The interplay between moral actions and moral

Consciousness and Cognition xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog

The interplay between moral actions and moral judgments in children and adults ⁎

Janani Prabhakar , Deena Skolnick Weisberg, Alan M. Leslie Rutgers University New Brunswick, 152 Frelinghuysen Road, Piscataway, NJ 08854, United States

A R T IC LE I N F O

ABS TRA CT

Keywords: Moral reasoning Moral judgments Moral choices

Do choices about which moral actions to take cohere with subsequent judgments of their outcomes? The first set of experiments (N = 60 preschoolers and 30 adults) directly compared whether moral choices and judgments reflect distinct considerations, and whether coherence varies based on the valence of the moral scenario. Participants’ responses suggested that moral principles may be applied differently for moral choices and judgments, and that harm-based situations are particularly demanding for children. To determine whether children’s difficulty with harm-based situations reflects demand characteristics, a second set of experiments presented forty-three preschoolers and thirty-nine adults with a moral dilemma wherein they could choose to omit an action and maximize harm or act to minimize harm. Both age groups acted to minimize harm when caused indirectly. These results suggest that making choices about harm are not unilaterally demanding for preschoolers, but they struggle to make choices that minimize harm in a forced-choice scenario.

1. Introduction There is an important difference between making moral judgments and engaging in moral actions. You make moral judgments when you blame yourself for an ethical mistake and when you approve of the efforts of a charitable organization. You act on those judgments when you refrain from making the same mistake in the future and when you donate to the charity. The latter behaviors concern moral decisions. If there exists any coarse normative standard of ethical behavior, a good candidate would be that people should make decisions in accordance with their moral judgments; it would be a mistake to act in a way that violates one’s beliefs. The current study aimed to determine the extent to which either children or adults practice this principle: Do their choices about which of two moral actions to take cohere with how they judge the outcomes of these actions? Prior research provides some support for a connection between moral judgments and decisions. For example, Baird and Astington (2004) found that young children could evaluate another individual’s moral actions based on their underlying motives. This result suggests that children can use their moral knowledge to provide evaluations of morally charged actions of others. Prior work has also shown that adults and children evaluate moral outcomes based on the intentions of the agent that caused the outcome as well as on the relative utility of the outcome itself (i.e., saving 5 and harming 1 is permissible only if the intentions of the agent carrying out the action are good) (Cushman, Sheketoff, Wharton, & Carey, 2013; Cushman, Young, & Hauser, 2006). However, these findings do not measure children’s ability to engage in self-driven moral actions nor do they provide a measure of children’s judgments of self-driven moral outcomes. We believe that it is important to consider one’s own moral actions because, as argued by Blasi (1980, 1983), moral actions arise

⁎

Corresponding author. E-mail address: [email protected] (J. Prabhakar).

https://doi.org/10.1016/j.concog.2018.05.007 Received 16 August 2017; Received in revised form 25 April 2018; Accepted 16 May 2018 1053-8100/ Published by Elsevier Inc.

Please cite this article as: Prabhakar, J., Consciousness and Cognition (2018), https://doi.org/10.1016/j.concog.2018.05.007


J. Prabhakar et al.

from a motivation to maintain consistency between one’s own set of beliefs and how one chooses to act. In contrast, moral judgments of outcomes may not hinge upon reconciling differences between beliefs and actions, as these evaluations consider only past events. Given this background, in the present paper, we sought to directly compare how one’s beliefs and values are brought to bear when making moral evaluations of outcomes and when making moral decisions about future actions. In these studies, we recruited both adults and preschool-aged children. Early childhood reflects a distinct time during which children begin to learn and accumulate a set of moral beliefs (Dunn, Cutting, & Demetrious, 2000; Lane, Wellman, Olson, LaBounty, & Kerr, 2010; Smetana, Jambon, Conry-Murray, & Sturge-Apple, 2012). We sought to determine whether these beliefs are utilized differently for moral choices and moral judgments during this formative period of moral development in comparison to adults. 1.1. Moral reasoning, moral actions, and their co-development Children internalize moral principles in early childhood (Smetana, 1981, 1985; Turiel, 1978). For example, preschoolers distinguish moral rules like “don’t hit other people” from conventional rules such as “don’t wear pajamas to school” judging only the former as universally bad (e.g., Leslie, Mallon & DiCorcia, 2006; Nucci, 1985; Smetana, 1981, 1985; Turiel, 1978). In addition, young children understand the role of intention, judging that breaking ten cups is worse than breaking two, for example, but breaking a cup as a result of helpful intentions is better than breaking it as a result of rule violations (Surber, 1977; see also Núñez & Harris, 1998; Siegal & Peterson, 1998). In fact, past work has shown that in early childhood development, young children begin to weigh intentions over outcomes in their moral evaluations (Baird & Astington, 2004; Cushman et al., 2013; Vaish, Carpenter, & Tomasello, 2010; Zelazo, Helwig, & Lau, 1996). Because these rules and principles reflect children’s growing knowledge about abstract moral situations, they reflect a form of semantic knowledge that can be accessed during moral evaluations. That is, preschoolers evaluate morally charged events based on a set of semantic knowledge about moral rules and principles (Gall, 1985; Smetana, 1983; Wellman, Larkey, & Somerville, 1979). These results are consistent with other work showing that preschoolers can use semantic information to think about themselves in the future in non-moral situations, as well (Atance & Meltzoff, 2005; Hudson, Shapiro, & Sosa, 1995; Prabhakar & Hudson, 2014). But can preschool-aged children use moral knowledge to organize self-motivated actions in a morally-charged future scenario? Some recent work has shown that by five years of age, children begin to use moral principles in guiding their decisions about how to allocate resources. For example, while three-year-old children display knowledge about the principles of fairness, they do not tend to allocate resources fairly between themselves and others (Blake, McAuliffe, & Warneken, 2014; Shaw & Olson, 2012). It is not until the age of five that children begin to divide allocations that reflect their own moral principles of fairness. These studies suggest that children under five may be hesitant to act in a manner that would result in a disadvantage to themselves, even if it violates known moral principles of behavior. Choices that result in a personal disadvantage reflect one possible consequence of a violation between a moral principle and moral decision. However, not all moral decisions have an immediate tangible impact on the self. For example, choosing not to help a friend who is being bullied may not result in a disadvantageous outcome to the self, but it likely contradicts certain moral principles and expectations one holds. Furthermore, whether children choose to withhold help may diverge from their judgment of their own action afterwards. In the present study, we examined the moral principles children bring to bear when the consequences of one’s own action do not impact the self. We compared the principles that guide children’s actions to those used to evaluate the outcomes of the actions. While preschool-aged children can reason about their own freedom of choice in ways that impact their moral decisions (Kushnir, Wellman, & Chernyak, 2009), these results do not necessarily indicate that they understand how the outcomes of these choices should be judged (see also Chernyak & Kushnir, 2014). The current study, which compares choices about and judgments of the same actions, complements this previous research on children’s ability to evaluate past outcomes and behaviors by asking children to engage in a forward-looking process where moral outcomes are determined by their own choices and actions and not by a third-party agent. 1.2. The interaction of moral decision making with future thinking In the current studies, we employ a modified version of the classic thought experiment known as the trolley problem (see Foot, 1978; Thomson, 1986). In this problem, a trolley is headed down a track towards five people. A bystander could pull a switch to divert the trolley to another track, thereby saving the five individuals. However, this action will cause one individual on the other track to be run over and killed. Most adults and children think it is morally permissible for the bystander to pull the switch in this situation to minimize harm (Cushman et al., 2006; Cushman, Knobe, & Sinnott-Armstrong, 2008; Greene, Morelli, Lowenberg, Nystrom, & Cohen, 2001; Pellizzoni, Siegal, & Surian, 2010). In contrast, they judge an alternative scenario where a bystander can push an individual with a heavy backpack over a footbridge, killing this individual to save the other five, thereby also minimizing harm, as morally impermissible. These two opposing judgments suggest that adults and children’s moral evaluations are sensitive to the manner by which the outcome was achieved. Furthermore, they indicate that these individuals recruit moral principles to make choices for others. However, it is not yet clear whether individuals recruit these same moral principles when they themselves are deciding which outcomes they will be responsible for. This is the focus of the two experiments presented here and its primary departure from previous work: to directly distinguish between one’s personal moral actions and one’s judgments of the outcomes of these same actions. Thus, our first aim was to determine whether there was any consistency in how children chose to act and how they evaluated the outcomes of those actions. We assessed actions and evaluations of outcomes on whether children’s behavior when faced with two options of differing moral value cohere with utilitarian principles. This is not to say that a utilitarian perspective is the sole 2


