Morality in Everyday Life Applications

Corrected 11 May 2015; see full text. RESEARCH | REPORTS

Fig. 4. Magneto-optical response 29. K. Vahaplar et al., Phys. Rev. B 85, 104402 (2012). 30. C. H. Back et al., Science 285, 864–867 (1999). in various applied magnetic field of a 15-nm FePtAgC granular film ACKNOWLEDGMENTS sample starting with an initially We thank R. Tolley and M. Gottwald for help on sample fabrication demagnetized sample. Shown are line and M. Menarini and V. Lomakin for helpful discussions. This work s+ was supported by the Agence Nationale de la Recherche (ANR), scans for circularly polarized light in “ ” s – ANR-10-BLANC-1005 Friends, and work at University of California the left column and circularly San Diego was supported by the Office of Naval Research (ONR) polarized light in the right column. The Multidisciplinary University Initiative (MURI) program and a grant laser power was 677 nW. The magnitude from the Advanced Storage Technology Consortium. It was also “ of the external magnetic field is given supported by the Partner University Fund Novel Magnetic Materials for Spin Torque Physics” as well as the European Project in the figures, and the orientation of the (OP2M FP7-IOF-2011-298060) and the Region Lorraine. Work at + field supports s polarization and the National Institute for Materials Science was supported by opposes the s – polarization. Advanced Storage Technology Consortium. Author contributions include the following: S.M., M.A., M.C., Y.S., and E.E.F. designed and coordinated the project; C.-H.L. grew, characterized, and optimized the thin films samples, while B.S.D.Ch.S.V., Y.K.T., and K.H. developed and grew the FePt-based granular media. C.-H.L., S.M., and Y.K.T. operated the Kerr microscope and the pump laser experiment. S.M. and E.E.F. coordinated work on the paper with contributions from C.-H.L, M.H., M.C., and G.M., with regular discussions with all authors. thesamplewilldemagnetizeduringcooling.For 25. S. Alebrand, A. Hassdenteufel, D. Steil, M. Cinchetti, SUPPLEMENTARY MATERIALS M. Aeschlimann, Phys. Rev. B 85, 092401 (2012). perpendicular magnetized films, there are strong www.sciencemag.org/content/345/6202/1337/suppl/DC1 26. A. Moser et al., J. Phys. D Appl. Phys. 35, R157–R167 dipolar fields within the film that support do- Materials and Methods (2002). Supplementary Text main formation. The dipolar energy gain for do- 27. L. Zhang, Y. K. Takahashi, K. Hono, B. C. Stipe, J. Y. Juang, Figs. S1 to S7 main formation is strongly suppressed in the M. Grobis, L1 -ordered FePtAg-C granular thin film for 0 References thermally assisted magnetic recording media. J. Appl. Phys. ultrathin film limit and explains the observation on May 18, 2015 of AO-HDS only in the thin-film limit (Figs. 1 and 109, 07B703 (2011). 17 March 2014; accepted 4 August 2014 28. H. J. Richter, A. Lyberatos, U. Nowak, R. F. L. Evans, Published online 21 August 2014; 2). Domain formation is also suppressed for low R. W. Chantrell, J. Appl. Phys. 111, 033909 (2012). 10.1126/science.1253493 magnetization materials, consistent with AO-HDS measurements of ferrimagnetic materials (11). The present results on ferromagnetic materials demonstrate a new and technologically im- SOCIAL PSYCHOLOGY portant class of materials showing AO-HDS and opens new directions in integrated magnetic- optical memory, data storage, and processing Morality in everyday life applications. This study further offers progress www.sciencemag.org toward a better understanding of the interac- Wilhelm Hofmann,1* Daniel C. Wisneski,2 Mark J. Brandt,3 Linda J. Skitka2 tion between pulsed polarized light and magnetic materials. The science of morality has drawn heavily on well-controlled but artificial laboratory settings. To study everyday morality, we repeatedly assessed moral or immoral acts and experiences REFERENCES AND NOTES in a large (N = 1252) sample using ecological momentary assessment. Moral experiences were 1. B. Koopmans et al., Nat. Mater. 9, 259–265 (2010). surprisingly frequent and manifold. Liberals and conservatives emphasized somewhat different 2. A. Kirilyuk, A. V. Kimel, T. Rasing, Rep. Prog. Phys. 76, 026501 moral dimensions. Religious and nonreligious participants did not differ in the likelihood or

(2013). Downloaded from quality of committed moral and immoral acts. Being the target of moral or immoral deeds 3. E. Beaurepaire, J. Merle, A. Daunois, J. Bigot, Phys. Rev. Lett. 76, 4250–4253 (1996). had the strongest impact on happiness, whereas committing moral or immoral deeds had 4. S. O. Mariager et al., Phys. Rev. Lett. 108, 087201 (2012). the strongest impact on sense of purpose. Analyses of daily dynamics revealed evidence for 5. A. Cavalleri et al., Phys. Rev. Lett. 87, 237401 (2001). both moral contagion and moral licensing. In sum, morality science may benefit from a 6. M. H. Kryder et al., Proc. IEEE 96, 1810–1835 (2008). closer look at the antecedents, dynamics, and consequences of everyday moral experience. 7. B. C. Stipe et al., Nat. Photonics 4, 484–488 (2010). 8. C. D. Stanciu et al., Phys. Rev. Lett. 99, 047601 (2007). 9. S. Alebrand et al., Appl. Phys. Lett. 101, 162408 (2012). ow people distinguish between actions moral vignettes, questionnaire data, and thought 10. A. Hassdenteufel et al., Adv. Mater. 25, 3122–3128 (2013). that are “right” and “wrong” affects many experimentssuchastrolleyproblems(10). As im- – 11. S. Mangin et al., Nat. Mater. 13, 286 292 (2014). important aspects of life. Morality science— portant as these approaches are, they are all lim- 12. J. P. van der Ziel, P. S. Pershan, L. D. Malmstrom, Phys. Rev. Lett. 15, 190–193 (1965). informed by philosophy, biology, anthro- ited to some extent by the artificial nature of the H — 13. A. Kirilyuk, A. V. Kimel, T. Rasing, Rev. Mod. Phys. 82, pology, and psychology seeks to understand stimuli used and the non-natural settings in which 2731–2784 (2010). how the moral sense develops (1, 2), how moral they are embedded. Despite considerable scientific 14. C. D. Stanciu et al., Phys. Rev. B 73, 220402 (2006). judgments are made (3, 4), how moral experiences and practical interest in issues of morality, virtually 15. I. Radu et al., Nature 472, 205–208 (2011). 16. T. A. Ostler et al., Nat. Commun. 3, 666 (2012). differ among individuals, groups, and cultures no research has taken morality science out of these 17. J. H. Mentink et al., Phys. Rev. Lett. 108, 057202 (2012). (5–8), and what the psychological implications artificial settings and directly asked people about 18. C. E. Graves et al., Nat. Mater. 12, 293–298 (2013). of the morally “good” or “bad” life are (9). whether and how they think about morality and 19. E. Turgut et al., Phys. Rev. X 3, 039901 (2013). Insights from contemporary morality research immorality in the course of their everyday lived 20. E. Turgut et al., Phys. Rev. Lett. 110, 197201 (2013). 21. M. Battiato, K. Carva, P. M. Oppeneer, Phys. Rev. Lett. 105, have mostly been gained through the analysis of experience. Here we present an attempt to cap- 027203 (2010). ture moral events, experiences, and dynamics as 22. G. Malinowski et al., Nat. Phys. 4, 855–858 (2008). 1Department of Psychology, University of Cologne, 50931 they unfold in people’snaturalenvironments. 2 23. B. Varaprasad, M. Chen, Y. K. Takahashi, K. Hono, IEEE Trans. Cologne, Germany. Department of Psychology, University of Using ecological momentary assessment (11), Magn. 49, 718–722 (2013). Illinois, Chicago, IL 60607, USA. 3Department of Social 24. O. Hellwig, A. Berger, J. B. Kortright, E. E. Fullerton, J. Magn. Psychology, Tilburg University, 5000, Tilburg, Netherlands. we addressed a number of fundamental key issues Magn. Mater. 319,13–55 (2007). *Corresponding author. E-mail: [email protected] in scientific and public debates about morality:

