Do Relatives With Greater Reproductive Potential Get Help First?: A Test of the Inclusive Fitness Explanation of Kin Altruism

Abstract

According to inclusive fitness theory, people are more willing to help those they are genetically related to because relatives share a kin altruism gene and are able to pass it along. We tested this theory by examining the effect of reproductive potential on altruism. Participants read hypothetical scenarios and chose between cousins (Studies 1 and 2) and cousins and friends (Study 3) to help with mundane chores or a life-or-death rescue. In life-or-death situations, participants were more willing to help a cousin preparing to conceive rather than adopt a child (Study 1) and a cousin with high rather than low chance of reproducing (Studies 2 and 3). Patterns in the mundane condition were less consistent. Emotional closeness also contributed to helping intentions (Studies 1 and 2). By experimentally manipulating reproductive potential while controlling for genetic relatedness and emotional closeness, we provide a demonstration of the direct causal effects of reproductive potential on helping intentions, supporting the inclusive fitness explanation of kin altruism.

Genetic Relatedness and Family Helping Behavior

Perhaps nowhere is helping behavior in greater supply than in family relationships. Relatives are often present when we enter this world, and we usually want to have them near as we get ready to leave it. In the memories which populate those intervening years, family members often play an important role in major life events and can usually be counted on for comfort, love, and support. It is then not surprising that a great deal of evidence indicates that people are more willing to help relatives than nonrelatives and to endure greater hardships to deliver that help. This tendency toward kin altruism, or preferential helping behavior directed toward organisms to whom one is genetically related, is well-documented in animal species (e.g., Baglione, Canestrari, Marcos, & Ekman, 2003; Clutton-Brock, 2009; Silk, Brosnan, Henrich, Lambeth, & Shapiro, 2013).

In research with humans, the majority of studies on kin altruism employ hypothetical scenarios to assess respondents’ intentions to help. For example, Gesselman and Webster (2012) asked respondents to imagine a scenario in which someone insulted their sibling, their cousin, or a stranger, and the researchers observed that respondents were more likely to express aggressive intentions to someone insulting their sibling or cousin than a stranger. Similarly, Jonason, Izzo, and Webster (2007) observed that when respondents were asked to imagine helping someone find a romantic partner, they reported greater willingness to help a family member find a long-term romantic partner than to help a nonfamily member to do the same. Preferential helping intentions to family members have also been demonstrated in higher risk scenarios. For example, Kruger (2003) observed that respondents reported a greater willingness to help a sibling than a friend escape from a hypothetical situation involving life-threatening fire. Similarly, participants in Curry, Roberts, and Dunbar’s (2013) study reported a greater willingness to donate their kidney to a family member than to a friend. When assessing actual helping behavior, rather than willingness to help in hypothetical situations, Madsen and colleagues (2007) observed that participants spent significantly more time enduring physical pain for the benefit of their parent or sibling than of a nonrelative (Study 2) and someone with whom they were more distantly related (Studies 1–3).

Although evidence cited so far suggests a greater willingness to help family than nonfamily, not all family members are helped equally. Humans, like other animals, have evolved mechanisms for identifying genetic relatives and use that kin detection mechanism to make helping decisions (Lieberman, Tooby, & Cosmides, 2007; Mateo, 2015). Evidence linking genetic relatedness to helping has been found in research examining people’s willingness to help different family members. For example, Tifferet, Pollet, Bar, and Efrati (2016) asked their participants with a younger sibling to indicate the perceived similarity between themselves and this younger sibling and their routine investment in this sibling (e.g., care for well-being, amount of communication, time and money spent). The researchers observed that perceived similarity positively predicted investment toward their younger sibling and interpreted this observation as perceived similarity cueing a stronger sense of kinship. Although Tiffert et al.’s findings are consistent with the notion that genetic relatedness may foster greater helping behavior, the actual genetic similarity was not assessed in the study. Webster (2003) observed that shared genetic relatedness positively predicted a greater allocation of lottery money when he asked participants to allocate hypothetical lottery money to various blood relatives and to specify each relative’s relationship to them (e.g., “father’s sister” for “aunt”). Although genetic relatedness was assessed in Webster’s study, it was not experimentally manipulated. Hence, it may be the case that factors related to genetic relatednesses (e.g., frequency of interaction, emotional closeness) rather than genetic relatedness per se that contributed to the willingness to help (assuming that people generally interact more with, and feel emotionally closer to, genetically close relatives than genetically distant relative). Fitzgerald and Whitaker (2009) experimentally varied genetic relatedness by randomly assigning participants to indicate their willingness to help a sibling, a cousin, or a friend in a hypothetical violent (e.g., burning house) or nonviolent (e.g., chased by an attacker) life-threatening situation; participants in this study reported higher likelihood of helping in the sibling scenario in comparison to the cousin scenario, which did not differ from the friend scenario, regardless of the violence of the situations.