J. Prabhakar et al.

determinant of children’s choices and judgments. Rather, we note that a utilitarian perspective offers a lens through which we can assess how children weigh two options against one another in a morally-charged scenario when making choices and judgments. Our second departure from previous work reflects our second aim. In this aim, we sought to understand how individuals make moral decisions about distributing benefits in addition to harms. The pioneering work of Carol Gilligan (1987) suggests that children may view benefits and prosocial actions differently from harms and antisocial actions. Research on moral decisions about benefits has shown that young children engage in prosocial and helping behaviors (Eisenberg & Fabes, 1999; Hepach, Vaish, & Tomasello, 2012; Warneken & Tomasello, 2006, 2007), even when given costly choices such as foregoing a highly preferred reward (Chernyak & Kushnir, 2013; Tasimi & Wynn, 2016). Young children also have intuitions about fairness when allocating varying amounts of harm or benefit (Blake et al., 2014; McCrink, Bloom, & Santos, 2010; Sheskin, Bloom, & Wynn, 2014; Smith, Blake, & Harris, 2013). For example, when asked to judge which of two puppets was nicer based on its allocation of resources to other puppets, children judged that the puppet that allocated the greater absolute amount was nicer (McCrink et al., 2010). Additionally, by the age of five, children employ principles of fairness in their own moral actions (Blake et al., 2014). However, prior work has not directly compared how children decide between options that result in outcomes of harm and benefit. Children and adults attribute greater agency to a character that produces negative side effects than one that produces positive side effects (Knobe, 2003; Leslie, et al., 2006; Pellizzoni, Siegal, & Surian, 2009), but these results do not tell us whether children’s personal choices distinguish between scenarios when the consequences do not impact the self. Asking about the distribution of benefits in the current studies thus allowed us to determine whether personal moral actions and moral judgments might be affected by the valence of the moral situation. 1.3. Experiments 1a and 1b: main research questions To understand how children employ moral principles when making a personal moral choice, we showed participants a game board with two hallways that diverged from one (Fig. 1). Participants were asked to decide down which hallway to send a dog: to a hallway with five characters or to a hallway with one. Critically, we varied whether the dog was friendly or angry, thereby creating two environments that differed in the valence of the moral situation but did not impact the valence of the child’s own environment. Our first main measure was participants’ choices: To which hallway would they send the angry or friendly dog? Our second measure was participants’ post-hoc evaluations of outcomes of their future-directed choices. To our knowledge, this is the first time that child participants have been asked to provide future-directed choices and post-hoc judgments about the outcomes of the same scenario, allowing us to determine whether there are distinct principles that guide forward-looking moral choices and backwards-looking moral judgments. 1.4. Experiments 1a and 1b: secondary research questions We included two additional manipulations to further understand differences in moral choices and moral judgments. First, we varied the method by which participants were asked to make their future choice: Half were asked to choose which hallway to block from receiving a harm or benefit, while the other half were asked to point to which hallway harm or benefit should be sent. This measure was introduced to examine whether children and adults would distinguish different ways of achieving an outcome in their future-directed moral choices, as they do in their moral evaluations of others’ actions (Cushman et al., 2006, 2008; Greene et al., 2008; Pellizzoni et al., 2010). Because we asked children to reason about both harm and benefit, inclusion of this manipulation also allowed us to determine whether individuals prefer to act in different ways based on the valence of the outcome. The second additional manipulation involved varying the number of characters about which participants were asked to make a

Fig. 1. This figure depicts the hallway that children were shown for the angry (a) and friendly (b) dog. All children were asked to make a choice between 1 and 0 characters, as well as 1 and 5 characters for both the friendly and angry dog, altogether providing four choices. This figure depicts a subset of these choices. 3


J. Prabhakar et al.

moral choice. We asked participants to choose between harming and benefitting 1 versus 5 characters and to choose between harming and benefitting 1 versus 0 characters. In the former situation, any choice would lead to a negative outcome for one set of characters. In the latter situation, one choice leads only to positive outcomes while the other leads only to negative outcomes. This numerical comparison, therefore, allowed us to determine whether children’s moral actions and subsequent judgments depend on their ability to compare relative outcomes of good and bad. We tested 4- and 5-year-old children using this experimental design (Exp. 1a), and included an adult sample as well to serve as a comparison group (Exp. 1b). Adults were told that we wanted to determine whether our experimental design would be viable for children, and to answer as truthfully and honestly as possible, even if the stimuli appeared juvenile. 2. Experiment 1a methods 2.1. Participants Sixty preschoolers (mean age = 56 months; range: 48–64 months; 29 females) participated in the study. One additional child was tested but excluded for having already participated in a pilot version of this study. Children were recruited from and tested at local preschools. Parental consent was obtained for all participants, and they were given stickers and a certificate for their participation. 2.2. Materials and procedures 2.2.1. Practice trials Children first engaged in two practice trials with the hallway. These trials first asked participants to send a soccer ball to an empty hallway or to a hallway with a goal and then to send a character to her teacher’s preferred food item or disliked food item. All participants always sent the ball to the goal and the character to the hallway with the teacher’s preferred food item. Participants were additionally trained to make judgments using a five-point Likert scale (Fig. 2). The scale ranged from “really bad” (denoted by X’s) to “really good” (denoted by stars), with “just okay” (denoted by an open circle) at the center. They were introduced to the positive, negative, and neutral points of the scale, and asked to judge simple, non-social scenarios that clearly fell on either end of the scale or in the middle. All participants provided appropriate judgments for each item. 2.2.2. Experimental trials There were four experimental trials. In each trial, participants were asked to make a moral choice, immediately followed by a post-hoc evaluation of the outcome using the scale. The study’s four trials were grouped into two blocks: a Harm block and a Benefit block. In the Harm block, participants were presented two stories about an angry dog (the Harm condition); in the Benefit block, they were presented two stories about a friendly dog (the Benefit condition). When the experimenter introduced each dog, she described the dog’s personality as either “mean” or “nice” and explained that the dog will either knock the characters over and make them sad or play with them and make them happy. We randomized the order of the blocks by subject. Within each harm and benefit block, participants were given two stories based on the number of children at the end of each hallway (order randomized within each block). In the 1vs0 condition, they were shown 0 characters at the end of one hallway and 1 character at the end of the other. In the 1vs5 condition, there was 1 character at the end of one hallway and 5 at the end of the other. Participants were therefore presented with 4 stories total: 2 numerical combinations (1vs0 characters and 1vs5 characters) within each of 2 valence blocks (one that reflected harm situations and one that reflected benefit situations). For each story, the experimenter first described the outcome that would result from making each choice: “If the mean dog goes

