6 Newman,Cain control condition along with a second paragraph that Morality.For ratings of morality,the predicted contrast the Gap was also signincant,(202) 99,1 003.The patterr in the well as an additional sen morality th the ontrol (M 5.86.SE= tence.which highlighted the profit to the company 0.19)and tainted-altruism (M=5.93.SE=0.24)condi Finally,in the counterfactual-information condition,pa tions compared with the altruism C=7 .12,5E=0.19 ame in the tainte and 6.91,SE=0.19 inded pa donate any mone to charity.(See the Supplemental Manipulativeness.Ratings of the Gap as selfish and Material for the full text used in each condition.) manipulative were consistent with the predicted pattern the company along the follow but the mora 202 ded) coded).They also rated how much they liked the com- Discussion pany,trusted it,and were willing to purchase its products All ratings were made on scales from 1 (low)to(high) Expe s were presented was ran experiment is notable for a number of reasons.First.it Results was ecologically valid in that it assessed evaluations in a real-world context (an ongoing charitable program by a otation indi ated tha th -kn any hat (ratings of liking.trust,and willingness to purchase the (merely calling attention to the fact that the com pany alsc ble,and altruistic the company Is tions of the company as well as participants 87).and manipulativen this sing their proc An 63).For each of these measures.we conducted a con terfactual thinking as an explanation of the effect.becaus trast analysis testing the prediction that ratings would be the salience of the profitability information was the criti significantly lower in the tainted-altruism condition than cal factor that differed across conditions in General Discussion Liking.The contrast analysis comparing the tainted- The experiments reported here demonstrate the exis altruism condition with the other three conditions was tence of a novel psychological phenomen on:the tainted 8, As we prec t.In sum.we that the presence 5.16.SE=0.29)than in the control (M=5.74.SE=0.23) ing.such that ple udged efforts that realized both altruism (M=5.91,SE=0.24),and counterfactual-infor personal and charitable benefits to be less moral than 5.83.SE=0.21).which did not sself-interested ehaviors that produced no This effect was observed across a Table 1.Mean Scores for Items Loading on Each Factor in Experiment 4 Factor Condition Liking Morality Manipulativeness 612 712135 Tainted altruisn 5.16(2.05 5.93(1.73 5.14(1.79) Counterfactual information 5.83(1.51) 6.91(1.34 5.46(1.97 Cai Xing on Jausry 21.2014 6 Newman, Cain control condition along with a second paragraph that described the Gap (RED) campaign. In the tainted-altru￾ism condition, participants read the same information as in the altruism condition as well as an additional sen￾tence, which highlighted the profit to the company. Finally, in the counterfactual-information condition, par￾ticipants read the same information as in the tainted￾altruism condition, along with an additional sentence that reminded participants that the Gap did not have to donate any money to charity. (See the Supplemental Material for the full text used in each condition.) Participants then rated the company along the follow￾ing dimensions: moral, ethical, acceptable, altruistic, manipulative (reverse-coded), and selfish (reverse￾coded). They also rated how much they liked the com￾pany, trusted it, and were willing to purchase its products. All ratings were made on scales from 1 (low) to 9 (high). The order in which the items were presented was ran￾domized for each participant. Results A factor analysis with varimax rotation indicated that the dependent measures loaded onto three factors: liking (ratings of liking, trust, and willingness to purchase the company’s products; α = .83), morality (ratings of how moral, ethical, acceptable, and altruistic the company is; α = .87), and manipulativeness (ratings of how manipula￾tive and selfish the company was perceived to be; α = .63). For each of these measures, we conducted a con￾trast analysis testing the prediction that ratings would be significantly lower in the tainted-altruism condition than in the control, altruism, or counterfactual-information conditions (see Table 1). Liking. The contrast analysis comparing the tainted￾altruism condition with the other three conditions was significant, t(202) = 2.38, p = .018. As we predicted, rat￾ings were lower in the tainted-altruism condition (M = 5.16, SE = 0.29) than in the control (M = 5.74, SE = 0.23), altruism (M = 5.91, SE = 0.24), and counterfactual-infor￾mation conditions (M = 5.83, SE = 0.21), which did not differ from one another. Morality. For ratings of morality, the predicted contrast was also significant, t(202) = 2.99, p = .003. The pattern, however, was somewhat different, because ratings of morality were lower in both the control (M = 5.86, SE = 0.19) and tainted-altruism (M = 5.93, SE = 0.24) condi￾tions compared with the altruism (M = 7.12, SE = 0.19) and counterfactual-information (M = 6.91, SE = 0.19) conditions. Manipulativeness. Ratings of the Gap as selfish and manipulative were consistent with the predicted pattern, but the contrast analysis was only marginally significant, t(202) = 1.43, p = .15. Discussion Results from Experiment 4 indicated that merely high￾lighting the potential for self-interest significantly under￾mined evaluations of a company’s charitable efforts. This experiment is notable for a number of reasons. First, it was ecologically valid in that it assessed evaluations in a real-world context (an ongoing charitable program by a well-known company). Second, the effect itself was shown to be quite powerful in that a subtle manipulation (merely calling attention to the fact that the company also earned a profit, the other 50%) had an effect on evalua￾tions of the company as well as participants’ behavioral intentions (interest in purchasing their products). And finally, this experiment provides further support for coun￾terfactual thinking as an explanation of the effect, because the salience of the profitability information was the criti￾cal factor that differed across conditions. General Discussion The experiments reported here demonstrate the exis￾tence of a novel psychological phenomenon: the tainted￾altruism effect. In sum, we found that the presence of self-interest in the charitable domain was seen as taint￾ing, such that people judged efforts that realized both personal and charitable benefits to be less moral than analogous self-interested behaviors that produced no charitable benefit. This effect was observed across a Table 1. Mean Scores for Items Loading on Each Factor in Experiment 4 Condition Factor Liking Morality Manipulativenessa Control 5.74 (1.65) 5.86 (1.33) 5.74 (1.80) Altruism 5.91 (1.73) 7.12 (1.35) 5.49 (1.77) Tainted altruism 5.16 (2.05) 5.93 (1.73) 5.14 (1.79) Counterfactual information 5.83 (1.51) 6.91 (1.34) 5.46 (1.97) Note: Standard deviations are given in parentheses. a Higher numbers indicate more-favorable evaluations. Downloaded from pss.sagepub.com by Cai Xing on January 21, 2014