1340 12 SEPTEMBER 2014 • VOL 345 ISSUE 6202 sciencemag.org SCIENCE Corrected 11 May 2015; see full text. RESEARCH | REPORTS

(i) How often do people commit moral and im- understudied dimensions of morality. (iii) Given wise boost momentary happiness and sense of moral acts in their daily lives? How often are they the ongoing debate about whether religion is a purpose? (v) Finally, our approach affords the pos- the targets of moral and immoral acts? How often necessary foundation for morality (15, 16), is there sibility to study the temporal dynamics of morality. do they witness moral and immoral acts in their evidence that religious peopleactuallycommit Forinstance,arepeoplemorelikelytodosometh- environment, or learn about them through indi- more moral or fewer immoral deeds than nonre- inggoodiftheyhavebecomethetargetsofa rect channels such as social media? (ii) What are ligious people? And can we replicate evidence for moral deed themselves (moral contagion)? And these moral experiences about? In particular, we a political morality divide between liberals and can we replicate moral self-licensing effects dem- examined how well an influential taxonomy of conservatives (6, 17)? (iv) What is the empirical onstrated in the lab (18) in the context of everyday moral dimensions, moral foundations theory connection between morality, momentary happi- social interaction, whereby committing a prior [MFT (12–14)], can account for descriptive content, ness, and meaning in life (i.e., sense of purpose)? moral act leads people to relax their moral stan- and whether everyday moral experiences highlight For instance, does committing moral deeds like- dards with regard to subsequent behavior? We recruited a large, demographically and ge- ographically diverse sample (1252 adults aged 18 to 68 years) from the United States and Canada. Each participant was randomly signaled five times daily on his or her smartphone for 3 days between 9 a.m. and 9 p.m. At each assessment, participants indicated whether they committed, were the target of, witnessed, or learned about a moral or immoral act within the past hour (they could also respond “none of the above”). For each moral or immoral event, participants described via text entry what the event was about. They also provided contextual information on the moral event (e.g., location) and completed state measures of nine distinct moral emotions such as guiltanddisgust[onascalefrom0(notatall) to 5 (very much)], momentary happiness [“How happy do you feel at the moment?” from –3 (very unhappy) to +3 (very happy)], and sense of purpose [“Do you feel that your life has a clear sense of purpose at the moment?” from 0 (not at all) to 4 (very much)]. Religiosity and political

Fig. 1. Overall distribution of responses (pie chart) and of the eight moral event categories (bar graph) in relation to total responses (left axis) and morally relevant responses (right axis). Type of event and moral valence (moral versus immoral) were statistically associated; see main text.

Fig. 3. Morality, happiness, and purpose. Moral acts were associated with relative gains in happiness relative to baseline, immoral acts with relative losses (upper panel). The happiness effect was strongest when people reported being the target of Fig. 2. Moral content and liberal versus conservative political ideology. Percentage distributions of moral/immoral acts. Sense of purpose was most moral and immoral event codings by political ideology (liberals versus conservatives) are shown on moral strongly affected by the commission of moral as dimensions. Moral as well as immoral events within each political ideology add up to 100%. Moral content compared to immoral acts (lower panel). *P < and political ideology were statistically associated; see main text. 0.05, **P <0.01,***P <0.001.

SCIENCE sciencemag.org 12 SEPTEMBER 2014 • VOL 345 ISSUE 6202 1341 Corrected 11 May 2015; see full text. RESEARCH | REPORTS ideology were assessed during an intake survey (13.9%), Honesty/Dishonesty (12.8%), Authority/ rated moral and immoral deeds committed by reli- upon study registration. Subversion (5.6%), Sanctity/Degradation (5.2%), giouspeopleasequallyrightandwrong,respec- Participants furnished a total of 13,240 reports Loyalty/Disloyalty (4.8%), Self-Discipline/Lack of tively (table S5). However, assessing the average (at a median response rate of 80%). On 28.9% Self-Discipline (3.8%), and Liberty/Oppression emotional “footprints” of these acts revealed that, of responses (n = 3828), participants reported a (3.3%). These results confirm MFT’soriginalcore relative to nonreligious people, religious people moral or immoral event. Moral events (15.3%; dimensions in spontaneously generated partici- experienced more intense self-conscious emotions n = 2029) and immoral events (13.6%; n = 1799) pant responses from everyday life, but also sug- such as guilt, embarrassment, and disgust in re- had similar overall frequencies. As shown in Fig. 1, gest additional categories associated with honesty sponse to the immoral deeds they had committed, type of event and moral valence were associated and self-discipline. andmoreprideandgratefulnessinresponseto such that people were more likely to report com- Political ideology was reliably associated with moraldeeds(fig.S3).Viewedinconcert,thesefind- mittingorbeingthetargetofamoralversusan moral content [c2(7) = 81.9, P <0.001,Cramer’s ings suggest that religious and nonreligious people immoral act, and were more likely to learn about V = 0.18]. Descriptive percentage distributions commit comparable moral and immoral deeds and an immoral rather than a moral act [c2(3) = 483.6, (Fig. 2) show that liberals mentioned events re- with comparable frequency. However, religious peo- P <0.001,Cramer’s V =0.36].Thefindingthat lated to Fairness/Unfairness, Liberty/Oppression, ple respond more strongly in psychological terms people were more than twice as likely to find out andHonesty/Dishonestymorefrequentlythandid to the immoral and moral deeds they commit. about immoral rather than moral acts through conservatives, whereas conservatives mentioned Does morality have implications for happiness personal communication (e.g., gossiping) and events related to Loyalty/Disloyalty, Authority/ and sense of purpose? Figure 3 (upper panel) other channels (see fig. S1 for more information Subversion, and Sanctity/Degradation more fre- displays momentary happiness levels for the four on this category) fits well with social-psychological quently than did liberals. More focused tests es- types of events relative to the baseline level of hap- theories of the function of gossip and evolutionary tablished that these differences were most reliable piness observed for nonmoral events. Moral acts theories of reputation management (19, 20). Ex- for Fairness/Unfairness, Loyalty/Disloyalty, Sanctity/ were associated with higher levels of momentary cluding learned-about acts (for which location Degradation, Liberty/Oppression, and Honesty/ happiness than immoral acts, as indicated by a data are often unknown), most moral or immoral Dishonesty, and remained largely stable when con- large main effect (F1,3441 =758.4,P <0.001).This acts happened in public settings (64.3%), fol- trolling for religiosity, which accounted for signif- happiness effect was moderated by type of event lowed by participants’ homes (23.4%), the homes icant portions of variance on four of the eight (F3,3382 = 22.7, P < 0.001) such that the strongest of close others such as family and friends (6.5%), dimensions (see table S3). Thus, our everyday-life gain and loss in happiness was observed when and other settings (5.7%). There was no statis- approach largely corroborates the idea that po- participants were the targets of moral and im- tical association between location and whether litical ideology relates to different moral emphases, moral acts, respectively (Cohen’s d = 1.34). Happi- an act was moral or immoral [c2(3) = 3.3, P = even though real-world effects appear to be more ness effect sizes were comparatively smaller when 0.343, Cramer’s V = 0.04]. amatterofnuanceratherthanstarkcontrast. committing moral or immoral acts (d=0.85), wit- Building on MFT (14), we reliably classified Comparing religious and nonreligious par- nessing them (d =0.73),orlearningaboutthem moral events into five originally proposed core ticipants, there was no discernible difference in (d = 0.57). In contrast, sense of purpose (Fig. 3, dimensions (Care/Harm, Fairness/Unfairness, the frequency of positive moral experience (both lower panel) was most strongly affected by the Loyalty/Disloyalty, Authority/Subversion, Sanctity/ overall and by event; table S4). Thus, we did not commission of moral as compared to immoral Degradation) as well as a newly proposed Liberty/ find evidence for religious people committing acts (interaction effect F3,2172 =5.1,P = 0.002). Oppression dimension (17) and two additional cat- moral acts more frequently than nonreligious Hence, whereas benefiting from others’ good deeds egories derived from our data (Honesty/Dishonesty, people. Religious people reported fewer immoral grants the highest observed levels of momentary Self-Discipline/Lack of Self-Discipline) (supple- experiences overall, but this difference was mostly happiness among types of events, doing good mentary materials and table S1). The first five core attributable to religious people reporting having lends the most purpose to people’s lives. dimensions accounted for 80.1% of moral events learnedaboutimmoralactslessoften—a possible Finally, we investigated whether the likelihood mentioned in daily life (86.1% of moral acts, 73.5% result of selective exposure—rather than having of committing a moral or immoral deed can be pre- of immoral acts). Care/Harm was by far the most committed immoral deeds less often than non- dicted by moral events that happened earlier to a frequently mentioned dimension (50.6%, in par- religious people (table S4). Moreover, a second person on a given day (table S6). In support for ticular accounting for a large share of morally sample of independent judges (N =249)whowere moral contagion, becoming the target of a moral good acts; table S2), followed by Fairness/Unfairness presented with participants’ open text descriptions act was associated with an above-average likelihood