Preference to invest in genetically related family members has also been observed outside of the lab in real-world giving and wealth distribution situation. For example, Mysterud, Devron, and Slagsvold (2006) found that the amount of money Norwegian graduate students spent on presents for family members during the preceding Christmas was significantly associated with their degree of relatedness to the recipient; the closer a graduate student was to a family member genetically, the more money they would spend on that person’s Christmas gift. These preferences are more clearly demonstrated in research on bequests. In their analysis of 1,000 probated wills from Vancouver, British Columbia, for example, Smith, Kish, and Crawford (1987) found that the deceased typically bequeath more to close relatives than distant relatives. In another study, Judge and Hrdy (1992) examined wills from Sacramento, CA, and found that women, but not men, who are survived by their spouse leave a greater share of their wealth to their children than to their spouse; and the researchers attributed this sex difference to women being more certain of their children’s genetic relatedness to themselves than are men. Bossong (2000) later replicated these findings in the laboratory demonstrating that women, but not men, who imagined being survived by their spouse reported greater willingness to leave their wealth to their children than to their spouse. Thus, when it comes to making helping decisions between family members, the degree of genetic similarity between the helper and the recipient appears to play an important role.

In summary, existing research evidence suggests that humans are more willing to help family than nonfamily members, and when choosing between family members, preference is given to family members with closer genetic relatedness.

Emotional Closeness and Family Helping Behavior

Although genetic relatedness appears to play a role in family helping decisions, there is more to being a family than biological commonalities. There is some evidence pointing to emotional closeness as a factor that plays a role in family helping. Emotional closeness refers to one’s concern for, trust in, and enjoyment of their relationship with another person (Korchmaros & Kenny, 2006). Having a need to belong with other people and feel emotional closeness with them would have given our ancestors a survival advantage over others without this need, and so emotional closeness may itself have evolutionary origins (Miller, 2015). When people feel a sense of connection to a group, regardless of their genetic relationship, they are more inclined to give help (Pavey, Greitemeyer, & Sparks, 2011). For example, Rachlin and Jones (2008) asked participants to imagine dividing a pot of money between themselves and different members of their social network and found that genetic relatives, the people whom participants are most willing to help, also tend to be the people that participants report feeling closest to within their social networks. In another example demonstrating the effect of emotional closeness on family helping behavior, Korchmaros and Kenny (2001) asked participants to pick, in round-robin style, relatives varying in emotional closeness to hypothetically save from a burning building; they found that the association between genetic relatedness and help intentions was partially mediated by emotional closeness. Furthermore, emotional closeness along with neediness and familial obligation accounted for over 60% of the variance in helping intention attributed to genetic relatedness (Korchmaros & Kenny, 2006), although Curry et al. (2013) found that genetic relatedness still predicted help intentions for family members when controlling for emotional closeness.

In summary, previous research has shown that people tend to help family over friends and genetically close family members over genetically distant family members, and the association between genetic relatedness and helping intention has been at least partially attributed to emotional closeness.

Inclusive Fitness, Kin Altruism, and Reproductive Potential

Inclusive fitness theory outlines an evolutionary account for the origins of family helping behavior or kin altruism (Hamilton, 1964). For any gene to increase in frequency in a population, it must either do so directly, by replicating itself, or indirectly, by promoting the survival and reproduction of other organisms who also have a copy of the gene. This second option is most likely to occur when the cost of promoting the other organism’s survival (c) is less than the product of the degree of relatedness between the two organisms (r) and the potential benefit of helping to the reproduction of the gene (B). Thus, according to Hamilton’s rule, kin altruism will occur when rB > c. According to the theory of inclusive fitness, family helping behavior developed as a way of promoting the indirect replication of a kin altruism gene through the survival and reproduction of related others (who are presumed to also be carrying the gene). As a consequence, humans have evolved to make choices about helping that are sensitive to the parameters mentioned above (the cost to helper, relatedness of the recipient, and reproductive benefit); and in the right set of circumstances, these choices can have a powerful effect on help intentions (Korchmaros & Kenny, 2006).

The principle of inclusive fitness may seem fairly innocuous, but it has at least one somewhat disturbing implication. If people are motivated to help their family members based on the probability that they will help reproduce the kin altruism gene (B) and some relatives have a very low probability of being able to reproduce, then people should be less willing to help those family members than others with a greater chance of producing biological offspring. There are some preliminary findings suggesting that this may be the case. For example, in research by Fitzgerald and Colarelli (2009), participants indicated their intentions to help fictional relatives with a variety of circumstances that limit their reproductive potential such as functional limitations (e.g., mental illness), physical limitations (i.e., malnutrition), and sexual limitations (e.g., homosexuality). Participants reported less willing to provide help to people who had one of these limitations but only when the cost of helping was high (e.g., helping someone escape a burning house). When the cost of helping was low (e.g., helping someone pick up a few items from the store), the limitations had little or no effect on help intentions. This finding makes evolutionary sense: If people are motivated to help relatives based on their likelihood of passing on genes, then recipients’ reproductive potential should have a greater influence on helping decisions in situations where their ability to reproduce is threatened than when it is not (Korchmaros & Kenny, 2006). Furthermore, according to Hamilton’s rule, kin altruism is most likely to occur when the cost of helping (c) is less than the product of the reproductive benefit to helping the organism (B) and the degree of relatedness (r). When the cost of helping is low, helping may still occur when the degree of relatedness or reproductive benefit are also low (as long as the product of the two is still higher than the cost). As the cost of helping increases, however, it becomes more important that the product of reproductive benefit and relatedness also increase to compensate. Thus, a family members’ potential to reproduce should have the greatest influence on helping behavior when the decision has direct consequences for the recipients’ potential to reproduce and the cost of helping is high. Indeed, Burnstein, Crandall, and Kitayama (1994) randomly assigned participants to imagine providing help to family members in either mundane or life-or-death situations, and found that although participants gave priority to family members based on who they perceived as needing help the most (such as the poor, the sick, the very old, or the very young) in mundane situations; in life-or-death situations, they gave priority to those who were most likely to successfully produce genetic offspring (such as the young, the healthy, the wealthy, and the premenopausal). Similarly, Stewart-Williams (2007) found that people were more willing to hypothetically donate a kidney to a sibling than to a friend but were equally likely to help friends and siblings who were ill and were more willing to give emotional support to friends than siblings. These patterns were again observed in Chuang and Wu’s (2017) study in which they asked participants to choose one family member from a triad to give help in either mundane or life-or-death hypothetical scenarios. In the hypothetical life-or-death scenarios, participants were most willing to provide help to relatives with the highest reproductive value or the greatest expectation of producing future offspring based on their relative age, even if they had indicated a preference to help another member of the triad in the hypothetical mundane help condition.

