Brief Communication | Published:

Warm glow is associated with low- but not high-cost sustainable behaviour

Nature Sustainabilityvolume 1pages2830 (2018) | Download Citation



Why do people contribute to important societal causes, such as sustainability? This study hypothesized that people are motivated to help because they anticipate a sense of warm glow from acting green. Although results reveal that ‘feel-good’ affect mostly drives low- rather than high-cost behaviour changes, harnessing people’s intrinsic motivation to help the environment may be an underleveraged mechanism for promoting sustainability.

Although the study of why people contribute—or fail to contribute—to important societal causes has captured the attention of behavioural science for many decades1, the motivational foundations of cooperative behaviour seem especially relevant in the face of what is perhaps the biggest social dilemma of our time: ensuring a sustainable future for our planet2,3. While the psychological factors that motivate sustainable behaviour have shown to be diverse4, one important observation is that humans are often intrinsically motivated to contribute to societal causes because people derive a sense of ‘warm glow’ from helping others5.

Yet, the ‘half-life’ of empathy-driven prosocial behaviour—whether it concerns raising support for debilitating diseases such as amyotrophic lateral sclerosis, conserving energy or humanitarian responses to conflict—tends to be short-lived6,7,8. This is problematic when it comes to issues that require long-term support and cooperation from people, such as sustainability. Accordingly, solutions that leverage people’s ‘intrinsic’ motivation to help the environment may sustain behaviour change longer than policy incentives that are contingent on extrinsic rewards6,7,8. Nonetheless, an important prerequisite of this assumption is that people actually experience or anticipate positive ‘internal’ rewards from acting green. This is a tenuous assumption for several reasons. First, compared to helping other humans in need, sustainability often includes non-human nature and so it is less clear whether people also derive a sense of warm glow from contributing to large-scale and relatively depersonalized collective action problems. To this extent, emerging research suggests that helping the environment does in fact make people feel good about themselves9,10,11. This is also evidenced by the fact that extrinsic incentives, such as financial rewards, often backfire12, presumably because they dilute the purity of the prosocial act13, which generally makes people feel less positive about their contribution to the environment14.

Second, prior studies have mostly measured ‘experienced’ emotions to help the environment retrospectively, often for one particular action at a single point in time and typically using ‘good intentions’ as a substitute for behaviour. It is therefore unclear to what extent people act sustainably because they anticipate that engaging in morally desirable behaviours will make them feel good about themselves. In other words, to what extent does a sense of anticipated warm glow from acting sustainably actually predict future green behaviour, especially when changing behaviour may be costly? The current study seeks to answer this novel question by drawing on a national panel study. Specifically, this study takes a prospective view and evaluates to what extent the anticipated ‘feel-good’ from helping to save the planet is predictive of a wide range of low- and high-cost sustainable behaviours over a four-week period.

Data are analysed from a UK national survey, which was administered (longitudinally) in two waves. In wave 1, participants were asked two questions to estimate the degree of positive affect (or warm glow) they would derive from acting sustainably. Green intentions were also measured by asking participants to what extent they intend to adopt a total of 21 green behaviours in the next four weeks (for example, buying locally grown and produced foods). Four weeks later, respondents were re-contacted and asked to report on the same behavioural items. The main findings are reported in Figs. 1,2.

Fig. 1: The influence of anticipated warm glow on green behaviour.
Fig. 1

a,b, The relationship between affect and behaviour without (a) and with (b) green intention. Coefficients are standardized and adjusted for covariates; gender, age, education, and income. Bootstrapped 95% confidence intervals are provided in parentheses. T 2 − T 1 = 4 weeks. To address panel attrition, a full information maximum likelihood (FIML) procedure was used to estimate the mediation models. *P < 0.05, **P< 0.01, ***P < 0.001, N = 808.

Fig. 2: The influence of anticipated warm glow on low-cost and high-cost green behaviour.
Fig. 2

 ad, The relationship between affect and behaviour without (a,b) and with (c,d) green intention. Coefficients are standardized and adjusted for covariates; gender, age, education, and income. Bootstrapped 95% confidence intervals are provided in parentheses. T 2 − T 1 = 4 weeks. To address panel attrition, an FIML procedure was used to estimate the mediation models. *P < 0.05, **P < 0.01, ***P < 0.001, N = 808.

