The promise of achieving “the greater good” has inspired numerous interventions designed to move society toward presumably desirable ends. Companies develop and market products to improve quality of life, organizations introduce policies to improve employees’ workplace experiences, and educators implement practices to improve learning outcomes. These interventions are frequently justified by claims that the benefits to many outweigh the potential harms to a few—a moral argument consistent with a utilitarian ethical framework.Footnote 1
A utilitarian approach to morality accepts inflicting harm onto some people if doing so increases the sum total of human happiness and well-being (e.g., Mill, 1861/2010; Singer, 1981, 2020). Guided by the classic tenets, Kahane et al. (2018) identified two elements that reflect the negative and positive features of utilitarian reasoning. The negative dimension—instrumental harm (colloquially known as collateral damage)—gives a moral agent permission to “instrumentally use, severely harm, or even kill innocent people to promote the greater good” (Kahane et al., 2018, p. 132). Impartial beneficence reflects the positive aspect of utilitarianism, requiring prioritization of the greater good above all else. In its purest form, this element demands people ignore personal ties, family loyalties, group memberships, special preferences, and emotional impulses that compromise impartiality and achieving this greater good (e.g., Hughes, 2017).
However, people frequently depart from such prescriptive moralities (Hughes, 2017; Kern & Chugh, 2009), seldom approaching the level of impartiality required to practice utilitarianism and accept sacrifices that may contribute to the greater good. Indeed, judgments about benefit and harm are highly subjective (Schein & Gray, 2018) and even malleable (Haslam, 2016; Rozin, 1999). This subjectivity, coupled with the difficulty of achieving consensus on what constitutes the greater good, undermines impartial calculi of costs and benefits requisite for upholding utilitarian principles.
The current investigation examined one factor that might compromise the impartial evaluation of social interventions: the gender of the person who experiences instrumental harm (Instrumental Harm). Based on prior research on perceptions of harm to women and men, we hypothesized that people asymmetrically support interventions inflicting collateral harm to men versus women. Such a bias violates the principle of impartial beneficence, potentially compromising the evidence-based advancement of men and women alike. As detailed below, our predictions are rooted in extant work on gender and moral decision-making and are extended to contexts depicting low-level harm.
Asymmetry in IH Acceptance as a Function of Gender
Perhaps the most compelling test of utilitarianism’s impartial beneficence tenet is the trolley problem (Foot, 1978). In the classic version of this moral dilemma, individuals must consider whether they would save a few people tethered to trolley tracks by derailing the trolley to instead crush a single individual (also tethered). Impartial beneficence dictates this single individual be sacrificed for the greater good; personal characteristics of this unfortunate individual—including their gender—should be irrelevant to decision-making. Yet, in line with the bounded nature of moral judgment (Kern & Chugh, 2009), the sacrificial individual’s gender sways observers’ judgments. Indeed, FeldmanHall et al. (2016) demonstrated that when responding to this trolley problem, people were more willing to sacrifice a man than a woman. These patterns have been replicated using virtual reality (Skulmowski et al., 2014).
Although these findings suggest people more readily accept physical harm to men than women in life-versus-death contexts, it remains unclear whether these results translate to lower-level, but nonetheless consequential forms of harm (e.g., psychological, health, educational, sexual). Extant evidence provides some indirect support to this possibility: people perceive men as less physically vulnerable and report lower desires to help them than women (Burnstein et al., 1994; Dijker, 2001, 2010). These patterns tentatively suggest people should more readily accept various forms of instrumental costs borne by men than by women.
One explanation for these patterns is gender stereotyping. Gender is a social category linked to numerous stereotypes relevant to moral decisions about harm. Throughout history (Bem, 1974; Hoffman & Borders, 2001) and still today, gender stereotypes conceptualize men as aggressive, self-sufficient, and risk-accepting, and women as gentle, tender, and yielding (Bhatia & Bhatia, 2021; Donnelly & Twenge, 2017; Eagly et al., 2020; Ellemers, 2018; Lewis & Lupyan, 2020). These assumptions have been further differentiated into the domains of agency and warmth, wherein men are more closely linked to agency and women to interpersonal warmth (i.e., communion; Eagly et al., 2020; Fiske et al., 1999, 2007).