Fig. 2. Children were asked to provide evaluations of scenarios with this 5-point scale. The stars reflected evaluations of “good,” the x’s reflected evaluations of “bad,” and the circle in the middle reflected an evaluation of “just okay.” The ends of the scale reflect “really bad” and “really good” evaluations. 4


J. Prabhakar et al.

this way [pointing to the appropriate hallway], it will knock over 5 kids and make them sad. And if the mean dog goes this way [pointing to the other hallway], it will knock over 1 kid and make him sad.” The experimenter then asked participants to make a choice. We manipulated the manner by which children made their choice as a between-subject measure. Half of the participants used a green building block to stop the dog from entering one of the hallways (Blocking condition). The rest of the participants simply pointed to the hallway where they wanted the dog to go (Pointing condition). Immediately following this choice, the experimenter moved the dog down the indicated hallway and explained the outcome (e.g., one kid is sad and five are happy). Participants were asked to judge this outcome using the pink scale: “Was that good, bad, or just okay?” If a participant said that the outcome was good or bad, a follow-up question was asked: “Was it really good/bad or just a little good/bad?” We incorporated one additional manipulation. Half of the children (n = 30) judged the outcome that they themselves had just caused, and the other half of the children (n = 30) judged the opposite outcome (i.e., the outcome that would have occurred if this participant had sent the dog down the other hallway). This manipulation was incorporated to determine the objectivity of children’s judgments, since they may display a bias to affirm any choice that they themselves made. 2.2.3. Coding For moral choices, the design included two within-subject conditions (Valence, harm and benefit, and Numerical Comparison, 1v0 and 1v5) and one between-subject condition (Manner of Choice, block or point). Choices could result in one of two outcomes: one that maximized utility (i.e., benefiting more or harming fewer, coded as 1) or one that minimized utility (i.e., benefiting fewer or harming more, coded as 0). For moral judgments, children’s responses were made on a 5-point scale. We converted this to a numerical scale for analysis, with −2 corresponding to “really bad,” +2 corresponding to “really good” and 0 corresponding to a “just okay” response. We additionally generated a second scoring scheme that converted these moral judgments to a binary variable to reflect utilitarian and non-utilitarian evaluations; this allowed us to directly compare judgments with choices. In this scheme, positive ratings (little good or really good) for a utilitarian outcome and negative ratings (little bad or really bad) for non-utilitarian outcomes were given a score of 1. Negative ratings (little bad or really bad) for a utilitarian outcome and positive ratings of non-utilitarian outcomes, as well as scores of “just okay”, were given a score of 0. We chose to score responses of “just okay” as 0 as they could not be clearly identified as positive or negative toward a utilitarian or non-utilitarian outcome, respectively. This binary scoring system was used only when comparing moral choices to moral judgments. 3. Results Initial analyses revealed that outcome judged (one’s own or the alternative) did not predict moral judgments, B = 0.002, p = 0.994. Therefore, the remaining analyses do not consider this variable. We divide the results into two sections. Our main research aims were to compare children's moral choices and judgments, and determine whether any differences between choices and judgments depend on the valence of the scenario. Thus, in the first section, we compare the effect of valence (harm or benefit) on children’s choices and on their judgments. The second section reflects supplementary analyses to address the secondary research questions about the potential roles of the numerical comparison (1v0 or 1v5) and manner of choice (point or block) conditions. 3.1. Primary analyses: effect of measure (choice or judgments) and valence (harm or benefit) A mixed model binomial regression was computed on the proportion of utilitarian responses with measure (choice or judgment) and valence (harm or benefit) as predictors (Fig. 3). The results indicate that measure was a significant predictor of utilitarian

Fig. 3. This figure depicts the proportion of responses that reflect maximizing benefits and minimizing harm by measure (choice, judgment) and valence (harm, benefit). The errors bars reflect standard error. 5


J. Prabhakar et al.

Fig. 4. This figure presents the average utilitarian choice by manner of choice, valence and numerical comparison. Error bars reflect standard error.

responses: children used utilitarian principles more often when making a moral choice (M = 0.68, SD = 0.49) than a moral judgment (M = 0.49, SD = 0.50), B = 1.39, p < 0.001. The results also found that valence was a significant predictor of utilitarian responses: children used utilitarian principles more often when reasoning about benefit (M = 0.62, SD = 0.49) than harm (M = 0.55, SD = 0.50), B = 0.84, p = 0.004. The regression also yielded a significant interaction between valence and measure, B = 1.08, p = 0.006. Overall, children made choices and judgments that favored maximizing benefit (M = 0.61, SD = 0.49) more often than choices and judgments that favored minimizing harm (M = 0.55, SD = 0.50), suggesting that benefits may be easier for children to reason about than harms. Looking just at their moral choices, children chose to maximize benefit (M = 0.77, SD = 0.42) more often than they chose to minimize harm (M = 0.59, SD = 0.49), t(59) = 2.79, p < 0.01, d = 0.36. That is, children were more likely to send the happy dog to the hallway with more characters than to send the angry dog to the hallway with fewer characters. However, children showed no difference in their moral judgments. That is, they did not rate maximizing benefits (M = 0.46, SD = 0.50) differently from minimizing harm (M = 0.52, SD = 0.50). Taken together, these results indicate that the extent to which children provide a utilitarian response (i.e., maximizing benefit and minimizing harm) differs based on whether one is making a moral choice versus moral judgment. 3.2. Supplementary analyses: effect of numerical comparison and manner of choice conditions on moral choices and judgments 3.2.1. Moral choices For children’s choices, a mixed model binomial regression was computed on the proportion of utilitarian choices with manner of choice, numerical comparison, and valence conditions as predictors. The regression found that manner of choice, B = 2.38, p = 0.031, was a significant predictor of selecting a utilitarian option (Fig. 4). This was qualified by a significant interaction between numerical comparison and manner of choice, B = 2.54, p = 0.040, and a significant interaction between manner of choice and valence, B = 2.96, p = 0.016. When children could block the dog from a hallway, they chose the utilitarian option to the same degree in both numerical comparison conditions, t(59) = 0.20, p = 0.843. In contrast, they were more likely to make a utilitarian choice when asked to point to 1vs0 versus point to 1vs5, t(59) = 2.09, p = 0.04, d = 0.27. When asked to block harm and block benefit, children showed no significant difference in their choices, t(59) = 0.60, p = 0.553. However, they made more utilitarian choices when asked where to point benefit versus where to point harm, t(59) = 3.46, p = 0.001, d = 0.45. Using single sample t-tests, we compared children’s performance to chance (1/2). We found that children were no different from chance when deciding where to send an angry dog either in the 1vs0 condition, t(59) = 1.57, p = 0.12, or in the 1vs5 condition, t (59) = 1.30, p = 0.20. They chose the utilitarian option significantly above chance in the benefit condition in both numerical comparison conditions (ps < 0.05). 3.2.2 Moral judgments To evaluate children’s judgments, we ran a mixed model linear regression to determine whether numerical comparison (1vs0 vs. 1vs5), manner of choice (point vs. block), valence (harm vs. benefit), or outcome type (utilitarian vs. non-utilitarian) predicted children’s moral ratings (see Fig. 5). A Type III Wald test found that valence, X = 10.37, p = 0.001, numerical comparison, X = 8.92, p = 0.003, and outcome type, X = 54.82, p < 0.0001, were significant predictors of children’s moral judgments. In addition, the Wald test found a significant interaction between manner of choice and numerical comparison, X = 3.96, p = 0.047, as well as between numerical comparison and outcome type, X = 44.14, p < 0.001. Critically, children rated outcomes that minimized benefit and maximized harm in the 1vs0 condition (M = −1.21, SD = 0.85) more negatively than outcomes that maximized benefit and minimized harm in the 1vs0 condition (M = 0.88, SD = 0.96), t(85) = 10.76, p < 0.001. The same difference between harm and benefit was not found for the 1vs5 condition, t(86) = 0.47, p = 0.64. 4. Discussion The primary goal of Experiment 1a was to contrast children’s moral actions and moral judgments about the outcomes of the same scenario. Because we asked the same participants to choose which action to take and to judge the resulting outcome, this design allowed us to separate children’s moral reasoning behavior for each task, and hence to determine whether they involve different moral considerations. A second innovation of this design was to explore how children reason about benefit and harm within the same 6