Fig. 4. Moral dynamics. Predicted probabilities with which a prior event is followed by at least one committed moral act (left panel) or immoral act (right panel) on the same day, relative to the average sample probability (vertical axis), are shown. Having been the target of an earlier moral act was associated with an above-average likelihood of subsequent moral behavior. Having committed an earlier moral act was associated with an above-average likelihood of subsequent immoral behavior. *P <0.05,**P <0.01.

1342 12 SEPTEMBER 2014 • VOL 345 ISSUE 6202 sciencemag.org SCIENCE Corrected 11 May 2015; see full text. RESEARCH | REPORTS of committing a moral act later (Fig. 4). In addition, BIODIVERSITY LOSS a moral self-licensing pattern emerged (18), such that committing a moral act earlier in the day was associated with an above-average likelihood of a Loss of avian phylogenetic diversity in subsequent immoral act and a decreased likelihood of a subsequent moral act (Fig. 4). Together, the analysis of everyday moral dynamics revealed evi- neotropical agricultural systems dence both for moral contagion through other people’sgooddeedsandmoralself-licensingthrough Luke O. Frishkoff,1,2*† Daniel S. Karp,3,4*† Leithen K. M’Gonigle,3 one’s own good deeds outside of the laboratory. Chase D. Mendenhall,1,2 Jim Zook,5 Claire Kremen,3 Given these different mechanisms, it seems impor- Elizabeth A. Hadly,1 Gretchen C. Daily1,2,6,7,8 tant to find out more about how the principles of moral contagion can be used in public policy inter- Habitat conversion is the primary driver of biodiversity loss, yet little is known about how ventions, and how moral slacking may be prevented. it is restructuring the tree of life by favoring some lineages over others. We combined a By tracking people’s everyday moral experiences, complete avian phylogeny with 12 years of Costa Rican bird surveys (118,127 detections we corroborated well-controlled but artificial labo- across 487 species) sampled in three land uses: forest reserves, diversified agricultural ratory research, refined prior predictions, and made systems, and intensive monocultures. Diversified agricultural systems supported 600 illuminating discoveries about how people expe- million more years of evolutionary history than intensive monocultures but 300 million rience and structure morality, as well as about fewer years than forests. Compared with species with many extant relatives, evolutionarily how morality affects people’s happiness and sense distinct species were extirpated at higher rates in both diversified and intensive of purpose. A closer, ecologically valid look at how agricultural systems. Forests are therefore essential for maintaining diversity across the morality unfolds in people’s natural environments tree of life, but diversified agricultural systems may help buffer against extreme loss of may inspire new models and theories about what phylogenetic diversity. it means to lead the “good” or “bad” life. s human-converted habitats expand over tree of life are at greater risk than others (5, 6, 16), REFERENCES AND NOTES Earth’s surface, the fate of global biodiver- although whether evolutionarily distinct species 1. L. Kohlberg, in Moral Development and Behavior, T. Lickona, Ed. – sity will depend increasingly on the quality aremoreatriskthanspecieswithmanyliving (Holt, Rinehart & Winston, New York, 1976), pp. 31 53. 6 16 17 2. F. Warneken, M. Tomasello, Science 311, 1301–1303 (2006). and characteristics of farming landscapes relatives remains contested ( , , ). A 1 2 3. J. Haidt, Psychol. Rev. 108,814–834 (2001). ( , ). Agricultural systems vary widely in We quantified changes in phylogenetic di- 4. J. Haidt, Science 316, 998–1002 (2007). their ability to support biodiversity, with many versity across multiple landscapes in Costa Rica, 5. J. Haidt, S. H. Koller, M. G. Dias, J. Pers. Soc. Psychol. 65, – species extirpated from some but sustained in combining a recent complete avian phylogeny 613 628 (1993). 1 3 18 6. J. Graham, J. Haidt, B. A. Nosek, J. Pers. Soc. Psychol. 96, others ( , ). Additionally, characteristics of the ( ) with temporally and spatially extensive trop- 1029–1046 (2009). species themselves, evolved over millions of years, ical bird censuses to assess how habitat conver- 7. E. Turiel, The Culture of Morality: Social Development, Context, maypredisposesomelineagestobenefit(orsuf- sion is restructuring the avian phylogeny (19). and Conflict (Cambridge Univ. Press, Cambridge, 2002). fer) from human environmental impacts (4–6). The data set comprised 44 transects, surveyed 8. L. J. Skitka, C. W. Bauman, E. G. Sargis, J. Pers. Soc. Psychol. 88, 895–917 (2005). Phylogenetic diversity, the total evolutionary in wet and dry seasons over 12 years (2001 to 9. K. Aquino, A. Reed 2nd, J. Pers. Soc. Psychol. 83, 1423–1440 history or phylogenetic branch lengths of all spe- 2012) across four regions in two biomes (fig. S1). (2002). cies in a community (7), is recognized as having Transects were located in three land-use types: 10. J. D. Greene, R. B. Sommerville, L. E. Nystrom, J. M. Darley, intrinsic conservation value (8, 9). Also, ecolog- forest reserves, diversified agricultural systems, J. D. Cohen, Science 293, 2105–2108 (2001). 11. S. Shiffman, A. A. Stone, M. R. Hufford, Annu. Rev. Clin. ical experiments in small plots indicate that com- and intensive monocultures. Compared with in- Psychol. 4,1–32 (2008). munities with more phylogenetic diversity are tensive monocultures, diversified agricultural 12. J. Graham et al., J. Pers. Soc. Psychol. 101, 366–385 (2011). more stable (10), possess higher productivity (11), systems had more crop types, complex configu- 13. J. Haidt, J. Graham, Soc. Justice Res. 20,98–116 (2007). 14. J. Graham et al., Adv. Exp. Soc. Psychol. 47,55–130 (2013). and support more species at other trophic levels rations of vegetation, and substantial surround- 15. R. Dawkins, The God Delusion (Bantam, New York, 2006). (12). Despite the known impact of agriculture on ing tree cover (1) (table S1). Our analysis focused 16. S. Harris, The End of Faith: Religion, Terror, and the Future of species loss, how habitat conversion affects phy- on three unresolved questions. First, do certain Reason (Norton, New York, 2004). logenetic diversity remains unknown. Studies of bird clades thrive in agriculture, or is this capac- 17. J. Haidt, The Righteous Mind: Why Good People Are Divided by Politics and Religion (Pantheon, New York, 2012). plants and invertebrates have established that ity broadly distributed across the tree of life? 18.D.T.Miller,D.A.Effron,Adv. Exp. Soc. Psychol. 43, 115–155 (2010). local environmental disturbances (e.g., lake acid- Second, how much phylogenetic diversity is lost 19. R. F. Baumeister, L. Zhang, K. D. Vohs, Rev. Gen. Psychol. 8, ification and species invasion) favor subsets of when native forest is replaced with agriculture? 111–121 (2004). 20. R. I. M. Dunbar, Rev. Gen. Psychol. 8, 100–110 (2004). closely related clades and often result in phylo- Last, are evolutionarily distinct species capable of genetic diversity loss (13–15). Further, some studies persisting in agriculture? ACKNOWLEDGMENTS that examine the global extinction risk of birds and We found that clades from across the bird Supported by a SAGE Award from the Foundation for Personality mammals suggest that particular branches of the phylogeny thrived in agriculture (Fig. 1). Affinity and Social Psychology and a research grant from the University of for different habitats showed phylogenetic sig- Chicago Booth School of Business (W.H.), where portions of this work 1 were completed. The study was approved by the institutional review Department of Biology, Stanford University, Stanford, CA nal, meaning that closely related species were 2 board of the University of Chicago. W.H. is a co-founder of SurveySignal, 94305, USA. Center for Conservation Biology, Stanford more likely to share habitat preferences than 3 a University spin-off of the Web-based experience-sampling software University, Stanford, CA 94305, USA. Department of species that were distantly related (table S2) (20). used for this project, and has been advising the development of Environmental Science, Policy, and Management, University 4 The phylogenetic signal was best described by SurveySignal on an unpaid basis. The replication data are available of California, Berkeley, CA 94720, USA. Nature 5 ’ at the Harvard Dataverse Network (doi:10.7910/DVN/26910). Conservancy, Berkeley, CA 94705, USA. Unión de using Pagel s lambda transformation of the phy- Ornitólogos de Costa Rica, Apartado 182-4200, Naranjo de logeny (21), which reduces the degree of correla- 6 SUPPLEMENTARY MATERIALS Alajuela, Costa Rica. Woods Institute for the Environment, tion of traits between species below the Brownian Stanford University, Stanford, CA 94305, USA. 7Global www.sciencemag.org/content/345/6202/1340/suppl/DC1 Economic Dynamics and the Biosphere, Royal Swedish motion expectation (across habitat types and Materials and Methods Academy of Sciences, SE-104 05 Stockholm, Sweden. seasons, l = 0.25 to 0.48; table S3). Although most Figs. S1 to S3 8Stockholm Resilience Center, University of Stockholm, Tables S1 to S6 taxonomic families had species associated with SE-106 91 Stockholm, Sweden. References (21–27) all habitat types, some families tended to affil- *These authors contributed equally to this work. †Corresponding 30 January 2014; accepted 5 August 2014 author. E-mail: [email protected] (L.O.F.), dkarp@berkeley. iate with particular habitats. For example, pi- 10.1126/science.1251560 edu (D.S.K.) geons, seedeaters, swallows, and blackbirds were