Eagly and Mladinic (1989, 1994) termed this divergence the “women are wonderful” effect, whereby women are regarded more positively due to their presumed communality, but men more negatively due to their boldness and relative lack of warmth. People espouse these beliefs implicitly, such that they more strongly dislike men due to their automatic associations between masculinity and potency to inflict destruction (e.g., rage-driven violence; Rudman et al., 2001). The stereotypes making women appear “wonderful” (Eagly & Mladinic, 1989; Glick et al., 2004) and communal have also been linked to people’s stronger inclination to perceive women as victims (Reynolds et al., 2020). Therefore, individuals may similarly apply a reflexive heuristic that women should be protected from harm, including even from the IH resulting from interventions potentially advancing a greater social good. In contrast, these harms will be viewed as more acceptable if borne by men. Accordingly, we test the following hypotheses:
Hypothesis 1
People will be more willing to endorse interventions when IH befalls men as opposed to women.
Gender Differences in Partiality
Although we predict a greater tolerance for instrumental harm borne by men than by women, not all individuals will espouse such an asymmetry to equal degrees. A substantial body of evidence finds women exhibit stronger in-group biases favoring their own gender than do men (Rudman & Goodwin, 2004), suggesting greater acceptance of IH to men than women will be especially pronounced among women. Across countries, women express stronger hostility toward men and lower hostility toward women (Glick et al., 2004), suggesting if anyone should exhibit the hypothesized gender bias in instrumental harm acceptance, it should be women. Supporting this prediction, laboratory experiments find women redistribute payments to favor low-earning female (but not male) workers, whereas men showed no such gender bias (Cappelen et al., 2019). In the courtroom, women filing workplace discrimination claims were more likely to win compensation when their case was adjudicated by a female judge (Knepper, 2018). These patterns might be explained by a stronger bias in moral typecasting among women, whereby women more easily recognize other women as victims and men as perpetrators of harm (Reynolds et al., 2020). However, it remains unclear whether women show stronger gender biases in their tolerance of instrumental harms in low-level contexts. We predicted the following:
Hypothesis 2
Female participants will show a stronger asymmetry in their endorsement of IH, such that compared to male participants, female participants will show more approval of interventions inflicting instrumental harm onto men than onto other women.
Boundary Contexts: Stereotypically Female Contexts
The same gender stereotypes contributing to the justification of women’s protection from harm (e.g., higher communality) may also highlight possible boundary contexts to Hypothesis 1. That is, the stronger tendency to protect women might disappear in contexts whereby gender stereotypes dictate that women should bear the costs of social progress, such as by sacrificing on behalf of infants, children, the elderly, and the infirm. If throughout history, women’s communal roles enhanced the well-being of vulnerable individuals (e.g., children and the elderly; Eagly & Wood, 1999; Geary, 2010), interventions benefitting those vulnerable individuals, but inflicting costs onto women, might be equally or more strongly tolerated than those inflicting harm onto men, who less often filled such caregiving roles. Indeed, people perceive women as more responsible than men for protecting their children from harm (Barry et al., 2020), suggesting that traditional gender roles contribute to perceptions of sacrificial obligations. Formally, we predicted:
Hypothesis 3
The bias to reject IH to women (Hypothesis 1)will be neutralized in caregiving domains whereby historically, women have been expected to sacrifice more than men.
Present Research
We tested Hypotheses 1 and 2 across three complementary experimental studies (see Appendix Table 1 for an overview of all three studies). Study 1 provided the first test of our primary predictions in an organizational context. Study 2 utilized a broad array of contexts and interventions to test the generalizability of these patterns. Study 3 relied on stereotypical female caregiving contexts to provide a conservative test of the hypotheses and examine a potential boundary condition: domains wherein women have been traditionally expected to sacrifice (Hypothesis 3).
All studies were approved by the human subjects review boards from authors’ institutions. Participants provided their consent before they commenced the studies. Links to data and pre-registration documents (Study 1 and 3) are available at the end of this manuscript. We direct readers to supplementary online materials (SOM) for our materials, additional explanations, and exploratory analyses.