J. Prabhakar et al.

Fig. 5. This figure presents the average ratings for outcomes of (a) 1 vs. 0 and (b) 1 vs. 5. In both figures, mean ratings based on outcome type (utilitarian vs. non-utilitarian), manner of choice (block vs. point), and valence (benefit vs. harm) are provided.

experimental design. Overall, our results suggest that children apply moral principles differently when making moral choices and moral judgments: Children provided more utilitarian responses when making moral choices than judgments. That is, they favored maximizing benefit and minimizing harm when making a moral choice, but were less likely to do so when providing a post-hoc moral evaluation. One possible explanation for this difference is that participants may have been critical of their own choices. However, we found no differences between the judgments children provided for the outcomes of their own choices and for the outcomes that they did not choose. This suggests that children’s post-hoc judgments were objective and did not necessarily favor their own decisions. Another possible explanation is that children focused more heavily on the outcomes rather than how the outcomes were achieved when they provided their ratings. An outcome-based strategy would, indeed, lead children to engage with providing moral judgments differently from making moral decisions. These results thus indicate that the process of making moral judgments may have some degree of independence from the process of making moral choices. Furthermore, these results cohere with Blasi’s (1983) notion that perceptions of personal responsibility may dictate how children choose to behave. Post-hoc evaluations of the outcome, which in this case had little personal value to or influence on the participant (cf., Miller, Starmans, Fugelsang, & Friedman, 2016), may be sufficiently divorced from these perceptions and from self-interest. Thus, children may make a choice to minimize harm and maximize benefit, but provide judgments that do not favor these decisions. We also found that children tended to maximize benefit more often than minimize harm. However, their tendency to provide such a response differed based on whether they were making a moral choice or moral judgment. Specifically, they tended to make choices that maximized benefit, but did not judge the corresponding outcome favorably. This suggests that children’s choices and subsequent judgments do not align when reasoning about a positive moral outcome, possibly because they had difficulty giving a positive rating to any outcome in which even a single character did not receive a benefit. By contrast, in the harm condition, they made fewer choices that minimized harm and judged all outcomes of harm negatively. This suggests that reasoning about harm, in general, may be more difficult for children than reasoning about benefits. The notion that reasoning about harm may be more demanding for children is supported by the fact that they performed poorly in both numerical comparison conditions. Children’s choices in the harm condition did not differ from chance, even in the 1vs0 condition. Further, when providing judgments, children rated any outcome where at least 1 character was harmed negatively. This result departs from previous work wherein children judged outcomes that minimize harm indirectly as permissible (Pellizzoni et al., 2010). In the present study, we show that even if an outcome is deemed permissible, children do not find it morally acceptable. Future research would benefit from distinguishing between notions of permissibility and moral acceptability. In general, Experiment 1a suggests that the set of moral beliefs that dictate children’s personal moral choices do not necessarily dictate how they evaluate the outcomes of these choices. Whether this difference reflects a unique aspect of early moral development or a more general cognitive characteristic of moral reasoning is the focus of Experiment 1b. In Experiment 1b, we replicated Experiment 1a with an adult sample to serve as a comparison group to the children. 5. Experiment 1b methods 5.1. Participants Thirty adults (mean age = 20 years; range: 18–29 years; 18 females) participated in the study. Adults were undergraduate students who participated in this study for credit in a Psychology course. 5.2. Materials and procedure The materials and procedure were identical to Experiment 1a except for one minor difference. Because we found no difference in ratings in the outcome rated condition (their own or alternative) in Experiment 1a, all adults in this experiment were only asked to rate the outcome of their own choice. In the end, all subjects provided four moral choice responses and four moral judgment 7


J. Prabhakar et al.

Fig. 6. This figure depicts the proportion of responses that reflect maximizing benefits and minimizing harm by measure (choice, judgment) and valence (harm, benefit) in adults. The errors bars reflect standard error.

responses. The resulting experimental design included two within-subject conditions (Valence and Numerical Comparison) and one between-subject condition (Manner of Choice). 6. Results We used the same two types of scoring as in Experiment 1a. The results section is again split into two sections. The first explores the effect of measure (moral choices, moral judgments) and valence (harm, benefit) on adults’ moral choices. The second examines any additional effects of the numerical comparison and manner of choice conditions. 6.1. Main analyses: effect of measure (choice or judgments) and valence A mixed model binomial regression was computed on the proportion of responses that reflected minimizing harm and maximizing benefit with measure and valence as predictors (Fig. 6). The regression found that valence was a marginally significant predictor of a utilitarian response pattern, B = 1.01, p = 0.056. Specifically, adults provided somewhat more utilitarian responses in the harm (M = 0.76, SD = 0.43) than benefit (M = 0.72, SD = 0.45) conditions. In addition, the regression yielded a marginally significant interaction between measure and valence, B = 1.23, p = 0.06. Adults provided significantly more utilitarian responses when making a choice (M = 0.90, SD = 0.30) than when making a judgment (M = 0.62, SD = 0.49) about harm, t(59) = 3.43, p = 0.001, d = 0.44. Similarly, adults provided more utilitarian responses when making a choice (M = 0.77, SD = 0.43) than when making a judgment (M = 0.67, SD = 0.48) about benefit; however, this difference was only marginally significant, t(59) = 1.76, p = 0.083, d = 0.23. 6.2. Supplementary analyses: effect of manner of choice and numerical comparison on moral choices and judgments 6.2.1. Moral choices Because all adults chose to maximize benefit and minimize harm in the 1vs0 condition, a mixed model binomial regression was computed on the proportion of utilitarian choices in the 1vs5 condition only with manner of choice and valence as predictors (Fig. 7). The regression found that valence was a significant predictor, B = 13.35, p = 0.003, as was manner of choice, B = 20.62, p < 0.001. Unlike children, when reasoning about 1vs5, adults were significantly more likely to make choices that minimized harm (M = 0.80, SD = 0.41) than maximized benefit (M = 0.53, SD = 0.51), and were more likely to make the utilitarian choice by blocking (M = 0.70, SD = 0.47) than by pointing (M = 0.63, SD = 0.49). 6.2.2. Moral judgments Next, we examined adults’ moral ratings. Because all adults chose to maximize benefit and minimize harm in the 1vs0 condition,