In summary, previous psychological research on kin altruism in support of inclusive fitness shows that people are more willing to help relatives than nonrelatives and genetically close relatives than genetically distant relatives. Additionally, when the cost of helping is high, relatives with distinct reproductive limitations or characteristics that make them less likely to reproduce are less likely to be helped than relatives without these limitations or characteristics. The logic of these studies is built on the assumption that people are less likely to help relatives who are less likely to reproduce and therefore pass on the kin altruism gene, and research findings so far are consistent with that assumption. To our knowledge, however, the precise effect of explicit reproductive potential on a person’s likelihood of receiving kin help has not yet been tested. To rigorously examine whether people’s perception of a family member’s reproductive potential can influence their intention to help this family member, experimental research in which the reproductive potential of a family member is manipulated is necessary.

Current Research

The present research uses an experimental design to directly and explicitly test the effect of a kin’s reproductive potential on their likelihood of receiving help. Based on the logic of inclusive fitness theory, we expect to find that the potential of reproducing (and passing on the kin altruism gene) will have a causal effect on kin help intention, particularly when the cost of helping is high. Although previous research suggests that this is the case by using proxies for reproductive potential (e.g., age, health, wealth, reproductive value, and the presence or absence of reproductive limitations), manipulating explicit reproductive potential information is essential for rigorously addressing the issue of reproductivity.

Additionally, the current research seeks to disentangle emotional closeness from family relationships to clearly examine the causal effect of emotional closeness on intentions to help while holding family relationship constant. Furthermore, the current research makes a unique contribution to the understanding of whether a kin’s reproductive potential drives helping for the kin independent of emotional closeness to the kin. By conducting the study with these parameters, we are able to strictly test the applicability of inclusive fitness theory as an explanation for helping intentions, as well as the unique effect of emotional closeness apart from genetic relatedness, and the unique effect of the kin’s reproductive potential apart from emotional closeness.

We used a series of decision-making tasks in a fully experimental design to examine the causal effects of reproductive potential, emotional closeness, and helping cost on intentions to engage in kin altruism. Because genetic relatedness and emotional closeness can covary with familial relationship, we held the relationship of the kin constant for Studies 1 and 2 (always a female cousin ¹ ) and manipulated it for Study 3 (either a female cousin or a female friend). We hypothesized that cousins more likely to reproduce would receive help earlier, particularly if the cost of helping was high (in a life-or-death situation) rather than low (in a mundane situation). We also manipulated emotional closeness in Studies 1 and 2 and hypothesized that relatives who were high in emotional closeness would receive help earlier than relatives who were low in emotional closeness, regardless of their reproductive potential or the type of help required.

In Study 1, we manipulated female cousins’ potential to pass on genes by presenting them as getting ready to bear a biological child or to adopt a child as well as emotional closeness. In Study 2, we manipulated reproductive potential, the capability to reproduce offspring, by varying each cousin’s ability to conceive a child as well as emotional closeness. In Study 3, holding emotional closeness constant, we compared intentions to help a cousin high in reproductive potential with intentions to help three other targets. In all three studies, participants were randomly assigned to make helping decisions regarding the targets in either a mundane or a life-or-death situation. All studies were approved by the University of New Brunswick Research Ethics Board.

Method

Participants

Participants were 203 residents of the United States and of the legal age of majority. Data on gender, ethnic background, age, and religion were not available for 15 participants who did not complete the demographic questionnaire due to a programing error. One hundred and eighty-eight participants completed the demographic questionnaire (103 men and 85 women; mean age = 34.77 years, SD = 11.80, range = 19–70). Of those, 68.7% were White, 6.2% Asian, 6.6% Hispanic/Latino, 5.7% African American, 1.9% American Indian (10.9% did not provide their ethnicity); 44.5% were Christian, 3.8% Buddhist, 1.4% Jewish, 1.4% Muslim, 0.5% Baha’i, 0.5% Hindu, 0.5% Zorostrian, and 36.0% indicated “other” with the majority of these identifying as atheist or not religious (11.4% did not report their religion). Among our participants, 86.3% reported having at least one cousin.

