Friday, 1 November 2013

Does Elevation Lead to Altruistic Behaviour? Commentary on Schnall, Roper and Fessler (2010).

Today, I presented an article at the Moral Psychology Journal Club at UniMelb, Elevation Leads to Altruistic Behavior by Simone Schnall (University Cambridge), Jean Roper (University of Plymouth), and Daniel Fessler (UCLA). I chose this article because I'm really interested in positive psychology, and I was eager to learn more about the elevation and prosocial behaviour. Some really interesting discussion (i.e. critique) ensued so I felt motivated to share some thoughts on this paper.

First, the paper.

What is elevation?

Elevation, coined by Haidt (2003), is "a positive emotion experienced upon witnessing another person perform a virtuous act, principally one that improves the welfare of other people" (Schnall, Roper, & Fessler, 2010, p. 315). Typically, people experiencing elevation will report that they feel inspired and uplifted, and motivated to perform a similarly prosocial act themselves.

Does elevation lead to altruistic behaviour?

Previous research had shown limited and indirect evidence for this hypothesis, for example through self-reported altruistic behaviour (Landis et al., 2009), more nurturant behaviours towards their babies in nursing mothers (Silvers & Haidt, 2008), and increased donations to a Black charity from Whites who are high in social dominance orientation, a construct linked to anti-Black racism (Freeman, Aquino, & McFerran, 2009). 

However, Schnall, Roper, and Fessler (2010) argued that self-report is subject to impression management considerations, that helping one's own child isn't true altruism, that it was still unknown whether elevation motivates helping outside of the context of group-based prejudice, and that none of these experiments had shown whether elevation motivates helping above and beyond the effects of general positive affect. Schnall, Roper, and Fessler (2010) therefore tried to establish whether elevation actually and unambiguously produces altruistic behaviour.

Experiment 1


Participants. 59 women, aged 18-26 years from the University of Plymouth.

Materials. For the elevation condition, they used a 7-min clip from "The Oprah Winfrey Show" where musicians thank the teachers who had mentored them as underprivileged youths. For the control condition, they used 7 mins of a nature documentary.

They used six self-report measures of feelings and cognitive appraisals associated with elevation: how much participants felt "moved," "uplifted," "optimistic about humanity," "warm feeling in chest," "want to help others," and "want to become a better person," on a 9-point scale (1 = didn't feel at all, 9 = felt very strongly). Using the same scale, they also asked participants to rate how happy they felt, to assess the effect of condition on general positive affect.

Procedure. Participants watched either the elevation-inducing clip or the control clip. They then spent 5 mins writing a short essay recalling as much as they could about it. The experimenter then left the room, ostensibly to photocopy another form. Before she left, she paid the participant and asked them to complete a payment receipt containing the dependent measure, a check box asking if the participant was willing to participate in an additional, unpaid study. On her return, the experimenter gave the participant the form with the seven self-report measures to complete.


Manipulation check. The mood induction of elevation was successful: Participants in the elevation condition reported higher ratings on all six items indicative of elevation ("moved," "uplifted," "optimistic about humanity," "warm feeling in chest," "want to help others," and "want to become a better person"), compared with the control condition. The groups didn't differ in their reported happiness.

Intention to volunteer. As predicted, more participants in the elevation condition volunteered for the unpaid study.


Three limitations of this experiment were:
  1. Their single measure of happiness may not have been enough to rule out the possibility that their results were driven by general mood differences, because elevation is positively-valenced.
  2. Since the dependent measure was yes/no, they couldn't assess dose-dependent effects of emotions
  3. The dependent measure only measured a commitment to help, not actual helping behaviour
Hence, enter Experiment 2...

Experiment 2


Participants. After 4 participants were excluded, 32 female participants remained.

Procedure & materials. To account for the positive valence of the elevation condition, they introduced a new condition, the mirth condition, induced using a comedy clip. So they had three conditions this time: elevation, mirth and control. Participants watched one of these clips, then completed the self-report measures from Experiment 1, which also included a rating of how amused they felt. The experimenter then faked some technical difficulties that made it "impossible to continue the experiment", so told the participant she was free to leave, but would receive the full hour's course credit. Then as an "afterthought", the experimenter asked the participants whether she would be willing to complete another questionnaire, emphasising that the questionnaire was rather boring, and that the participant was under no obligation and was free to stop whenever she wanted, but that completing any number of the items would greatly assist the experimenter. Participants who agreed to help were given 85 elementary math problems and secretly timed on how many minutes they spent "helping" the experimenter.


Manipulation check. Elevation and amusement were induced in their respective conditions, whilst reported happiness didn't differ between the three conditions.

Helping behaviour. Participants in the elevation condition spent roughly twice as much time on the questionnaire (~ 40 mins)  as participants in the control or mirth conditions; the latter conditions didn't differ. The dependent variable, minutes spent on the questionnaire, was significantly positively correlated with 5/6 of the elevation variables, but not significantly correlated with amusement or happiness.