Fig. 7. This figure presents adults’ proportion of utilitarian choices by manner of choice (block vs. point), numerical comparison (1vs0 vs. 1vs5) and valence (harm vs. benefit). Error bars reflect standard error. 8


J. Prabhakar et al.

Fig. 8. This figure presents ratings of outcomes of (a) 1 vs. 0 and (b) 1 vs. 5. For 1 vs. 0, all adults judged outcomes of utilitarian ratings. This figure therefore shows ratings based on valence and manner of choice. For outcomes of 1 vs. 5, adults rated both. As a result, this figure shows ratings based on outcome type, valence and manner of choice. Errors bars reflect standard error.

we examined their ratings in the 1vs0 (Fig. 8A) and 1vs5 (Fig. 8B) condition separately. First, we ran a mixed model linear regression with valence (harm vs. benefit) and manner of choice (block vs. point) as predictors of ratings of utilitarian outcomes (i.e., those that minimize harm and maximize benefit) for 1vs0 scenarios. The regression yielded no significant effects. Next, we ran a mixed model regression with valence (harm vs. benefit), manner of choice (block vs. point) and outcome type (utilitarian vs. non-utilitarian outcome) as predictors of adults’ moral ratings in the 1vs5 condition. A Type III Wald test found that valence, X = 34.52, p < 0.001, was a significant predictor of moral ratings in the 1vs5 condition. The test also yielded a significant interaction between valence and outcome type, X = 5.38, p = 0.02. While adults rated benefit to 1 versus 5 as a little good (M = 1.00, SD = 0.78) and harm to 1 versus 5 as only slightly negative (M = −0.54, SD = 1.14), t(36) = 4.64, p < 0.001, they rated benefit to 5 versus 1 (M = 0.63, SD = 0.89) as only slightly positive and harm to 5 versus 1 as more than a little bad (M = −1.33, SD = 0.52), t(20) = 5.06, p < 0.001. 7. Discussion Experiment 1b demonstrated that the set of moral guidelines that underlie moral actions and moral judgments appear to be differentiable in adults, as they are in children. Adults were more utilitarian in their choices than their judgments across both valence conditions. However, adults were also more utilitarian when reasoning about harm than benefit, but only in terms of their choices. These results illustrate some of the differences we found between children and adults’ performance patterns. Adults chose to minimize harm more often than did children, and children were more utilitarian in the benefit condition than were adults. Adults may have found the option to save the greater number of characters from harm more intuitive, in line with past work on adults’ moral evaluations (Cushman et al., 2006). When reasoning about benefit, adults may have acted on a fairness principle such that allocating all the benefit to one set of characters may not have been a preferred option. The valence of the outcome thus influences the choices that individuals make, but this influence manifests itself differently by age. In these experiments, we did not ask participants to tell us why they made their decisions. As such, we cannot say definitively that different sets of moral principles are used to guide moral behavior in these tasks. However, our paper provides the first step in this direction by showing that the principles that would guide one to favor maximizing benefit and minimizing harm are applied differently when one is faced with taking on responsibility for the actions (i.e., moral actions) versus when one must evaluate the outcomes of the actions (i.e., moral judgments). Future research is required to determine whether distinct principles are used in these two tasks, or whether the same principles manifest differently based on the task. Future work should also explore when children violate these principles and the processes that underlie the lack of consistency between one’s set of moral beliefs and their choice of action. 8. Experiment 2a One unexpected result from the first set of experiments was that children struggled to make adult-like choices in the harm condition. Even when asked to make a choice between 1 and 0 characters, they chose to minimize harm only 60% of the time. One possible explanation for these age-related differences is that children may have found it difficult to be responsible for causing harm to any group of characters. These results contrast with prior work that has shown that children and adults are sensitive to the manner by which harm is caused (Cushman et al., 2006; Pellizzoni et al., 2010). One reason for this discrepancy could be that children may have thought that the moral choice was just a game. That is, they may not have entirely appreciated that their choice would ultimately lead to a real outcome. As such, they may have chosen randomly between the two options, even in the 1vs0 condition, and then judged the subsequent outcome negatively, having recognized that harming someone is morally bad. However, in the benefit condition, children could make a choice between 1 and 0 characters, sending the friendly dog to 1 character more often. 9


J. Prabhakar et al.

A different possibility is that the forced-choice nature of the task in Experiment 1a and 1b made it difficult for children to make a choice that involved causing harm to a character. If this were the case, it is possible that they opted to not consider the options presented to them in the forced-choice task, and behave in such a way that displays random choice behavior. To address this possibility, Experiment 2a gave children the option to not act and thereby avoid being the agent of harm. In this experiment, participants were presented with a dilemma: They could act in a utilitarian manner to save 5 characters from harm by causing harm to 1 character through indirect (harm as a side effect) or direct means (harm as a main effect) or they could choose not to act altogether (omission of action), resulting in harm to 5 characters (a non-utilitarian outcome). This design allows us to test whether children would let events occur without interference in order to avoid being agents of harm, even though that would result in an outcome that maximizes harm. It could be the case that comparing an outcome to a default (i.e., a larger number of characters harmed without action) may reduce the cognitive demands that result from having to make a forced choice about harm, allowing children to consider other possible outcome options since the onus to act has been removed. However, if children generally wish to avoid causing any harm themselves, then they should always prefer a violation by omission. Given that children can distinguish between violations by omission and commission when evaluating post-hoc outcomes of other’s actions (Pellizzoni et al., 2010), we predicted that children would be able to do the same when asked to make a forward-looking choice about a similar situation. In addition to making this choice, as in Experiment 1a, children were asked to make moral judgments to obtain another comparison of moral actions and judgments. Because we did not find a difference between judgments of their own outcomes or the opposite outcome in Experiment 1a, here children evaluated only the outcomes of their own choices.

9. Methods 9.1. Participants Participants were 43 preschoolers (mean age: 57 months; range: 48–65 months; 20 females) recruited from and tested at local preschools. Parental consent was obtained for children, and they were given stickers for their participation.

9.2. Materials and procedures Experiment 2a was like Experiments 1a and 1b except the task was given on a touchscreen tablet. At the start of the procedure, children saw an angry dog headed toward 5 characters. They were told that the angry dog was going to bark at these kids and make them sad. They could let the dog bark at the 5 kids (omission), or they could act in one of two ways. They could move a gate to block the dog from running up the path toward the 5 characters, causing the dog to run down a path toward 1 character (side effect). Or, they could push the 1 character toward the dog, distracting him from the 5 (main effect). With both options to act, the same outcome would occur (5 characters would be saved, 1 character would be harmed). Like Experiment 1a, the outcome of each choice was verbally described with the pathways in view. In addition, children saw images that depicted the outcome of each choice before they could select one. Children were always given the option not to act first and then the order of the other two options were counterbalanced between subjects. Following their selection, we asked children why they made their choice. Then, they were asked to evaluate the outcome using the same 5-point Likert scale as in Experiment 1a, now presented on the touchscreen. Above the scale, children saw a set of pictures of the steps that lead to the outcome: the starting scenario, the dog starting to move, and the final outcome. This was included to reduce any memory demands for the outcome of their choice.

