example, the outcome may be a gain of 20 for self and a loss of 20 for the other. So, outcomes varied in terms of valence but not in terms of magnitude. Hence, the amount of gain or loss was always restricted to 20. The experiment involved four conditions: 1) self-gain (+20) and other-gain (+20), 2) self-loss (-20) and other-loss (-20), 3) self-gain (+20) and other-loss (-20), and 4) self-loss (-20) and other-gain (+20). The experiment was constructed in order to elicit brain response to an outcome, involving both self and other, which was presented to the participant (or decision-maker). In addition, the authors were also interested to see whether individuals who were more individualist and competitive in nature viewed outcome distributions in contrast to individuals who were cooperative and valued equality of distribution. Therefore, they categorised these two types of individuals as pro-selfs and pro-socials, respectively. This categorisation was carried out with the help of the Triple-Dominance Scale.
In order to understand the different ERP measures the following few terminologies and definitions will be of use. The Event-Related Potential (ERP) approach is a way of understanding the neural correlates of brain processes. In other words, ERPs are brain waves measured using electrodes. These waves are brain potentials measured in microvolts (µV), along the time axis. Hence, they are temporal in nature. Also, ERPs are generated as a response by the brain to a stimulus. Since ERPs are related to an event, which can also be a stimulus, they are called Event-Related Potentials.
Therefore, it is clear that a given stimulus will generate a certain waveform. Let’s take an example illustrated in figure 1. The graph shows four ERP waveforms. These are waveforms generated in the brain upon presentation of a stimulus. For the purposes of this study, the stimulus is the outcome presentation screen, which includes the outcome for both the self and the other.
Now, let us try to understand the specifics of this graph so that a few terminologies are clear. The x-axis represents time in milliseconds (ms) and the y-axis represents potential in microvolts (µV). Now, it is easy to understand that an ERP waveform is temporal in nature. On the y-axis, it is important to note that negative is plotted upwards and positive is plotted downwards. On the x-axis, zero (0) is the point in time at which the outcome screen is presented. Additionally, this graph concentrates on the total time window. But, the total time window can be further broken up into several smaller time windows.
For the purpose of understanding a cognitive process, in figure 2 (panel A) and figure 3 (panel A), we will focus only on that portion of the waveform which lies within the shaded time window. The waveform within a time window has a certain amplitude. Amplitude of a wave is the difference in potential between the x-axis and the peak of the wave. It can also be said that amplitude is a measure of how prominent a waveform is. All statistical analyses carried out for waveforms in this study is with reference to amplitude.
Every ERP component is measured by a waveform. A waveform can be called negative-going if it peaks in the negative direction, along the y-axis, but not necessarily peaking in the negative section of the graph. Also, the more negative-going a waveform is the stronger or more prominent the corresponding ERP component is. The phrases ‘negative-going’ and ‘negativity’ will be used interchangeably in this post. The same definition holds for positive-going as well.
In order to understand the underlying cognitive processes involved during outcome evaluation the authors analysed three temporally separated processes as identified by three different ERP components. The ERP components that the authors focused on were the Feedback-Related Negativity (FRN), the P3 and the Late Positive Component (LPC).
The first of these components, the FRN, measures losses. In other words, the greater the monetary loss the greater will be the negativity of FRN. The FRN waveforms for both groups have been presented in figure 2 (panel A). Clearly, the FRN waveform is negative-going for the selected time window, indicated by a vertical grey bar.
In order to understand the differences observed in the four conditions, follow the bar graphs in figure 2 (panel B). It was observed that FRN was more negative-going for self-loss than for self-gain, for both pro-selfs and pro-socials. But, it was observed that only pro-socials distinguished between other-gain and other-loss. In other words, only for pro-socials, the FRN amplitude was more negative-going for other-loss than for other-gain. The pro-selfs did not draw this distinction. The asterisks in the graph indicate statistically significant difference between conditions.
According to Taylor et al. (2006), the FRN component represents a quick and crude assessment of cognitive processes at a very early stage of outcome evaluation. The above observation supports previous findings that pro-socials pursue outcomes that is beneficial to both self and other. In other words, this is proof of individual reward-maximisation for pro-selfs and collective reward-maximisation for pro-socials.
Next, the P3 component measures feedback valence as well as reward size. In other words, P3 amplitude is larger (or more positive-going) for positive feedback than for negative feedback and also for a larger reward than for a smaller one. Also, according to previous studies, the P3 reaches its peak within the time window of 300-600 ms post outcome presentation. Therefore, P3 represents an intermediate evaluation of outcome distribution. In figure 3 (panel A), P3 is represented by the dark grey time window. The time window extends from 350-450 ms. Clearly, P3 is a positive-going waveform.
The authors found that P3 amplitude was greater for self-gain than for self-loss, for both pro-selfs and pro-socials. Moreover, pro-selfs did not draw any further distinction between gain and loss for the other individual. But, pro-socials were able to distinguish between other-gain and other-loss, but only during self-gain. In other words, pro-socials behaved like pro-selfs under the self-loss condition. These results are illustrated in figure 3 (panel B).
The pro-socials in this intermediate evaluation stage evaluated outcomes differently from what they do in the first or early stage. At this stage, they did not draw any distinction about the other’s outcome when they were undergoing losses themselves, unlike in the first stage. This insensitivity to other’s outcome by the pro-socials can be referred to as a positivity bias.