Procedure

Participants were recruited from the population of registered users on the Amazon M-Turk website and received US$3 in credits on Amazon.com for completing the study. M-Turk allows Amazon members (“workers”) to perform human intelligence tasks in exchange for monetary compensation. Data collected from M-Turk meet acceptable psychometric standards in terms of internal consistency and test–retest reliability and are more diverse than samples collected from introductory psychology course pools (Buhrmester, Kwang, & Gosling, 2011). The study was listed on the M-Turk site among thousands of other human intelligence tasks available for participation. Participants completed an informed consent form and were then directed to the scenario task and a demographic questionnaire.

Help type manipulation

Participants were randomly assigned to either the mundane helping condition or the life-or-death helping condition. Participants were asked to imagine a scenario where four female cousins either required their help to complete important errands (mundane help condition) or were asleep in a burning building and required their help to escape (life-or-death help condition). Participants were told that if they did not help, the cousins would not be able to complete their errands (in the mundane condition) or would die in the burning building (life-or-death condition), only one cousin could be helped at a time, and those helped later would be less likely to complete their errands (mundane condition) or live (life-or-death condition). Thus, only the cousin helped first was guaranteed to be helped (in the mundane condition) or saved (in the life-or-death condition). The four cousins were presented in a randomized order.

Reproductive potential manipulation

To manipulate the potential of passing on kin altruism genes, we varied the cousins’ childbearing status. Two cousins in each scenario were described as “preparing to conceive a biological child”; the other two were described as “preparing to adopt a child.”

Emotional closeness manipulation

Two cousins in each scenario were described as someone with whom the participant is “emotionally close”; the other two were described as someone with whom the participant is “not emotionally close.”

Participants’ task was to indicate the order in which they would help the four cousins described in the scenario, to which they were randomly assigned. Participants then completed the demographics questionnaire and were directed to a debriefing form.

Results and Discussion

We conducted a 2 (help type: mundane, life-or-death) × 2 (reproductive potential: conceive, adopt) × 2 (emotional closeness: high, low) mixed analysis of variance (ANOVA) on the order of helping, ² with reproductive potential and emotional closeness as within-subjects variables and help type as a between-subjects variable. ³ See Table 1 for the descriptive statistics.

Table 1.

Mean and Median Helping Order by Target and Condition in Study 1.

Target	Type of Help
	Mundane		Life-or-Death
	Mean (SD)	Median	Mean (SD)	Median
High closeness, preparing to conceive	1.70 (0.09)	1.0	1.45 (0.75)	1.0
Low closeness, preparing to conceive	3.18 (0.08)	3.0	2.98 (0.84)	3.0
High closeness, preparing to adopt	1.80 (0.08)	2.0	2.03 (0.76)	2.0
Low closeness, preparing to adopt	3.32 (0.09)	4.0	3.54 (0.74)	4.0

Note. SDs are given in parentheses.

A significant main effect of reproductive potential revealed that across the two types of helping scenarios, cousins preparing to conceive a child were generally helped earlier (M = 2.33, SE = .04) than cousins preparing to adopt a child (M = 2.68, SE = .04), F(1, 201) = 17.13, p < .001, ${\eta }_{\text{p}}^2$ = .08. However, we had predicted that the effect of reproductive potential would be stronger in the life-and-death condition than in the mundane condition. As predicted, the Reproductive Potential × Help Type interaction was significant, F(1, 201) = 7.19, p < .01, ${\eta }_{\text{p}}^2$ = .04. In the life-or-death condition, participants helped cousins preparing to conceive a child (M = 2.22, SE = .06) earlier than cousins preparing to adopt a child (M = 2.78, SE = .06), F(1, 103) = 25.44, p < .001, ${\eta }_{\text{p}}^2$ = .2. However, in the mundane condition, cousins who were preparing to conceive a child (M = 2.44, SE = .06) were helped no earlier than cousins preparing to adopt a child (M = 2.56, SE = .06), F(1, 98) = 0.97, p > .05, ${\eta }_{\text{p}}^2$ = .01 (see Figure 1). The cousins preparing to conceive stand to transmit a partial copy of a participant’s genes, but the cousins preparing to adopt do not. Consistent with inclusive fitness theory, it seems that people are willing to improve their overall fitness by being more willing to save the lives of those who can transmit shared genetic material.

Figure 1.

Mean intention to help as a function of reproductive potential and help type in Study 1.

Not surprisingly, we also found a main effect for emotional closeness, with cousins high in emotional closeness (M = 1.74, SE = .04) receiving help significantly earlier than cousins low in emotional closeness (M = 3.25, SE = .04), F(1, 201) = 474.01, p < .001, ${\eta }_{\text{p}}^2$ = .70. No other effects were statistically significant (all p ≥ .50). As predicted, across the two help conditions, emotional closeness led to helping earlier. Also as predicted, the main effect of emotional closeness was not qualified by the type of help required, nor the reproductive potential of the target, but had an overall effect above and beyond these two variables. Previous research showed that the relationship between kinship and help intentions is at least partially mediated by perceived emotional closeness (Korchmaros & Kenny, 2001, 2006). By experimentally manipulating emotional closeness, we demonstrated a causal relationship between emotional closeness and helping intentions when holding kinship constant.