Fig. 9. This figure presents the number of children who selected each of the three response options: omission, side effect and main effect. 10


J. Prabhakar et al.

10. Results 10.1. Choices A one-way Chi-Square test was conducted on the frequency of omission, side effect, and main effect choices (see Fig. 9). The results indicate a marginal difference in children’s choices among these 3 options, χ2(2) = 5.91, p = 0.052. Children (n = 21) chose to act by side effect significantly more than by chance (1/3), p = 0.01. Children did not choose main effect (n = 14) significantly more or less than what would be expected by chance, p = 1. Finally, children selected to omit action (n = 8) significantly less than by chance, p = 0.006. 10.2. Justifications Participants’ justifications after a choice to act were coded based on whether they reflected a utilitarian principle (e.g., “So the 5 kids don’t get sad,” “So he won’t bark at 5 kids,” etc.) or not (e.g., “So the 1 kid doesn’t get sad,” “I wanted the dog to bark at 1 kid,” etc.). Justifications that did not fit either of these principles were categorized as other (e.g., “I liked it,” “I don’t know,” “The dog is angry,” etc.). As such, we coded children’s responses into one of three categories: utilitarian response, non-utilitarian response, and other; chance performance is thus 1/3. We examined children’s justifications after a choice to act by side effect, choice to act by main effect, and choice of omission. We found that more than one third of the children who chose to harm by side effect provided a utilitarian justification (x = 14, N = 21), p = 0.002. Furthermore, we found that more than one third of children who chose to act by a main effect provided justifications that went in the “other” category (x = 10, N = 14), p = 0.004, as did children who chose omission (x = 6, N = 8), p = 0.20. 10.3. Judgments A one-way ANOVA was conducted to determine the effect of choice on children’s judgments of outcomes (Fig. 10). The ANOVA did not yield a significant effect of choice, F(2, 40) = 1.02, p = 0.37. A series of single sample t-tests comparing children’s ratings to ratings of just okay (0) found that children rated all outcomes of harm negatively (omission: t(7) = 3.99, p = 0.005; side effect: t (20) = 3.25, p = 0.004; main effect: t(13) = 2.92, p = 0.012). 11. Discussion In Experiment 2a, participants could choose two methods of minimizing harm: moving a gate to divert the dog to another path with only 1 character, or pushing the 1 character to the dog, thereby distracting it from the 5 characters. They could also choose not to act, in which case a greater number of characters would be harmed. We predicted that, with the option to omit action, children would face fewer demands in evaluating potential future outcomes of harm and would be less reticent to allow greater harm to occur, and instead decide to act in a way that reduces harm to the greater number. As predicted, children chose to act by side effect more often than to omit action. However, they did not choose to act by a main effect significantly more or less than by chance. Children’s justifications of their choices reflected utilitarian principles only when they had chosen the side effect option. This suggests that those children who chose to cause harm as a side effect may have had conscious access to principles that guided the manner by which they could reduce harm. However, children’s justifications of main effect actions did not reflect reference to outcomes to either group of characters. Because the number of children who chose to act by main effect did not significantly differ from chance, it could be the case that these children had difficulty choosing between the two options to act and thereby did not have a specific reason for doing so that pertained to the outcome.

Fig. 10. This figure presents children’s ratings on the 5-point pink scale of judgments of outcomes following a main effect, omission, and side effect choice. Error bars reflect SE. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 11


J. Prabhakar et al.

Like the Harm condition of Experiment 1a, children rated all outcomes negatively, regardless of the manner by which the outcome was achieved. Although previous literature reported children’s judgments are sensitive to the manner by which harm was caused (Pellizzoni et al., 2010), those studies asked children to judge the permissibility of an actor’s intentions. Here, we asked children to evaluate the acceptability of an outcome that they played a role in effecting. This change in the locus of the decision process may have resulted in greater sensitivity to the existence of any harm in the outcome environment. Furthermore, a focus on outcome over intention may have substantiated sensitivity toward harmed characters. As in Experiment 1b, we chose to include an adult sample with identical methods to Experiment 2a, both to obtain a comparison sample to children in Experiment 2a and also to establish an initial understanding about adults’ choices and subsequent judgments in a moral dilemma. 12. Experiment 2b methods 12.1. Participants Participants were 39 adults (mean age: 19.2 years; range: 19 to 29 years; 18 females) who were undergraduate students. Adults provided written consent and participated in exchange for credits in a Psychology course. 12.2. Materials and procedures Experiment 2b was identical to Experiment 2a. Adults were asked to choose between three options (omission, side effect, or main effect) before providing a judgment of the outcome of their choice. 13. Results 13.1. Choices Fig. 11 displays the number of adults who selected each choice: omission, side effect, and main effect. A one-way Chi-Square test was conducted on the frequency of omission, side effect, and main effect choices. The results indicate differences in adults’ (χ2(2) = 9.85, p = 0.007) choices among these 3 options. Adults (n = 21) chose to act by side effect significantly more than by chance (1/3), p = 0.01. A adults chose to act by main effect significantly less than by chance (n = 5), p = 0.006 and omitted action at rates that did not differ from chance (n = 13), p = 1. 13.2. Justifications Adults’ justifications were coded in a similar manner as Experiment 2a. Overall, adults provided a utilitarian justification after choosing to act by side effect (x = 19, N = 21), p < 0.001. Their justifications after choosing to act by a main effect all reflected a utilitarian justification (x = 5, N = 5). Finally, when adults acted by omission, their responses largely reflected non-utilitarian justifications (x = 12, N = 13), p < 0.001. 13.3. Judgments A one-way ANOVA was conducted to determine the effect of choice on adults’ judgments of outcomes (Fig. 12). The ANOVA yielded a marginally significant effect of choice, F(2, 36) = 2.76, p = 0.077. Specifically, adults rated outcomes of side effect actions (M = -0.67, SD = 0.66) significantly more negatively than outcomes of main effect actions (M = 0.20, SD = 0.84), t(24) = 2.52, p = 0.019, d = 0.9747824. A series of single sample t-tests comparing adults’ ratings to ratings of just okay (0) found that adults rated the outcomes of omissions and side effects negatively (omission: t(12) = 2.55, p = 0.025; side effect: t(20) = 4.64, p < 0.001), but

Fig. 11. This figure displays the number of adults who chose each response option: omission, side effect and main effect. 12


J. Prabhakar et al.

Fig. 12. This figure presents the average ratings by adults on the 5-point pink scale following outcomes of main effect, omission and side effect. The error bars reflect SE. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