Key conclusions
  1. The two experiments provided "convincing evidence" that elevation leads to increased altruistic behaviour.
  2. Elevation is distinct from mere positive mood.
  3. Elevation inspired helping in spirit, not in kind (i.e. imitation), as their helping behaviours (volunteering for a study/completing a questionnaire) bore no resemblance to that in the stimulus (mentoring underprivileged youths).
  4. By eliciting elevation, even brief exposure to other individuals' prosocial behaviour increases altruism. This is a potential way to increase the general level of prosociality in society.

Now, some questions.

When I first read the paper, I thought it was great! But I totally expected that the more experienced members of the Journal Club would tear it apart somehow, and of course they did, and I am truly grateful for the chance to learn from them and develop my skills in critically reading and evaluating psychological research. So these were our key concerns:

Gender. They only had female participants in this study, because of concerns that men might exhibit more helping behaviour because the experimenter was female. However, we thought that this wasn't really justified, unless you expected that men would be so eager to help that there would be a ceiling effect. In only testing females, they've excluded half the human population in being able to generalise these findings.

Sample size. In Experiment 2, there were only 32 participants, across 3 conditions! That means about 10 people in each condition. Surely at least 20 in each condition would be considered a minimal requirement here.

Mediation analyses (or lack thereof). This was the most important issue. The purpose of these experiments, as suggested by the title, was to show that elevation leads to (i.e. causes) altruistic behaviour, but they didn't really show that at all. Rather, they showed that the elevation condition induced elevation, and that the elevation condition increased altruistic behaviour, but they didn't do (or report) any basic regression analyses to show that elevation was actually mediating the relationship between the condition and the increased altruistic behaviour. In other words, they didn't rule out the possibility that elevation may have been epiphenomenal, and that something else may have mediated the relationship between witnessing the prosocial act and helping behaviours. For example, the Oprah Winfrey clip could have simply established a helping norm, basically priming helping behaviours. (In fact, we were wondering if it would be possible to ever get past this confound - is there a way to induce elevation without witnessing prosocial behaviour, or is that by definition impossible?)

On a final note regarding the mediation analyses, it looks like Thomson and Siegel (2013) more recently established elevation as a mediator between thinking about a time when they witnessed someone performing a moral act benefiting someone else, and donation behaviours, in samples with both genders. Take this with a grain of salt though, or check it out yourself, because I haven't had a chance to read the paper in-depth yet.

Overall, we weren't convinced that the experiments established that elevation leads to altruistic behaviour. Perhaps a more accurate title would have been, "Witnessing prosocial behaviour leads to elevation and altruistic behaviour in females". I would be extremely interested to see what happens if we (1) test both males and females, (2) have an adequate sample size, and (3) ran the mediation analyses. Would it then show that elevation leads to altruistic behaviour? Given the current crisis of confidence in social psychology, it's more important than ever to ensure that studies are as rigorous and unconfounded as possible.

Having said all this, in line with a positive psychology perspective, I want to recognise the strengths in this paper. The math questionnaire dependent measure was quite clever, as it was a way to quantitatively measure extents of helping behaviour. I also liked how they introduced the mirth condition to try and match the affective valence (although I suspect that elevation may still be higher in valence and intensity than mirth). And in the end, at least they showed that witnessing prosocial behaviour inspires prosocial behaviour. If we want to increase prosocial behaviour, that is, to encourage a society where people more regularly help one another out, then that is still an important, relevant and practically applicable finding in terms of publicising morally praiseworthy events, regardless of whether elevation, or something else, is mediating that relationship.


Freeman, D., Aquino, K., & McFerran, B. (2009). Overcoming beneficiary race as an impediment to charitable donation: Social dominance orientation, the experience of moral elevation, and donation behavior. Personality and Social Psychology Bulletin, 35, 72-84. doi: 10.1177/0146167208325415

Haidt, J. (2003). Elevation and the positive psychology of morality. In C.L.M. Keyes & J. Haidt (Eds.), Flourishing: Positive psychology and the life well-lived (pp. 275–289). Washington, DC: American Psychological Association.

Landis, S.K., Sherman, M.F., Piedmont, R.L., Kirkhart, M.W., Rapp, E.M., & Bike, D.H. (2009). The relation between elevation and self-reported prosocial behavior: Incremental validity over the five-factor model of personality. Journal of Positive Psychology4, 71–84. doi: 10.1080/17439760802399208

Schnall, S., Roper, J. & Fessler, D.M.T. (2010). Elevation leads to altruistic behavior. Psychological Science, 21(3), 315–320. doi: 10.1177/0956797609359882

Silvers, J.A., & Haidt, J. (2008). Moral elevation can induce nursing. Emotion, 8, 291–295. doi: 10.1037/1528-3542.8.2.291

Thomson, A.L., & Siegel, J.T. (2013). A moral act, elevation, and prosocial behavior: Moderators of morality. The Journal of Positive Psychology, 8(1), 50-64, doi: 10.1080/17439760.2012.754926

No comments:

Post a Comment