SCIENCE sciencemag.org 12 SEPTEMBER 2014 • VOL 345 ISSUE 6202 1343 Morality in everyday life Wilhelm Hofmann et al. Science 345, 1340 (2014); DOI: 10.1126/science.1251560

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.

The following resources related to this article are available online at www.sciencemag.org (this information is current as of May 18, 2015 ):

A correction has been published for this article at: http://www.sciencemag.org/content/348/6236/aac5401.full.html Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/345/6202/1340.full.html on May 18, 2015 Supporting Online Material can be found at: http://www.sciencemag.org/content/suppl/2014/09/11/345.6202.1340.DC1.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/345/6202/1340.full.html#related This article cites 18 articles, 5 of which can be accessed free: http://www.sciencemag.org/content/345/6202/1340.full.html#ref-list-1 www.sciencemag.org This article has been cited by 8 articles hosted by HighWire Press; see: http://www.sciencemag.org/content/345/6202/1340.full.html#related-urls This article appears in the following subject collections: Psychology http://www.sciencemag.org/cgi/collection/psychology Downloaded from

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.

www.sciencemag.org/content/345/6202/1340/suppl/DC1

Supplementary Materials for

Morality in Everyday Life

Wilhelm Hofmann,* Daniel C. Wisneski, Mark J. Brandt, Linda J. Skitka

*Corresponding author. E-mail: [email protected]

Published 12 September 2014, Science 345, 1340 (2014) DOI: 10.1126/science.1251560

This PDF file includes: Materials and Methods Figs. S1 to S3 Tables S1 to S6 References Morality in Everyday Life  SOM 2

Supporting Online Materials

Materials and Methods

Sample

Participants were recruited between January 2013 and June 2014 via various forms of advertising (Craigslist, Backpage, Facebook, Twitter, various local newspaper ads, blog and forum ads, crowdsourcing pages, university mailing lists) in two major waves throughout the US

(93.2%) and Canada (6.8%). 1,252 participants (51.8% female) who successfully registered and verified their smartphone and completed the screening and intake survey were retained for analysis. 123 participants (9.8%) did not report any moral or immoral experience (37% of these had overall response rates below 20%). The average age was 31.9 years (SD = 9.96), ranging from 18 to 68. 73.0% of participants were Caucasian, 5.7% were African American, 7.3%

Hispanic/Latino, 9.0% Asian, 0.6% Native American or Pacific Islander, and 4.5% were of other backgrounds. Regarding the highest level of education, 0.3% indicated “some high school,”

4.2% “completed high school,” 35.2% “some college,” 31.1% “completed college,” and 29.2% indicated “advanced/post-graduate studies.” Overall, 32.9% of participants indicated that they are currently a college student. Using the modified MacArthur 10-rung ladder measure of social class (21), class was approximately normally distributed around an average of 5.96 (SD = 1.67).

Regarding smartphone operating systems, 53.1% of participants used smartphones running iOS

(i.e., iPhone), 44.4% Android, 1.4% Windows Mobile, 0.4 Blackberry (RIM), 0.3% Symbian, and 0.5% other. Taken together, the sample can be described as relatively heterogeneous compared to the typical university student sample employed in much laboratory research.