Introduction

In Study 1, some cousins were described as “preparing to conceive a biological child,” but no mention was made of whether they were actually capable of doing so. In Study 2, therefore, we made a small but important adjustment to our design and manipulated cousins’ actual reproductive potential—their chance to reproduce—rather than the intention to adopt or conceive. A second addition to the design of Study 2 was the inclusion of an open-ended question to examine participants’ rationale behind their decisions. Because one’s reproductive potential has a direct effect on their ability to pass on the kin altruism gene and thus their reproductive benefit (B), we hypothesized that a cousin’s higher reproductive potential would lead to being helped earlier in the life-or-death condition, but not necessarily in the mundane condition. Additionally, we hypothesized that emotional closeness would again have an unqualified effect on the order in which the cousins were helped.

Method

Participants

Two hundred and seven participants were recruited for the second study (100 women, 102 men, 1 transgender person, and 4 declined to give their gender; mean age = 34.42, SD = 11.67, range = 18–70, and 3 participants did not report their age). All participants were residents of the United States and of the legal age of majority. Ethnic backgrounds were 80.2% White, 7.2% Asian, 5.8% African American, 3.4% Hispanic/Latino, and 1.4% American Indian or Alaska Native; 2.0% declined to give their ethnic background. Participants were 36.7% Christian, 2.4% Hindu, 1.0% Buddhist, 2.4% Jewish, 1.0% Muslim, and 0.5% Baha’i. Of the remaining sample, 51.2% reported “other” which was again predominantly atheist or agnostic, 2.9% reported “no religion,” and 1.9% gave no information about their religion. Among our participants, 92.3% reported having at least one cousin.

Procedure

As in Study 1, participants were recruited on the Amazon M-Turk website and received US$3 in credits on Amazon.com for completing the study. After completing an informed consent form, they were directed to the scenario. The type of help and emotional closeness manipulations were identical to those of Study 1, as was the task. The four cousins were again presented in a randomized order.

Reproductive potential manipulation

Two cousins were described as having “a 95% chance to conceive a child”; the other two were described as having “a 95% chance to not conceive a child.” Thus, two cousins had a high potential to reproduce and thus pass on the kin altruism genes and two cousins had a low potential of doing so.

After completing the ordering task, participants answered an open-ended question for the cousin helped first only, asking them to “tell us a little about what you were thinking when deciding to help this cousin.” Participants then completed a demographics questionnaire and were directed to a debriefing form.

Result and Discussion

We had hypothesized that high reproductive potential would lead to getting help earlier in the life-or-death scenario but not in the mundane scenario. To test this hypothesis, we conducted a 2 (help type: mundane, life-or-death) × 2 (emotional closeness: high, low) × 2 (reproductive potential: high, low) mixed ANOVA on the ordering data, with emotional closeness and reproductive potential as within-subjects variables and help type as a between-subjects variable. ⁴ See Table 2 for the descriptive statistics. As predicted, and consistent with the findings of Study 1, there was a significant interaction effect between a cousin’s reproductive potential and type of help, F(1, 203) = 68.69, p < .001, ${\eta }_{\text{p}}^2$ = .25. Replicating Study 1, in the life-or-death scenario, cousins high in reproductive potential (M = 2.11, SE = .04) were helped significantly earlier than cousins low in reproductive potential (M = 2.89, SE = .04), F(1, 98) = 99.29, p < .001, ${\eta }_{\text{p}}^2$ = .50. In the mundane scenario, however, cousins high in reproductive potential (M = 2.67, SE = .05) were helped significantly later than cousins low in reproductive potential (M = 2.33, SE = .05), F(1, 105) = 9.89, p < .01, ${\eta }_{\text{p}}^2$ = .09 (see Figure 2). In Study 1, there was no effect of reproductive potential in the mundane condition; however, the effect is significant in Study 2, with the cousins who are most likely to reproduce being the last to receive help. According to inclusive fitness theory, when the cost of helping is high, the ability to reproduce and pass on genes has a strong influence on helping intentions. As the cost of help decreases, however, the ability to pass on genes becomes less important, allowing other factors to have a greater relative influence on helping intentions. This might explain the finding from Burnstein et al. (1994) that, in mundane situations, participants gave priority to family members based on perceived need, not their ability to pass on genes. It might also explain why, even though people are more willing to give a kidney to a family member than to a friend, the latter is much more likely to be the recipient of emotional support than the former (Stewart-Williams, 2007). Similarly, in our studies, in the life-or-death conditions, the cost of helping was high and so the ability to pass on genes had a strong influence on helping intentions and hence cousins who were likely to conceive were helped earlier than cousins unlikely to conceive. In the mundane conditions, however, the cost of helping was low and so the ability to pass on genes had less importance. We speculate that participants might have felt sorry for the cousins with a low chance of conceiving and hence were willing to help them first with a mundane task. That is, the decisions of participants in the mundane condition might have been influenced by sympathy—a common motivator of helping behavior (Dovidio, Piliavin, Gaertner, Schroeder, & Clark, 1991).

Table 2.

Mean and Median Helping Order by Target and Condition in Study 2.