According to Buunk and Gibbons (2007), individuals receiving a negative self-outcome behave in order to protect one’s positive self-concept from harm. In order to do so, people tend to avoid attending to other’s outcome as that might lead to their self-concept getting hurt. Also, the observed positivity bias can play a useful role in maintaining a high self-esteem for pro-socials in social interactions under the condition of self-loss, as they are prone to being influenced by others.
Finally, the Late Positive Component (LPC) is a measure of emotional arousal at a later stage of outcome evaluation. According to previous studies, the LPC amplitude is larger (or more positive-going) for higher arousal levels. Figure 3 (panel A) illustrates the LPC waveform, shaded with a light grey vertical rectangle. The time window for this waveform is 500-700 ms. Although the waves are not very prominent, they are positive-going, just like the P3 waveform.
The results obtained at this stage were similar to those found during the early or first stage of outcome evaluation. The authors found that the LPC amplitude was larger for self-gain than for self-loss, for both groups. Moreover, only pro-socials drew a distinction between other-gain and other-loss. In other words, for pro-socials, the LPC was larger for other-gain than for other-loss irrespective of self-outcome. These findings have been illustrated in figure 3 (panel C).
The results indicate that pro-selfs experienced higher arousal level for self-gain than for self-loss. On the other hand, pro-socials experienced higher arousal level not only for self-gain compared to self-loss, but also for other-gain compared to other-loss. It can be clearly noted that pro-socials drew a distinction about other-outcome even during the self-loss condition. Therefore, we observe a change in outcome evaluation for pro-socials going from the intermediate stage to the later stage.
Previous studies have associated the LPC with the Theory of Mind. According to Van Lange (2000), comparatively pro-socials have stronger Theory of Mind ability and altruistic motivations for others. Hence, these altruistic motivations lead them to feel greater arousal for other-gain than for other-loss. Moreover, the change in outcome evaluation observed for pro-socials, going from the intermediate stage to the later stage, can be explained by their change of focus from self to the other due to their strong ability of Theory of Mind and altruistic motivations.
There were a few limitations to the study. The experimental situation created was too pure to capture real-life social factors that can regulate outcome evaluation. It can be of interest to find how people respond in different social contexts, such as a bargaining situation (Alexopoulos et al., 2013). In addition to a change in context, one crucial difference in the study by Alexopoulos et al. (2013) is that a choice or decision led to an outcome with certainty, whereas Hu et al. (2017) investigated a gambling scenario where a choice led to certain outcomes probabilistically. Therefore, varying agency control can lead to changes in valuation for other-outcome.
The other major drawback of this study was that all participants were from China, which is considered to be a collectivist culture. It is possible that individuals from a Western culture may respond in a different manner while evaluating outcomes. Another dimension of comparison could be across gender. An interesting question to study would be to investigate whether women care for other-outcome more than men. Therefore, it may be worthwhile to investigate the combined effect of social orientation and gender on outcome evaluation.
In conclusion, to answer the central question of interest, not all individuals value the consequences of their actions on others. According to the findings of this study, it can be inferred that only pro-socials care about such consequences on others. Even then, pro-socials do not all throughout give weight to the consequences on others. If cognitive processes during outcome evaluation can be divided into three temporally separated segments, it was observed that pro-socials valued other-outcome in all three stages except for the intermediate stage, where they stopped processing the other-outcome when undergoing self-loss. Therefore, processing other-outcome is not only modulated by our social orientation, but also by the outcome we experience. Does such a conclusion makes sense in a collectivistic socio-cultural environment? Moreover, would the findings take a turn towards a more rationalist perspective in a more individualistic society? What do you think?
This post has been written after a Journal Club presentation on the same article. If you are interested in knowing the details pertaining to the method and discussion sections, kindly refer to the original article by Hu et al. (2017). If you have suggestions related to this post, please leave them as comments. Also, feel free to comment on the findings of this study.
(Disclaimer: Any views or opinions represented in this blog article are personal and belong solely to the author, and do not represent those of people, institutions or organizations that the author may be associated with in professional or personal capacity, unless explicitly stated.)
Alexopoulos, J., Pfabigan, D. M., Göschl, F., Bauer, H., & Fischmeister, F. P. S. (2013). Agency matters! Social preferences in the three-person ultimatum game. Frontiers in human neuroscience, 7, 312. https://doi.org/10.3389/fnhum.2013.00312
Buunk, A. P., & Gibbons, F. X. (2007). Social comparison: The end of a theory and the emergence of a field. Organizational Behavior and Human Decision Processes, 102(1), 3-21. https://doi.org/10.1016/j.obhdp.2006.09.007
Hu, X., Xu, Z., & Mai, X. (2017). Social value orientation modulates the processing of outcome evaluation involving others. Social Cognitive and Affective Neuroscience, 12(11), 1730-1739. https://doi.org/10.1093/scan/nsx102
Leicht, G., Troschütz, S., Andreou, C., Karamatskos, E., Ertl, M., Naber, D., & Mulert, C. (2013). Relationship between oscillatory neuronal activity during reward processing and trait impulsivity and sensation seeking. PLoS One, 8(12), e83414. https://doi.org/10.1371/journal.pone.0083414
Taylor, S. F., Martis, B., Fitzgerald, K. D., Welsh, R. C., Abelson, J. L., Liberzon, I., ... & Gehring, W. J. (2006). Medial frontal cortex activity and loss-related responses to errors. Journal of Neuroscience, 26(15), 4063-4070. https://doi.org/10.1523/JNEUROSCI.4709-05.2006
Van Lange, P. A. (2000). Beyond self-interest: A set of propositions relevant to interpersonal orientations. European review of social psychology, 11(1), 297-331. https://doi.org/10.1080/14792772043000068