Ecological Momentary Assessment Procedure 2

Morality in Everyday Life  SOM 3

Mobile Application

To allow using people’s own smartphones independent of operating systems for online ecological momentary assessment, we chose an SMS survey distribution approach. The approach utilizes cellphone text messaging as a signaling device; the text signals contain a hyperlink that directs the participant to an online survey displayed on the smartphone’s browser, allowing for secure online data collection. We developed an extensive web-based experience-sampling application (SurveySignal) with the help of a professional programmer, using an asp.net framework in conjunction with a third-party SMS gateway. The application controlled all aspects of the smartphone experience-sampling phase including signup, the scheduling of signals as text messages to participants’ smartphones, link timeout, reminder SMS, and the registration of responses. After having passed a mobile compatibility test (passed by 91% of interested participants), a short screening survey (e.g., regarding age, country, language skills, smartphone ownership, data plan) and having provided informed consent, participants registered and verified their smartphone in the system. They then filled out an intake survey assessing demographic and personality information and received more information on the upcoming experience-sampling phase, starting on the following day. On each day of the three-day experience sampling schedule, five signals were distributed throughout a time window of 12 hrs. from 9 am to 9 pm. Following recommendations of Hektner, Schmidt, and Csikszentmihalyi (22), this time window was divided into five blocks of 144 minutes each; within each block, an exact signal time was randomly selected with the condition that two consecutive signals be at least 45 minutes apart.

Embedded in each text message was an individualized link directing participants to the online experience sampling survey (“mobile survey”) which was created and optimized for mobile display. The link carried information on the schedule day, signal number, send time, as well as a 3

Morality in Everyday Life  SOM 4

recipient identifier that allowed matching a given participant’s response.

Each survey link was valid for a maximum duration of 2 hrs. However, participants were

encouraged in the mobile phase information booklet to respond as soon as possible to signals and

to try their best to minimize the number of times they needed to delay responding. If a response was not received within the first 15 minutes after the signal was dispatched, the application sent out a reminder SMS. The median delay in responding was 5 minutes (including situations in which participants may have turned off their smartphone and received the text signal at a later point in time than originally sent out). Only one mobile survey response per measurement occasion was allowed. To reduce participant burden, data on the nine specific moral emotions was assessed on a random subset of 50% of occasions only. Two items (sense of purpose, feelings of pride) were only included for the second wave of data collection.

The overall number of returned mobile surveys was 13,240. On average, participants replied to 10.6 out of the 15 signals sent, indicating a satisfactory average response rate of 71%

(median response rate: 80%). Excluding participants who showed a very low response rate did not change our conclusions, so we report the findings for the full sample here. The average completion time per mobile survey was 5.0 minutes (excluding extreme outliers). A small

fraction of mobile surveys was only partially completed (3.8%).

Coding of Participant Responses

We built on moral foundations theory (12, 13) to develop a coding scheme that allows

categorizing the reported moral and immoral event text entries according to the original

dimensions (Care/Harm; Fairness/Unfairness; Loyalty/Disloyalty, Authority/Subversion;

Sanctity/Degradation), as well as a recently proposed Liberty/Oppression dimension (17). In

light of the current discussion surrounding further possible candidates for foundationhood, we 4

Morality in Everyday Life  SOM 5

assigned events related to Honesty/Dishonesty a dimension of its own (12, 23, 14). Furthermore,

a preliminary screening of responses showed that a considerable number of participants

mentioned strong displays of effort and restraint (e.g., resistance to temptation) as moral and lack

of effort and restraint as immoral events. In line with recent theoretical proposals (24), this observation led us to add the dimension of Self-Discipline (see Supplementary Table S1 for an overview of all eight coding dimensions).

We made a small number of informed coding decisions regarding events at the boundaries of these key moral dimensions or events with possible overlap among dimensions.

For instance, we coded instances of infidelity as disloyalty (to the partner), and we classified the

enactment of religious activities such as praying as related to sanctity. Table S1 summarizes the

taxonomy and provides concrete examples. For a small fraction of responses, participants did not

provide a text entry (1.5% of responses), so these entries were excluded from content analyses.

Moreover, an entry was classified as uncodeable if the response was too vague to allow for any

content coding (3.8% of all morally relevant responses). One coder (Wilhelm Hofmann) blind-

coded all responses according to the final taxonomy. A second coder (Daniel Wisneski) blind-

coded 50% of all responses to determine interrater-agreement which was very good, kappa = .85.

During the coding, a small number of erroneous participant categorizations (2.2 %) were

detected and corrected based on the written content and the valence ascribed to the event (e.g.,

participant categorization: “I committed a moral act”; Text entry: “I stole something”; How

right/wrong was the act?: “very wrong” was reclassified as “I committed an immoral act”). In a

similar vein, we checked location text entries under the category “other” for instances that could

be recoded into an existing category (e.g., “in a restaurant” was assigned to the category “public

setting”; 4.3% of location responses). 5

Morality in Everyday Life  SOM 6

Independent Judges’ Ratings of Committed Moral and Immoral Deeds

We collected outside observer’s ratings for all committed moral and immoral acts in the

database to (a) gauge the extent to which perceptions of what constitutes moral and immoral

deeds are socially shared versus purely subjective, and to (b) get at an independent measure of the “intensity” of moral and immoral deeds reported. Each open-ended text responses was rated by 11 to 13 independent raters from an overall pool of 249 US participants recruited via the

Mechanical Turk crowdsourcing website (50.6% female; mean age = 36.92, SD = 13.5). Raters were blind to the purpose of the study and did not have any information on the person who had written the original text entry. Each rater received a manageable subset of about 60 statements to code. For each presented statement, raters were asked “How morally right or wrong do you consider each of the following acts or deeds?” on a scale from -3 (totally wrong/immoral) to +3

(totally right/moral), with the midpoint of zero marked neutral/neither-nor, and were provided with the additional response option “can’t say”.

On 74.9 % of all judgments, individual raters judged an act as falling into the moral or immoral side of the scale in a way that was consistent with the event classification by the participant as moral or immoral. On 20.5% of judgments, the neutral midpoint was used and

1.9% of the time, the can’t say option was chosen. The remaining 2.7% were cases for which raters judged the reported act in a way that mismatched the participant’s categorization (e.g., participant classifies as moral, rater as immoral or vice versa). Consensus among raters was high on average, as indicated by a mean intraclass correlation coefficient of .89, so we averaged the

available ratings for each act. A kernel density plot displayed in Figure S2 shows that the

distribution of average ratings were clearly separable for moral and immoral committed acts.

Using these average ratings, 8.5% of statements were judged to be either not enough morally 6

Morality in Everyday Life  SOM 7 relevant (defined as a mean morality rating falling around the midpoint (0) of the scale from -.5 to +.5). 39.8 percent of cases were rated as falling around the “somewhat moral/immoral” scale point (i.e., an absolute mean rating between 0.5 and 1.5), 46.2% of cases as falling around the

“moderately moral/immoral” scale point (i.e., absolute mean rating between 1.5 and 2.5), and

5.2% of cases achieved average ratings close to the endpoints of the scales (i.e., absolute mean rating between 2.5 and 3). On the aggregate level, only 0.4% of cases were judged by raters in a way that mismatched the participant’s own categorization. Taken together, these results suggest considerable consensus among moral agents and independent judges, speaking against the notion that participants in our study held entirely subjective constructions of their own moral or immoral acts. At the same time, the consensus as to what constitutes morality was far from perfect, especially at the level of individual judgments.