Target	Type of Help
	Mundane		Life-or-Death
	Mean (SD)	Median	Mean (SD)	Median
High closeness, likely to conceive	1.92 (0.81)	2.0	1.34 (0.77)	1.0
Low closeness, likely to conceive	3.42 (0.86)	4.0	2.88 (0.64)	3.0
High closeness, unlikely to conceive	1.62 (0.87)	1.0	2.13 (0.68)	2.0
Low closeness, unlikely to conceive	3.04 (0.80)	3.0	3.64 (0.79)	4.0

Note. SDs are given in parentheses.

Figure 2.

Mean intention to help as a function of reproductive potential and help type in Study 2.

To assess whether sympathy played a role in helping the low reproductive potential cousin in the mundane condition, we had two independent coders analyze the content of open-ended responses for mentions of feelings of sympathy, empathy, or pity (specifically, statements that mentioned feeling bad for, feeling pity for, feeling sorry for, empathy and sympathy or sympathize); agreement between the two coders was 100%. We found that 19 of 106 participants in the mundane condition mentioned sympathy for cousins low in reproductive potential (e.g., “The fact that there is a 95% chance she is unable to have a child makes me feel sympathetic toward her and want to help her”), but only 1 participant of 99 in the life-or-death condition mentioned sympathy for cousins low in reproductive potential. None of the participants reported feeling sympathy, empathy, or pity for cousins high in reproductive potential. It appears, then, that sympathy had motivated participants to help the cousin with low reproductive potential when it comes to a mundane task, but not when it comes to saving their life.

We had also hypothesized that emotional closeness would again have an unqualified effect on one’s likelihood of receiving help. Consistent with the findings in Study 1, there was a main effect of emotional closeness, with cousins high in emotional closeness (M = 1.75, SE = .04) helped earlier than cousins low in emotional closeness (M = 3.25, SE = .04), F(1, 203) = 385.70, p < .001, ${\eta }_{\text{p}}^2$ = .66. Again, this main effect was not qualified by the reproductive potential of the targets, nor the type of help required, demonstrating the unqualified causal effect of emotional closeness on family helping intentions. No other effects were significant (ps > .50).

In summary, the results from Study 2 are consistent with the results from Study 1 in suggesting that both emotional closeness and reproductive potential influence helping. In general, people give preference to helping cousins to whom they feel emotionally close. Also, when the cost of help is high, cousins with a higher chance of having a genetically related offspring receive help earlier. When help cost is low, however, the effect is reversed: Cousins with a low chance of having genetically related offspring were more likely to be helped first, likely due to participants’ sympathy toward them. ⁵

Introduction

In Studies 1 and 2, we held genetic relatedness constant so that it would be possible to demonstrate the causal effect of emotional closeness on family helping intentions. Because genetic relatedness has been held constant so far, we cannot know if the effect of reproductive potential demonstrated in Studies 1 and 2 was motivated by a desire to pass on one’s kin altruism gene or simply by participants’ inclination to save the lives of women who can become pregnant—regardless of whether or not they are relatives. In Study 3, we manipulated both reproductive potential and relatedness. In line with the premises of the inclusive fitness theory, we predicted that reproductive potential should affect helping only relatives, not friends, in the life-or-death condition. In this study, participants selected the order in which they would help four targets: a high-fecundity cousin, a low-fecundity cousin, a high-fecundity friend, and a low-fecundity friend, all equal in emotional closeness. Because the high-fecundity cousin is the only target with a high likelihood of passing on one’s kin altruism genes and thus a high reproductive benefit (B) to compensate for the increase in cost (c), we hypothesized the high-fecundity cousin in life-or-death scenario would be helped earlier than any of the other targets and the other targets would not differ in order of receiving help.

Method

Participants

There were 191 participants (96 women, 93 men, 1 transgender person, 1 unspecified; mean age = 34.98 years, SD = 10.98, range = 18–70), who were residents of the United States and of the legal age of majority. Ethnic backgrounds of the participants were 75.4% White, 8.4% African American, 7.9% Asian, 7.3% Hispanic/Latino, and 1.0% American Indian or Alaska Native. Participants were 45.6% Christian, 3.1% Buddhist, 0.5% Jewish, 0.5% Zoroastrian, and 0.5% Hindu; of the remaining participants, 46.1% indicated “no religion” while the 3.7% who chose “other” mostly reported atheist or agnostic. Among our participants, 93.7% reported having at least one cousin.

Procedure

As in previous studies, participants were recruited from the Amazon M-Turk website. They received a US$1 credit on Amazon.com. Participants completed an informed consent form and then were randomly assigned to imagine either a life-or-death or mundane helping scenario. The type of help manipulation was identical to the previous two studies. Participants were presented with four targets and asked to list the order of helping. Emotional closeness was held constant by describing all targets as someone “you are close to emotionally.” Participants then completed a demographics questionnaire and were directed to a debriefing form.

Target manipulation

As in Study 2, to manipulate reproductive potential, two targets were described as having “a 95% chance of being able to conceive a child,” while the other two targets were described as having “a 95% chance of not being able to conceive a child.” To manipulate relationship, two targets were presented as a “female cousin” and two targets were presented as a “female friend.” Finally, to manipulate helping condition, half of the participants had to make helping decisions in a life-or-death scenario and half in a mundane scenario. All targets were described as a person “you are very close to emotionally.” Targets were presented in a randomized order.