not main effects (t(4) = 0.52, p = 0.621). 14. Discussion In Experiment 2b, adults chose to act by side effect more often than acting by main effect or omitting action. Between the other two options, they chose to omit action more than to act by main effect. That is, while adults did prefer to act, they avoided doing so if reduction of harm was caused by a main effect. In contrast, children were at chance in deciding whether to act by main effect. So, while both children and adults preferred to reduce harm by a side effect action, only adults preferred to act only in this case and did not choose to act if harm were caused as a main effect, like previous work. Like children in Experiment 2a, adults in Experiment 2b rated all outcomes negatively. The possible focus on the outcome may have led to adults’ negative ratings, considering that adults rated outcomes of omission of action and side effect action negatively. (The ratings of “just okay” after a main effect action are unreliable given the small number of adults who chose to act in this way.) 15. General discussion The experiments presented here sought to explore the nature of children’s and adults’ moral choices and judgments to achieve two primary goals. Our first goal was to determine whether children and adults employ similar sets of beliefs when they engage in personal moral actions and when they provide judgments of the outcomes of these actions. Our results are consistent with the notion that different moral considerations might underlie these types of moral cognitions; moral decisions do not always cohere with moral evaluations. That is, the semantic principles that guide one’s moral behavior may have a degree of independence from those that guide one’s moral evaluations. As such, this result may reflect the differential requirements of forward-looking versus backwardlooking processes. When making a future-oriented decision, individuals apply semantic knowledge not necessarily to select the most desirable option, but rather to select the least undesirable one that coheres with their own moral principles and code of conduct. When providing post-hoc evaluations of the outcomes, on the other hand, individuals may not be as encumbered by feelings of responsibility and can afford more scrutiny to the goodness or badness of the moral outcome while ignoring the actions or intentions that led to it. The current studies begin to illustrate the interplay between moral choices and moral judgments. However, more work is needed to determine exactly why these processes differ. One fruitful direction for future work would be to directly assess the coherence between moral choices and judgments. In the present experiment, we asked participants to focus on the outcomes of the choices when they made their judgments. Future work could ask participants to evaluate their choice itself rather than the outcome of that choice. Indeed, prior work has shown that preschool-aged children provide evaluations of a character’s helping and harming behavior based on the intentions of the character’s choice, and evaluations about goodness may emerge earlier than evaluations about badness (Margoni & Surian, 2016, 2017). Whether this translates to evaluations of their own choices has not yet been tested and is an avenue for future research. We described moral choices as being informed by an underlying set of moral principles. However, in this study, we did not ask children to first identify moral principles before acting. Instead, children chose an action and then provided an evaluation of its outcome. We did not counterbalance this order because our goal was to determine whether children’s own actions influenced how they felt about the subsequent outcomes. However, the alternative question is also relevant: Do children’s judgments of outcomes influence how they then choose to act? The bidirectionality of how choices and judgments interact is an open question that can be answered using the novel methodology presented in this paper. Another goal of the current studies was to determine whether distinctions between moral actions and judgments are particularly apparent in early moral development. Again, we found that this was the case. While we found that adults and children displayed similarities in terms of their choices and judgments, they differed in terms of how they responded to outcomes of benefit and harm. Adults tended to minimize harm more often than did children, and children tended to maximize benefit more often than did adults. Further, the contrast between children’s behavior in Experiments 1a and 2a suggests that children may find it particularly demanding 13


J. Prabhakar et al.

to make a forced-choice decision about how to allocate harm. When they are given the option to avoid action by maximizing harm, they behave similarly to adults and choose to act to minimize harm instead. These results indicate that while both adults and children utilize utilitarian principles when reasoning about harm, the ability to retrieve these principles to guide moral decisions in a forcedchoice scenario matures. Similarly, principles that guide decisions about benefit change between childhood and adulthood. While children prefer to maximize benefit, adults are torn between maximizing utility and notions of fairness. Future studies would benefit from exploring the developmental trajectory of these abilities into middle childhood when moral principles are utilized in social environments to a greater extent. Future studies should also address several limitations to the present paradigm. First, participants in the current experiments were asked to judge the outcome of only one choice. As a result, this method does not allow us to determine other aspects of the environment that might influence post-hoc ratings. For example, our focus on outcomes themselves means we do not know whether children distinguish, post-hoc, the underlying intentions of the decision from the manner by which the outcome came about or from the inherent goodness or badness of the outcome itself. Future research would benefit from directly asking participants to rate all of these factors in order to determine their independent contributions to children’s judgments. Second, Experiment 1a provided children with four stories split between valence and numerical comparison conditions. In contrast, Experiment 2a asked children to make choices about just one story. There may thus have been more fatigue effects in Experiment 1a than 2a, creating greater demands in the harm condition in Experiment 1a. While we do not believe that fatigue effects can entirely explain the current findings, future experiments could administer the forced-choice paradigm (Experiment 1a) with valence as a between-subject measure to directly measure task demands. Third, children’s choices cohered with adults’ in Experiments 2a and 2b but not Experiments 1a and 1b. One possibility is that having the actions shown in a continuous, movie-like fashion on the tablet in Experiments 2a and 2b may have promoted understanding of their choices. Thus, this paradigm offers insight into how different types of media sources can be harnessed to assess children’s cognitive capacities. Together, these two studies provide important clues about the connections between children and adults’ abilities to make choices as the agent of moral action and to evaluate the outcome that resulted. Future work should continue these within-subjects comparisons to more fully explore the development of moral principles and how they are employed across development and into adulthood. Acknowledgements We would like to thank all families and preschools that volunteered their time for this project. We would also like to thank all the research assistants for their efforts in assisting with this project. Finally, we would like to thank Sangeet Khemlani and all members of the Cognitive Development Laboratory for their advice and feedback on this work. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.concog.2018.05. 007. References Atance, C. M., & Meltzoff, A. N. (2005). My future self: Young children's ability to anticipate and explain future states. Cognitive Development, 20(3), 341–361. http:// dx.doi.org/10.1111/cdep.12128. Baird, J. A., & Astington, J. W. (2004). The role of mental state understanding in the development of moral cognition and moral action. New Directions for Child and Adolescent Development, (103), 37–49. http://dx.doi.org/10.1002/cd.96. Blake, P. R., McAuliffe, K., & Warneken, F. (2014). The developmental origins of fairness: The knowledge–behavior gap. Trends in Cognitive Sciences, 18(11), 559–561. http://dx.doi.org/10.1016/j.tics.2014.08.003. Blasi, A. (1980). Bridging moral cognition and moral action: A critical review of the literature. Psychological Bulletin, 88(1), 1–45. http://dx.doi.org/10.1037/00332909.88.1.1. Blasi, A. (1983). Moral cognition and moral action: A theoretical perspective. Developmental Review, 3(2), 178–210. http://dx.doi.org/10.1016/0273-2297(83) 90029-1. Chernyak, N., & Kushnir, T. (2013). Giving preschoolers choice increases sharing behavior. Psychological Science, 24(10), 1971–1979 doi: 0956797613482335. Chernyak, N., & Kushnir, T. (2014). The self as a moral agent: Preschoolers behave morally but believe in the freedom to do otherwise. Journal of Cognition and Development, 15(3), 453–464. http://dx.doi.org/10.1080/15248372.2013.777843. Cushman, F., Knobe, J., & Sinnott-Armstrong, W. (2008). Moral appraisals affect doing/allowing judgments. Cognition, 108(1), 281–289. http://dx.doi.org/10.1016/j. cognition.2008.02.005. Cushman, F., Sheketoff, R., Wharton, S., & Carey, S. (2013). The development of intent-based moral judgment. Cognition, 127(1), 6–21. http://dx.doi.org/10.1016/j. cognition.2012.11.008. Cushman, F., Young, L., & Hauser, M. (2006). The role of conscious reasoning and intuition in moral judgment testing three principles of harm. Psychological Science, 17(12), 1082–1089. http://dx.doi.org/10.1111/j.1467-9280.2006.01834.x. Dunn, J., Cutting, A. L., & Demetriou, H. (2000). Moral sensibility, understanding others, and children's friendship interactions in the preschool period. British Journal of Developmental Psychology, 18(2), 159–177. http://dx.doi.org/10.1348/026151000165625. Eisenberg, N., & Fabes, R. A. (1999). Emotion, emotion-related regulation, and quality of socioemotional functioning. In L. Balter, & C. S. Tamis-LeMonda (Eds.). Child psychology: A handbook of contemporary issues (pp. 318–335). New York, NY: Psychology Press. Foot, P. (1978). The problem of abortion and negative and positive duty: A reply to James LeRoy Smith. The Journal of Medicine and Philosophy, 3(3), 253–255. http:// dx.doi.org/10.1093/jmp/3.3.253. Gall, S. N. L. (1985). Help-seeking behavior in learning. Review of Research in Education, 12, 55–90. Gilligan, C. (1987). Moral orientation and moral development. In E. V. Kittay, & D. T. Meyers (Eds.). Women and moral theory (pp. 19–23). Landham, MD: Rowman & Littlefield.