Religiosity and political ideology

Religiosity was measured with the question “How religious are you?” on a scale from 1

(not at all) to 7 (very much). Because our main interest was in the general comparison of non- religious and religious people and because the religiosity measure was very far from a normal distribution, we created a simple contrast by classifying those participants as “non-religious who indicated that they were not religious at all (43.4%) and the remaining participants who indicated at least some degree of religiosity as religious (56.6%). The religious category had the following distribution of stated religious affiliations: 27.7% Catholic, 22.5% Protestant, 1.8% Orthodox,

29.5% “Christian” without further specification, 6.5% Jewish, 3.0% Muslim, 2.4% Buddhist,

1.9% Hindu, and 4.7% other world religions.

Political ideology was assessed with a hybrid measure incorporating a scale from 1 (very liberal) to 7 (very conservative) with 4 as the middle point (“middle of the road”), as well as the 7

Morality in Everyday Life  SOM 8

additional response options “libertarian”, “other”, and “apolitical” (“don’t know/not political”).

To conduct a focused test of the hypothesized effect of liberal versus conservative political ideology on moral content (Table S2), we computed a liberal vs. conservative contrast by classifying those participants as liberal who indicated weak to strong liberal views on the scale

(scores of 1 to 3; 49.4% of respondents) and those as conservative who indicated weak to strong conservative views (scores of 5 to 7; 20.6% of respondents; n = 258). The remaining distributions were: moderate (13.4%), libertarian (7.5%), apolitical (4.9%), and other (4.2%). To explore associations of political orientation with moral frequencies (Table S3) or control for political orientation as a covariate in other analyses involving the full sample, we employed a set of five effects-codes with the category “other” as the base category in the coding scheme. As in other research, religiosity and political ideology were correlated, r = .40, so we controlled for the influence of one demographic variable on the other in main or supplementary analyses.

Data analysis strategy

Descriptive content analyses were conducted by tabulating moral content codings by political ideology or religiosity. To evaluate the overall statistical association between liberal vs. conservative political orientation, we computed a Chi²-test of independence and Cramer’s V as a correlation coefficient (comparable to the Pearson coefficient for continuous measures).

Subsequent zero-inflated negative binomial models (with a logit link), capable of handling excess zeros and over-dispersion (25), were estimated for each moral dimension to determine which specific contrasts between liberal and conservative orientation were significant using the

R package MASS (26). In a similar vein, we used zero-inflated negative binomial models to explore the effects of religiosity and political ideology on moral frequencies. In both cases, univariate models were advocated over multivariate analyses because the average correlation 8

Morality in Everyday Life  SOM 9 among the dependent outcome variables was very low (average r = .06 for the analysis of moral content; average r = .05 for the analysis of moral event frequencies). Ordinary least-squares regression analysis was used to determine whether independent coders’ ratings of committed moral and immoral deeds, averaged for each participant in case of multiple deeds, could be predicted by participant religiosity and political ideology. A conceptually similar multilevel regression analysis using valence (moral/immoral) as a BY-variable yielded identical conclusions.

Due to the larger number of observations per participant, happiness and sense of purpose analyses were conducted with linear multilevel regression using the MIXED command in SPSS treating valence (moral/immoral) and type of event (committed, target of, witnessed, learned about) as categorical BY-variables. In a similar vein, we used multilevel modeling to estimate the intensity of nine moral emotions for committed moral and immoral acts as a function of religiosity.

To model temporal dynamics of prior moral events on the likelihood to commit a later moral or immoral act, we chose the day as the natural unit people may be applying to mentally account for and behaviorally trade-off moral behavior (later sensitivity analyses showed that conclusions also held when considering the whole time-window of the experience sampling phase). For each measurement occasion corresponding to the first, second, third, or fourth signal for each day (including occasions, on which no moral event was reported), we identified whether a moral or immoral act was committed at least once on any of the subsequent signals replied to on that day. If so, that measurement occasion was assigned an outcome value of 1 on the respective event (committed moral act/committed immoral act, respectively), otherwise a value of 0 was assigned. For instance, if a person was the target of a moral act on signal 2, and 9

Morality in Everyday Life  SOM 10

committed a moral act on signal 4, the measurement occasion for signal 2 received an entry of 1

for the dependent variable “committed a moral act later the same day.” Because events occurring

on signal 5 could not logically serve as prior events, the database for the dynamic analysis was

reduced to 10,713 observations. We ran logistic multilevel regression using HLM (27) to predict the likelihood of moral/immoral act occurrence from prior events (Table S6). Each of the eight morally relevant events was represented with an effects-coded variable and non-moral events formed the reference category of the coding scheme. Thus, each effect code tests the difference in predicted likelihoods from the sample average likelihood of occurrence. The underlying logic is that, if subsequent events are truly independent from prior events, these effect codes should not reliably differ from the grand average (represented as the vertical axes in the panels of Figure

4).

Morality in Everyday Life  SOM 11

Learned about moral act Learned about immoral act

Percentage 0% 5% 10% 15% 20% 25%

Personal communication

Social media

Internet news/magazine

Internet: other

TV: news

TV: other

Radio: news

Radio: other

Newspaper/magazine

Book

Other

Supplementary Figure S1. Sources of learned about moral and immoral acts as a proportion of

the overall number of responses (across moral and immoral acts). Overall, immoral acts were much more frequent than moral acts (see main text). The majority of acts were found out about through personal communication, internet news, and social media.

Morality in Everyday Life  SOM 12

1.0 Moral deeds committed by participants Immoral deeds committed by participants

0.8

0.6 Density

0.4

0.2

0.0

-3 -2 -1 0 1 2 3 totally immoral totally mor Degree of morality rated by judges

Supplementary Figure S2. Kernel density plot or independent judges’ ratings of moral deeds

(red line) and immoral deeds (black line) reported by participants via open text entry on a scale from -3 [totally wrong/immoral]to +3 [totally right/moral] (x-axis).

Morality in Everyday Life  SOM 13

Committed moral deeds Non-religious Religious

elevated 3.0 proud* 2.5 grateful* 2.0 1.5 1.0 disgusted 0.5 angry 0.0

embarrassed contemptuous

shameful guilty

Committed im moral deeds

Non-religious Religious

elevated 3.0 proud 2.5 grateful 2.0 1.5 1.0 disgusted* 0.5 angry* 0.0

embarrassed* contemptuous*

shameful* guilty*

Supplementary Figure S3. Emotional “footprints” of religious and non-religious participants as captured for the commission of moral deeds (upper panel) and immoral deeds (lower panel).

Emotional states were measured on a scale from 0 to 4. Emotions marked with an asterisk show a significant group difference at P < .01. 13

Morality in Everyday Life  SOM 14

Supplementary Table S1. Coding Taxonomy based on Moral Foundations Theory (12) and expanded by the dimensions of

Honesty/Dishonesty (14) and Self-Discipline/Lack of Self-Discipline (28).

Dimension Key Elements Example Responses

Care / Kindness, gentleness, nurturance, generosity, help Assisted a tourist with directions because he looked lost.

with regard to other living beings or environment. I gave a homeless man an extra sandwich that I had.

Harm (Threat of) physical or emotional harm to living A woman was driving and smoking a cigarette with small children in the car.

beings, environmental harm, lack of care. Hired someone to kill a muskrat that's ultimately not causing any harm.

Fairness / Concern for justice, equality/equity, and Talked to someone about treating others equally.

reciprocity. Reminded waitress I did not pay for my bill when she thought I did.

Unfairness Injustice, stealing, inequality/inequity, non- Congress making cuts across the board and not solving debt problems for the country.

reciprocity. At work someone stole my partner's nice balsamic vinegar while he was off shift and

most likely took it home with them.

Loyalty / Patriotism, self-sacrifice for the ingroup (e.g., Since this is Memorial Day, I've read a number of posts paying tribute to our veterans

family, friends, nation). and families that have lost a loved one.