Results and Discussion

We conducted a 2 (reproductive potential: high or low) × 2 (relationship: cousin or friend) × 2 (condition: life-or-death or mundane) mixed ANOVA. ⁶ See Table 3 for descriptive statistics. We found a significant main effect of relationship, F(1, 189) = 20.36, p < .001, ${\eta }_{\text{p}}^2$ = .10, with cousins (M = 2.25, SE = .06) being helped earlier than friends (M = 2.75, SE = .06). As in previous studies, we also found a significant effect of reproductive potential, F(1, 189) = 14.56, p < .05, ${\eta }_{\text{p}}^2$ = .07, with targets likely to conceive (M = 2.33, SE = .05) being helped earlier than targets unlikely to conceive (M = 2.67, SE = .05). This was qualified, however, by a significant Reproductive Potential × Condition interaction, F(1, 189) = 24.58, p < .001, ${\eta }_{\text{p}}^2$ = .12, showing that although the effect of reproductive potential was not significant in the mundane condition, F(1, 95) = 0.54, p > .05, ${\eta }_{\text{p}}^2$ = .01, it was significant in the life-or-death condition, F(1, 94) = 48.77, p < .001, ${\eta }_{\text{p}}^2$ = .342, with targets likely to conceive (M = 2.10, SE = .06) helped earlier than targets not likely to conceive (M = 2.90, SE = .06).

Table 3.

Mean and Median Helping Order by Target and Condition in Study 3.

Target	Type of Help
	Mundane		Life-or-Death
	Mean (SD)	Median	Mean (SD)	Median
Cousin, likely to conceive	2.32 (1.13)	2.0	1.80 (1.14)	1.0
Friend, likely to conceive	2.78 (1.12)	3.0	2.40 (0.94)	3.0
Cousin, unlikely to conceive	2.26 (1.03)	2.0	2.60 (0.83)	2.0
Friend, unlikely to conceive	2.64 (1.13)	3.0	3.20 (1.09)	4.0

Note. SDs are given in parentheses.

We specifically predicted that the high reproductive potential cousin in the life-or-death condition would be helped earlier than any other targets, which would not differ from each other. This means that we expected one cell to be different from the seven other cells that do not differ from one another. Although our Reproductive Potential × Relationship × Condition interaction did not reach significance (p > .05), a three-way ANOVA is not the proper way of assessing our prediction because the F-ratio compares variance in a model due to the manipulation against variance in a model due to chance. A large F-ratio requires the manipulation to be responsible for more differences between individual groups than would be expected at random (Field, 2009). Hence, the distinct variance of one cell (the one with the high-fecundity cousin in the life-or-death scenario) would be likely drowned out by the similarities between the other seven cells. In an attempt to adjust for this, we conducted a 4 (target: high-fecundity cousin, high-fecundity friend, low-fecundity cousin, low-fecundity friend) × 2 (condition: life-or-death, mundane) ANOVA with planned comparisons to assess the prediction that only in the life-or-death condition the high-fecundity cousin would be helped earlier than the other targets, but not in the mundane condition. These preliminary results demonstrated an overall significant interaction, F(3, 187) = 8.12, p < .001, ${\eta }_{\text{p}}^2$ = .12. The simple effect of target was not significant in the mundane condition, F(3, 93) = 2.48, p > .05, ${\eta }_{\text{p}}^2$ = .07, but it was significant in the life-or-death condition, F(3, 92) = 17.27, p < .001, ${\eta }_{\text{p}}^2$ = .36. In the life-or-death condition, planned comparisons revealed a statistically significant difference in intention to help from cousins who were likely to conceive (M = 1.80, SD = 1.14) to the average of cousins unlikely to conceive (M = 2.60, SD = 0.83), friends likely to conceive (M = 2.4, SD = 0.94), and friends unlikely to conceive (M = 3.20, SD = 1.09) with a mean difference of −.93, 95% confidence interval (CI) [−1.23, −0.63], p < .001. In the mundane condition, however, there was not a significant difference, with cousins likely to conceive (M = 2.32, SD = 1.13) no more likely than the average of cousins unlikely to conceive (M = 2.26, SD = 1.03), friends likely to conceive (M = 2.78, SD = 1.12), and friends unlikely to conceive (M = 2.64, SD = 1.13) to receive help, with a mean difference of −.24, 95%, CI [0.13, −0.54], p = .13 (see Figures 3 and 4). ⁷

Figure 3.

Mean intention to help as a function of reproductive potential and relatedness to the recipient in the life-or-death condition.

Figure 4.

Mean intention to help as a function of reproductive potential and relatedness to the recipient in the mundane condition.

General Discussion

Inclusive fitness theory holds that people help family members because they are likely to pass on a shared kin altruism gene. Previous psychological research on kin altruism in support of this theory shows that people with limitations or characteristics that may make it less likely that they will reproduce are also less likely to receive help from family members than people without those limitations or characteristics (Curry, Roberts, & Dunbar, 2013; Fitzgerald & Colarelli, 2009). No research to date, however, has directly and experimentally tested the effect of targets’ reproductive potential on their likelihood of receiving help.