Results show that the anticipated feel-good affect from helping the environment in Time 1 (T 1) directly predicts a wide range of self-reported pro-environmental behaviours (PEBs) four weeks later (Fig. 1a). Because intentions to engage in a specific behaviour are often the most proximal determinant of behaviour15, the influence of positive affect on PEB is also examined while controlling for green intentions. Although mediation analysis reveals that about 75% of the total effect is mediated by intentions, positive affect still exerts a significant direct impact on PEB (Fig. 1b). Positive affect also explains about 30% of the variance in ‘green intentions’.

To examine whether the same pattern holds for more costly PEBs, the model was estimated separately for low- and high-cost sustainable behaviours (Fig. 2). The high-cost index included 10 items (for example, insulate your home, purchase green energy, buy carbon offsets and so on). Interestingly, compared with low-cost green behaviours (Fig. 2a), anticipated warm glow from helping the environment at T 1 had a much smaller direct effect on more costly green behaviours four weeks later (Fig. 2b). Moreover, while still playing a (small) role in green intention formation, anticipated positive affect no longer exerts an independent influence on high-cost PEB (Fig. 2d).

Lastly, given political polarization on environmental issues16, the role of ideology was examined. Although liberals anticipated significantly more positive affect from acting green than conservatives (M = 4.99, s.e.m. = 0.07 versus M = 4.35, s.e.m. = 0.13, t(654) = 4.73, p < 0.001), the correlation between positive affect and behaviour did not differ by ideology (r p = 0.39 versus 0.44, P = 0.56), indicating it is likely that anticipatory warm glow fulfills the same psychological function for conservatives and liberals alike.

Drawing on national panel data, this study demonstrates that the anticipated warm glow from contributing to a sustainable future actually predicts a range of green behaviours four weeks later. These findings evidence that, at least in part, people are intrinsically motivated to (intend) to help the environment because they expect that engaging in morally desirable behaviours will generate a feel-good warm glow. This study constructively builds on prior work7,9,10, but crucially, highlights that anticipated feel-good affect is primarily predictive of ‘low-cost’ actions (for example, switching off lights) and not of more-involved green behaviours (for example, buying green energy). This is consistent with the finding that when helping behaviour is costly and at odds with other (self-serving) motives, more cognitive mediation tends to occur17.

Although this study does not provide causal evidence to suggest that intrinsic motivation is a more durable driver of green behaviour than extrinsic incentives, the fact that people actually anticipate that behaving sustainably will make them feel good suggests that behavioural policies that appeal to intrinsically valued motives could help engage people with important societal causes. Especially because extrinsic rewards, such as economic incentives, are often costly, may backfire, and are difficult to implement and maintain7,18. Moreover, the results presented here may be conservative, as other work suggests that people frequently underestimate how much they enjoy engaging with the environment19. In short, the norm of pure ‘self-interest’ often portrays a misleading picture of human nature. Nonetheless, it is of course possible that people can derive positive emotions from ‘extrinsic’ incentives too (for example, feeling good about saving money).

These results invite future experimental work to assess whether positive emotions can be leveraged to sustain more long-term behaviour change, and whether this in turn can increase the degree of anticipated warm glow and, as such, result in a positive self-reinforcing ‘feedback’ loop. More generally, this line of inquiry is important, as it can help inform whether behavioural policies that harness people’s intrinsic motivation to forge a more sustainable future are a viable low-cost alternative to help change human behaviour and decision-making in the long-term.


Data collection

The data set is based on a national quota sample (N = 808) of the population of the United Kingdom (England, Scotland, Wales and Northern Ireland) obtained from Survey Sampling International. Quotas were based on gender, age and region. Multi-stage randomization was used to select participants from a large mixed panel of people who were willing to participate in web-based research for a small (non-monetary) reward. The survey was administered (longitudinally) in two waves. In the first wave N = 808 responses were collected, and N = 501 (or 62%) of panel members responded in the second wave, four weeks later. More details on the sample and methodology are available in the Supplementary Information.


Participants were asked to estimate the degree of positive affect (or warm glow) they derive from acting sustainably by evaluating the following statements: “I would feel good about myself if I decided to take personal action to help reduce climate change” and “I would feel positive if I did my bit to help tackle climate change” (1 = very unlikely, 7 = very likely). Both items were combined and averaged to form a reliable measure (Pearson's correlation r = 0.83, Cronbach's α = 0.90).