I am putting my family before my own fun (chance to get drunk).

Disloyalty Lack of loyalty towards ingroup, betrayal of Gave up on my team.

ingroup. Arranging adulterous encounter.

Authority / Leadership, deference to legitimate authority, Enforced a rule.

Morality in Everyday Life  SOM 15

respect for traditions, obedience to laws and Appropriately disciplined a youth not my own.

regulations.

Subversion Disregard of legitimate authority, disrespect for Disrespecting my mother.

traditions, disobedience of laws and regulations. Had drinks with a colleague during work hours without the boss knowing.

Sanctity / Concern for elevation (including religious Talked about God with a family member.

activities) and standards of purity and sanctity. Yoga Nidra meditation class.

Degradation Violation of purity, decency, and religious Caught my teenage son looking at hard core porn.

standards. Woman made 3 year old eat her feces for having an accident.

Liberty / Concern for liberty, freedom I argued on behalf an oppressed population in a public setting.

An organization freed Beagles that had never seen daylight or felt grass, due to a life

of captivity for animal testing.

Oppression Reactance and resentment towards those who A female student in my daughter's class was denied a ticket to prom because she

dominate and restrict liberty wanted to take another girl as her date.

Kidnapped schoolgirls in Nigeria.

Honesty / Concern for honesty and truth, truth-telling, Honest about a sales initiative.

sincerity I found a lost cell phone and returned it to its owner.

Dishonesty Dishonesty, lying, cheating, insincerity Lied to someone by saying their suggestion was good. it wasn’t.

Student faked bomb threat to cover up not graduating.

Self-Discipline / Display of high levels of industriousness, effort, I followed through on completing a work commitment I made, even though I completed

Morality in Everyday Life  SOM 16

persistence, not giving up easily, controlling it after hours.

impulses, resisting temptation. Did not read confidential info not for me at work even though had access.

Lack of Self- Slacking, lack of effort and persistence on tasks and Child was sloppy in her work because she didn't want to do it

Discipline goals, failure to resist temptation. I sneaked and got fast food though I promised someone I wouldn't have it.

Morality in Everyday Life  SOM 17

Supplementary Table S2. Percentage distribution of coded responses along eight moral dimensions (overall, and separately for moral and immoral acts).

Overall Percentage Percentage Dimension Percentage moral immoral Care / Harm 50.6 70.0 29.0 Fairness / Unfairness 13.9 7.3 21.3 Loyalty / Disloyalty 4.8 3.4 6.4 Authority / Subversion 5.6 1.5 10.1 Sanctity / Degradation 5.2 3.9 6.7 Liberty / Oppression 3.3 1.2 5.6 Honesty / Dishonesty 12.8 8.6 17.5 Self-Discipline / Lack of Self- 3.8 4.1 3.4 Discipline

Morality in Everyday Life  SOM 18

Supplementary Table S3. Zero-inflated Negative Binomial regression analyses (with logit link) predicting the frequency of occurrence of moral dimensions from political ideology (liberal [0] vs. conservative [1] contrast), and political ideology controlling for religiosity (effects-coded). Results are in the original logit metric used for linear regression analysis.

Single Prediction Model Controlling for Religiosity

Political Ideology (lib. vs cons.) Political Ideology (lib. vs cons.) Religiosity

Outcome (Intercept) Blog SE P (Intercept) Blog SE P Blog SE P

Care / Harm 0.39 0.10 0.08 .213 0.41 0.12 0.09 .188 -0.04 0.08 .664 Fairness / Unfairness -0.80 -0.58 0.15 <.001 -0.67 -0.42 0.16 .006 -0.33 0.13 .008 Loyalty / Disloyalty -2.04 0.44 0.21 .034 -2.24 0.28 0.23 .225 0.40 0.23 .085 Authority / Subversion -1.78 0.14 0.20 .478 -1.93 0.01 0.22 .980 0.33 0.21 .113 Sanctity / Degradation -2.16 0.95 0.20 <.001 -2.48 0.71 0.21 .001 0.62 0.24 .010 Liberty / Oppression -1.73 -0.75 0.31 .015 -1.57 -0.36 0.33 .275 -0.82 0.26 .002 Honesty / Dishonesty -0.91 -0.47 0.15 .001 -0.75 -0.28 0.16 .075 -0.40 0.13 .002 Self-Discipline / Lack -1.68 0.16 0.23 .468 -1.61 0.24 0.25 .324 -0.19 0.23 .425 of Self-Discipline

Morality in Everyday Life  SOM 19

Supplemenary Table S4. Zero-inflated negative binomial regression analyses (with logit link) predicting the frequency of committed, received (target of), witnessed, and learned about moral and immoral acts from religiosity (non-religious [0] vs. religious [1] contrast) and political ideology (set of four effects-codes with “other” as base). Results are in the original logit metric used for linear regression analysis.

Religiosity Political Ideology (effects-coded)

Liberal Moderate Conservative Libertarian Apolitical

Outcome Blog (SE) P Blog (SE) P Blog (SE) P Blog (SE) P Blog (SE) P Blog (SE) P Grand total -0.07 (0.05) .157 0.02 (0.04) .572 0.04 (0.06) .515 0.06 (0.05) .242 0.10 (0.07) .174 -0.13 (0.09) .138 Moral Acts Overall Moral 0.00 (0.06) .986 -0.05 (0.06) .353 -0.01 (0.08) .907 0.13 (0.07) .052 0.07 (0.10) .472 -0.16 (0.12) .192 Committed 0.01 (0.09) .935 0.01 (0.08) .947 -0.06 (0.11) .599 0.08 (0.10) .423 0.19 (0.13) .141 -0.28 (0.18) .119 Target of 0.17 (0.15) .262 -0.07 (0.13) .579 -0.33 (0.20) .095 0.05 (0.16) .731 0.00 (0.23) .991 0.16 (0.26) .527 Witnessed 0.01 (0.12) .934 -0.16 (0.11) .124 0.11 (0.14) .461 0.32 (0.12) .009 0.08 (0.18) .643 -0.20 (0.23) .401 Learned About -0.16 (0.12) .157 -0.03 (0.10) .751 0.22 (0.13) .103 0.14 (0.13) .265 -0.16 (0.19) .389 -0.15 (0.23) .498 Immoral Acts Overall Immoral -0.14 (0.06) .019 0.11 (0.05) .034 0.10 (0.07) .186 -0.03 (0.07) .610 0.14 (0.09) .121 -0.09 (0.12) .423 Committed 0.12 (0.14) .363 0.21 (0.11) .068 -0.36 (0.19) .056 -0.04 (0.15) .769 0.19 (0.20) .354 -0.14 (0.26) .597 Target of -0.11 (0.18) .550 0.27 (0.18) .130 0.37 (0.23) .110 0.13 (0.22) .558 0.24 (0.29) .400 -0.30 (0.42) .472 Witnessed -0.04 (0.13) .763 0.02 (0.11) .875 0.19 (0.15) .194 -0.01 (0.14) .952 -0.06 (0.20) .765 -0.10 (0.24) .681 Learned About -0.28 (0.08) .001 0.10 (0.07) .174 0.16 (0.10) .100 -0.06 (0.10) .533 0.18 (0.12) .127 -0.04 (0.16) .798 Note. Religiosity was coded 0 (non-religious) versus 1 (religious). Significant (P < .05) regression weights are printed in bold.