The results of our studies demonstrate that, when helping affects relatives’ chances of survival and the cost of helping is high, people give priority to the relatives with the greatest likelihood of passing on the kin altruism gene: a cousin who is preparing to conceive rather than adopt a child (Study 1) and cousins with a high likelihood of conceiving rather than a low likelihood (Studies 2 and 3). When the situation does not directly affect survival and the cost of helping is low, a relative’s likelihood of passing on the shared kin altruism gene has less of an effect on helping. Across three studies, we found that cousins who were less likely to reproduce were helped later with mundane tasks (Study 1), earlier (Study 2), or at the same time (in Study 3) as cousins who were more likely to reproduce. When the cost of helping is high, it is important that the reproductive benefit of helping is also high, but when the cost is low, having a high reproductive benefit becomes less important. In these situations, reasons for helping associated with social norms or other practical and social considerations, such as reciprocity, proximity, or sympathy (as was found in Study 2), may have more of an effect. Hence, when help does not affect a target’s likelihood of survival, the effect of reproductive potential on kin altruism becomes less certain.

The role of emotional closeness in family helping intentions has been well-documented and is demonstrated in this research as well. In both Studies 1 and 2, emotional closeness led to receiving help earlier regardless of target reproductive potential or type of help required. Although previous research on helping intentions and behavior in support of inclusive fitness theory has shown preference for helping family over friends and close family members over distant family members (Curry et al., 2013; Kruger, 2003; Madsen et al., 2007; Smith, Kish, & Crawford, 1987; Webster, 2003), it is unclear if this helping is due to emotional closeness or familial relationship. In the current research, we disentangle emotional closeness from family relationships, by manipulating the former while holding the latter constant. We thereby contribute to the literature by showing the causal effect of emotional closeness on helping intentions in family relationships. Thus, future research that assumes that the effect of emotional closeness on kin altruism is bound inexorably to familial closeness should take note of these findings.

More importantly, however, by manipulating reproductive potential while controlling for genetic relatedness and emotional closeness, we demonstrate that reproductive potential has a causal effect on helping intentions when helping directly affects the survival of the person being helped, strongly supporting the inclusive fitness explanation of kin altruism. Previous studies have used proxies for reproductive potential, such as age, health, wealth, reproductive value, and the presence or absence of reproductive limitations to infer the effect of reproductive potential, but none have assessed the effect directly. By experimentally manipulating this variable directly, the present research adds to the literature by demonstrating its causal effect on the intention to provide help, corroborating the assumptions on which these previous studies and inclusive fitness theory itself are founded. To our knowledge, this is the most direct and precise incorporation of an inclusive fitness account of kin altruism into a fully experimental design to date.

There are several limitations to what can be said about family helping intentions based solely on the results of this research. Although our data suggest that reproductive potential plays a role in how people decide which family members to help, it is beyond the scope of that data to determine how that process takes place. One possibility is that the process is largely unconscious—knowledge about the reproductive potential of family members is accumulated over time and can affect the way we treat relatives, even if we seldom reflect on that information consciously. In our study, the only information participants had about the targets was their degree of reproductive potential and level of emotional closeness, and so participants were restricted to making decisions based only on that information. Another possibility is that people begin with the assumption that their family members are willing and able to reproduce, and that assumption is qualified by evidence to the contrary. In our study, then, the knowledge that family members were unlikely to reproduce gave them a special status that made them receive help later in the life-or-death condition but sooner in the everyday condition. Future research should look for ways to follow up on the comparative plausibility of each explanation.

Although our experiments allowed us to make causal inferences, hypothetical scenarios set limitations on the external validity of the research. The cousins that participants were asked to help were imagined, unidimensional characters, and so there might have been less information processing required in making helping decisions for those characters than there would be for actual family members. There is also the possibility that our participants might think and behave differently if they were facing down a real burning building rather than a hypothetical one. The results may also have been different if we had only used participants who had cousins and used the actual names of those cousins in the scenarios. Thus, although the use of real burning buildings with real cousins in them will be clearly frowned upon by most ethics committees, clever and safe designs should test these effects in less tightly controlled environments and should seek to replicate our results in more real-world contexts to examine the generalizability of the phenomenon.

Additionally, whereas this study focused exclusively on female cousins (with the addition of female friends in Study 3), investigating the relationship between male reproductive potential and intentions to help noncousin family members (e.g., siblings, aunts) would serve to further extend this research. Finally, most of our participants were from a modern Western society, which is characterized by an emphasis on independence and self-actualization (Markus & Kitayama, 1991). How might inclusive fitness look in a collectivist culture, where social groups are given greater emphasis, or in more traditional cultures that may place different values on childbearing? Future research should investigate the degree to which these observed effects hold across cultures.

Conclusion

Taken together, the present research demonstrates that people prefer to help family members who are more likely to reproduce biologically. This set of responses may be due to evolved decision-making rules which are sensitive to the parameters outlined by inclusive fitness theory, however, there are two important caveats to these findings. The first caveat is that it is only when one’s life is in danger that the ability to reproduce has a clear bearing on who receives help. In the majority of situations, a person’s ability to bear children may have no relation to or even increase their likelihood of receiving help. The second caveat is that the effect of emotional closeness still has a clear, significant effect on helping intentions. It is comforting, then, to know that even if our helping decisions are clearly and directly affected by the survival strategies of our ancient ancestors, how we feel about our family still plays an undeniably important role in kin altruism.