14


J. Prabhakar et al.

Greene, J. D., Morelli, S. A., Lowenberg, K., Nystrom, L. E., & Cohen, J. D. (2008). Cognitive load selectively interferes with utilitarian moral judgment. Cognition, 107(3), 1144–1154. http://dx.doi.org/10.1016/j.cognition.2007.11.004. Hepach, R., Vaish, A., & Tomasello, M. (2012). Young children are intrinsically motivated to see others helped. Psychological Science, 23(9), 967–972. http://dx.doi. org/10.1177/0956797612440571. Hudson, J. A., Shapiro, L. R., & Sosa, B. B. (1995). Planning in the real world: Preschool children's scripts and plans for familiar events. Child Development, 66(4), 984–998. http://dx.doi.org/10.1111/j.1467-8624.1995.tb00917.x. Kushnir, T., Wellman, H. M., & Chernyak, N. (2009). Preschoolers’ understanding of freedom of choice. In N. Taatgen, & H. van Rijn (Eds.). Proceedings of the 31st annual conference of the cognitive science society (pp. 87–92). Austin, TX: Cognitive Science Society. Lane, J. D., Wellman, H. M., Olson, S. L., LaBounty, J., & Kerr, D. C. (2010). Theory of mind and emotion understanding predict moral development in early childhood. British Journal of Developmental Psychology, 28(4), 871–889. http://dx.doi.org/10.1348/026151009X483056. Leslie, A. M., Mallon, R., & DiCorcia, J. A. (2006). Transgressors, victims, and crybabies: Is basic moral judgment spared in autism? Social Neuroscience, 1, 270–283. http://dx.doi.org/10.1080/17470910600992197. Margoni, F., & Surian, L. (2016). Explaining the U-shaped development of intent-based moral judgments. Frontiers in Psychology, 7. http://dx.doi.org/10.3389/fpsyg. 2016.00219. Margoni, F., & Surian, L. (2017). Children’s intention-based moral judgments of helping agents. Cognitive Development, 41, 46–64. http://dx.doi.org/10.1016/j.cogdev. 2016.12.001. McCrink, K., Bloom, P., & Santos, L. R. (2010). Children's and adults' judgments of equitable resource distributions. Developmental Science, 13(1), 37–45. http://dx.doi. org/10.1111/j.1467-7687.2009.00859.x. Miller, C., Starmans, C., Fugelsang, J., & Friedman, O. (2016). It's personal: The effect of personal value on utilitarian moral judgments. Judgment and Decision-Making, 11(4), 326–331. Nucci, L. P. (1985). Children's conceptions of morality, societal convention, and religious prescription. In C. G. Harding (Ed.). Moral dilemmas: Philosophical and psychological issues in the development of moral reasoning (pp. 137–174). Chicago, Illinois: Precedent Publishing Inc. Núñez, M., & Harris, P. L. (1998). Psychological and deontic concepts: Separate domains or intimate connection? Mind & Language, 13(2), 153–170. http://dx.doi.org/ 10.1111/1468- 0017.00071. Pellizzoni, S., Siegal, M., & Surian, L. (2009). Foreknowledge, caring, and the side-effect effect in young children. Developmental Psychology, 45(1), 289–295. http://dx. doi.org/10.1037/a0014165. Pellizzoni, S., Siegal, M., & Surian, L. (2010). The contact principle and utilitarian moral judgments in young children. Developmental Science, 13(2), 265–270. http:// dx.doi.org/10.1111/j.1467-7687.2009.00851.x. Prabhakar, J., & Hudson, J. A. (2014). The development of future thinking: Young children’s ability to construct event sequences to achieve future goals. Journal of Experimental Child Psychology, 127, 95–109. http://dx.doi.org/10.1016/j.jecp.2014.02.004. Shaw, A., & Olson, K. R. (2012). Children discard a resource to avoid inequity. Journal of Experimental Psychology: General, 141(2), 382–395. http://dx.doi.org/10. 1037/a0025907. Sheskin, M., Bloom, P., & Wynn, K. (2014). Anti-equality: Social comparison in young children. Cognition, 130(2), 152–156. http://dx.doi.org/10.1016/j.cognition. 2013.10.008. Siegal, M., & Peterson, C. C. (1998). Preschoolers' understanding of lies and innocent and negligent mistakes. Developmental Psychology, 34(2), 332. http://dx.doi.org/ 10.1037/0012-1649.34.2.332. Smetana, J. G. (1981). Preschool children's conceptions of moral and social rules. Child Development, 52(4), 1333–1336. http://dx.doi.org/10.2307/1129527. Smetana, J. G. (1983). Social-cognitive development: Domain distinctions and coordinations. Developmental Review, 3(2), 131–147. http://dx.doi.org/10.1016/02732297(83)90027-8. Smetana, J. G. (1985). Preschool children's conceptions of transgressions: Effects of varying moral and conventional domain-related attributes. Developmental Psychology, 21(1), 18. http://dx.doi.org/10.1037/0012-1649.21.1.18. Smetana, J. G., Jambon, M., Conry-Murray, C., & Sturge-Apple, M. L. (2012). Reciprocal associations between young children's developing moral judgments and theory of mind. Developmental Psychology, 48(4), 1144. http://dx.doi.org/10.1037/a0025891. Smith, C. E., Blake, P. R., & Harris, P. L. (2013). I should but I won’t: Why young children endorse norms of fair sharing but do not follow them. PloS one, 8(3), 1–11. http://dx.doi.org/10.1371/journal.pone.0059510. Surber, C. F. (1977). Development processes in social inference: Averaging of intentions and consequences in moral judgment. Developmental Psychology, 13(6), 654. http://dx.doi.org/10.1037/0012-1649.13.6.654. Tasimi, A., & Wynn, K. (2016). Costly rejection of wrongdoers by infants and children. Cognition, 151, 76–79. http://dx.doi.org/10.1016/j.cognition.2016.03.004. Thomson, J. J. (1986). Rights, restitution, and risk: Essays. Cambridge, Mass: Harvard University Press. Turiel, E. (1978). Social regulations and domains of social concepts. New Directions for Child and Adolescent Development, 1, 45–74. http://dx.doi.org/10.1002/cd. 23219780105. Vaish, A., Carpenter, M., & Tomasello, M. (2010). Young children selectively avoid helping people with harmful intentions. Child Development, 81(6), 1661–1669. http://dx.doi.org/10.1111/j.1467-8624.2010.01500.x. Warneken, F., & Tomasello, M. (2006). Altruistic helping in human infants and young chimpanzees. Science, 311(5765), 1301–1303 doi: 0.1126/science.1121448. Warneken, F., & Tomasello, M. (2007). Helping and cooperation at 14 months of age. Infancy, 11(3), 271–294. http://dx.doi.org/10.1080/15250000701310389. Wellman, H. M., Larkey, C., & Somerville, S. C. (1979). The early development of moral criteria. Child Development, 50(3), 869–873. http://dx.doi.org/10.2307/ 1128956. Zelazo, P. D., Helwig, C. C., & Lau, A. (1996). Intention, act, and outcome in behavioral prediction and moral judgment. Child Development, 67(5), 2478–2492. http:// dx.doi.org/10.1111/j.1467-8624.1996.tb01869.x.

15