Morality in Everyday Life  SOM 20

Supplemenary Table S5. Regression analyses of independent judges’ ratings of moral and immoral deeds committed by participants, measured on a scale from -3 [totally wrong/immoral] to +3 [totally right/moral], from religiosity (Step 1), and controlling for political ideology (Step 2). Participants’ religiosity did not account for significant (P < .05) variation in judges’ ratings of the rightness or wrongness of committed deeds.

Committed Moral Deeds Committed Immoral Deeds B SE P B SE P Step 1 (Intercept) 1.68 .037 < .001 -1.23 .058 < .001 Religiosity (non-religious [0] vs. religious [1]) -0.05 0.05 .346 0.13 0.08 .088 Step 2 (Intercept) 1.68 .047 < .001 -1.28 0.08 < .001 Religiosity (non-religious [0] vs. religious [1]) -0.06 0.05 .260 0.14 0.09 .121 Political Ideology (effects-coded) Liberal 0.00 0.05 .948 0.06 0.07 .374 Moderate 0.09 0.07 .172 0.02 0.12 .833 Conservative 0.02 0.06 .713 0.10 0.10 .319 Libertarian 0.01 0.08 .872 0.02 0.12 .843 Apolitical 0.00 0.10 .980 -0.22 0.17 .193

Morality in Everyday Life  SOM 21

Supplemenary Table S6. Logistic multilevel regression analyses (with logit link) predicting the likelihood with which a prior event is followed by at least one committed moral (left columns) or immoral act (right columns) on the same day, relative to the average sample probability (represented by the intercept).

Predicted likelihood of committing at Predicted likelihood of committing at least one moral act later that day least one immoral act later that day

Blog SE p Blog SE p (Intercept) -2.45 0.08 < .001 -3.72 0.11 < .001 Prior Event Predictors Committed, moral -0.50 0.11 <.001 0.46 0.17 .008 Committed, immoral -0.01 0.18 .956 -0.46 0.22 .035 Target of, moral 0.38 0.18 .033 -0.05 0.32 .889 Target of, immoral -0.04 0.25 .874 -0.50 0.46 .285 Witnessed, moral 0.03 0.16 .861 -0.03 0.29 .906 Witnessed, immoral -0.31 0.20 .116 0.09 0.28 .735 Learned about, moral 0.29 0.16 .070 0.21 0.26 .429 Learned about, immoral 0.15 0.12 .226 0.11 0.20 .581 Note. Significant (P < .05) regression weights are printed in bold. Likelihoods have been transformed into probabilities for illustration in Figure 4.

References

1. L. Kohlberg, in Moral Development and Behavior, T. Lickona, Ed. (Holt, Rinehart & Winston, New York, 1976), pp. 31–53.

2. F. Warneken, M. Tomasello, Altruistic helping in human infants and young chimpanzees. Science 311, 1301–1303 (2006). Medline doi:10.1126/science.1121448

3. J. Haidt, The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychol. Rev. 108, 814–834 (2001). Medline doi:10.1037/0033- 295X.108.4.814

4. J. Haidt, The new synthesis in moral psychology. Science 316, 998–1002 (2007). Medline doi:10.1126/science.1137651

5. J. Haidt, S. H. Koller, M. G. Dias, Affect, culture, and morality, or is it wrong to eat your dog? J. Pers. Soc. Psychol. 65, 613–628 (1993). Medline doi:10.1037/0022-3514.65.4.613

6. J. Graham, J. Haidt, B. A. Nosek, Liberals and conservatives rely on different sets of moral foundations. J. Pers. Soc. Psychol. 96, 1029–1046 (2009). Medline doi:10.1037/a0015141

7. E. Turiel, The Culture of Morality: Social Development, Context, and Conflict (Cambridge Univ. Press, Cambridge, 2002).

8. L. J. Skitka, C. W. Bauman, E. G. Sargis, Moral conviction: Another contributor to attitude strength or something more? J. Pers. Soc. Psychol. 88, 895–917 (2005). Medline doi:10.1037/0022-3514.88.6.895

9. K. Aquino, A. Reed 2nd, The self-importance of moral identity. J. Pers. Soc. Psychol. 83, 1423–1440 (2002). Medline doi:10.1037/0022-3514.83.6.1423

10. J. D. Greene, R. B. Sommerville, L. E. Nystrom, J. M. Darley, J. D. Cohen, An fMRI investigation of emotional engagement in moral judgment. Science 293, 2105–2108 (2001). Medline doi:10.1126/science.1062872

11. S. Shiffman, A. A. Stone, M. R. Hufford, Ecological momentary assessment. Annu. Rev. Clin. Psychol. 4, 1–32 (2008). Medline doi:10.1146/annurev.clinpsy.3.022806.091415 12. J. Graham, B. A. Nosek, J. Haidt, R. Iyer, S. Koleva, P. H. Ditto, Mapping the moral domain. J. Pers. Soc. Psychol. 101, 366–385 (2011). Medline doi:10.1037/a0021847

13. J. Haidt, J. Graham, When morality opposes justice: Conservatives have moral intuitions that liberals may not recognize. Soc. Justice Res. 20, 98–116 (2007). doi:10.1007/s11211- 007-0034-z

14. J. Graham, J. Haidt, S. Koleva, M. Motyl, R. Iyer, S. P. Wojcik, P. H. Ditto, Moral foundations theory: The pragmatic validity of moral pluralism. Adv. Exp. Soc. Psychol. 47, 55–130 (2013). doi:10.1016/B978-0-12-407236-7.00002-4

15. R. Dawkins, The God Delusion (Bantam, New York, 2006).

16. S. Harris, The End of Faith: Religion, Terror, and the Future of Reason (Norton, New York, 2004).

17. J. Haidt, The Righteous Mind: Why Good People Are Divided by Politics and Religion (Pantheon, New York, 2012).

18. D. T. Miller, D. A. Effron, Psychological license: When it is needed and how it functions. Adv. Exp. Soc. Psychol. 43, 115–155 (2010). doi:10.1016/S0065-2601(10)43003-8

19. R. F. Baumeister, L. Zhang, K. D. Vohs, Gossip as cultural learning. Rev. Gen. Psychol. 8, 111–121 (2004). doi:10.1037/1089-2680.8.2.111

20. R. I. M. Dunbar, Gossip in evoluationary perspective. Rev. Gen. Psychol. 8, 100–110 (2004). doi:10.1037/1089-2680.8.2.100

21. E. Goodman, N. E. Adler, I. Kawachi, A. L. Frazier, B. Huang, G. A. Colditz, Adolescents’ perceptions of social status: Development and evaluation of a new indicator. Pediatrics 108, e31 (2001). Medline doi:10.1542/peds.108.2.e31

22. J. M. Hektner, J. A. Schmidt, M. Csikszentmihalyi, Experience Sampling Method: Measuring the Quality of Everyday Life (Sage, Thousand Oaks, CA, 2006).

23. R. Iyer, “The Case for Honesty as a Moral Foundation.” Blog posting at PoliPsych.com (2010). 24. R. Janoff-Bulman, N. C. Carnes, Surveying the moral landscape: Moral motives and group- based moralities. Pers. Soc. Psychol. Rev. 17, 219–236 (2013). Medline doi:10.1177/1088868313480274

25. A. Zeileis, C. Kleiber, S. Jackman, Regression models for count data in R. J. Stat. Softw. 27, 1 (2008).

26. W. N. Venables, B. D. Ripley, Modern Applied Statistics with S (Springer, New York, ed. 4, 2002).

27. S. Raudenbush, A. Bryk, Y. F. Cheong, R. Congdon, HLM 6 for Windows (Scientific Software International, Skokie, IL, 2004).