Green intentions were measured by asking participants to what extent they intend to adopt each of the following 21 behaviours in the next four weeks (example items: “reduce my driving by using more public transportation”, “purchase home energy from a green source”, “buy locally grown and produced foods”). Response format was a 7-point scale (very unlikely–very likely) and intentions were averaged (mean M = 4.44, s.d. = 1.37) to form a reliable index (α = 0.92). Four weeks later, respondents were re-contacted and presented with the same behavioural items and asked to report whether and how often they performed (M = 3.90, s.d. = 1.35) each of the behaviours (never–very frequently, α = 0.91). The sub-scales for low- (M = 4.30, s.d. = 1.40) and high-cost (M = 3.23, s.d. = 1.59) green behaviour also indicated good reliability (α = 0.87 and 0.90, respectively). The ideology measure contrasted support for the Conservative and UK Independence parties (24%) versus the Labour Party and Liberal Democrats (36%)—the remainder declined to answer. The full list of measures is available in the Supplementary Information.

Data analysis

The data were analysed with a structural equation modelling approach using STATA v. 14.2 ( To address panel attrition (38%), a full information maximum likelihood (FIML) procedure was used to estimate the mediation models. Although the results are robust to estimation method, FIML is recommended because it uses the full sample variance-covariance matrix to estimate the missing data20. Additional discussion of the statistical approach and analyses are available in the Supplementary Information.

Data availability

The data that support the findings of this study are available from the author upon request.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Batson, C. D. & Powell, A. A. in The Handbook of Psychology (ed. Weiner, I. B.) 463–479 (Wiley, 2003).

  2. 2.

    Oskamp, S. Am. Psychol. 55, 496–508 (2000).

  3. 3.

    van der Linden, S., Maibach, E. & Leiserowitz, A. Perspect. Psychol. Sci. 10, 758–763 (2015).

  4. 4.

    Kollmuss, A. & Agyeman, J. Environ. Educ. Res. 8, 239–260 (2002).

  5. 5.

    Andreoni, J. Econ. J. 100, 464–477 (1990).

  6. 6.

    van der Linden, S. Nat. Hum. Behav 1, 0041 (2017).

  7. 7.

    van der Linden, S. Nat. Clim. Change 5, 612–613 (2015).

  8. 8.

    Slovic, P., Västfjäll, D., Erlandsson, A. & Gregory, R. Proc. Natl Acad. Sci. USA 114, 640–644 (2017).

  9. 9.

    Bolderdijk, J. W., Steg, L., Geller, E. S., Lehman, P. K. & Postmes, T. Nat. Clim. Change 3, 413–416 (2013).

  10. 10.

    Taufik, D., Bolderdijk, J. W. & Steg, L. Nat. Clim. Change 5, 37–40 (2015).

  11. 11.

    Evans, L. et al. Nat. Clim. Change 3, 122–125 (2013).

  12. 12.

    Asensio, O. I. & Delmas, M. A. Proc. Natl Acad. Sci. USA 112, E510–E515 (2015).

  13. 13.

    Ariely, D., Bracha, A. & Meier, S. Am. Econ. Rev. 99, 544–555 (2009).

  14. 14.

    Schwartz, D., Bruine de Bruin, W., Fischhoff, B. & Lave, L. J. Exp. Psychol. Appl. 21, 158–166 (2015).

  15. 15.

    Ajzen, I. Organ. Behav. Hum. Decis. Process 50, 179–211 (1991).

  16. 16.

    McCright, A. M., Xiao, C. & Dunlap, R. E. Soc. Sci. Res. 48, 251–260 (2014).

  17. 17.

    Moll, J. et al. Proc. Natl Acad. Sci. USA 103, 15623–15628 (2006).

  18. 18.

    Deci, E. L., Koestner, R. & Ryan, R. M. Psychol. Bull. 125, 627–668 (1999).

  19. 19.

    Nisbet, E. K. & Zelenski, J. M. Psychol. Sci 22, 1101–1106 (2011).

  20. 20.

    Enders, C. K. & Bandalos, D. L. Struct. Equ. Modeling 8, 430–457 (2001).

Download references


I would like to thank the UK’s Economic and Social Research Council (ESRC), the Grantham Foundation for the Protection of the Environment, the Andlinger Center for Energy and the Environment and the Yale Program on Climate Change Communication for their support.

Author information


  1. Department of Psychology, Downing Site, University of Cambridge, Cambridge, UK

    • Sander van der Linden


  1. Search for Sander van der Linden in:


S.v.d.L. is the sole author of this article and is fully responsible for its content.

Competing interests

The author declares no competing financial interests.

Corresponding author

Correspondence to Sander van der Linden.

Electronic supplementary material

  1. Supplementary Information

    Supplementary Methods, Supplementary Tables 1–3, Supplementary Fig. 1 and Supplementary References

About this article

Publication history





Further reading