Exploring the effects of similarity in basic needs on moral judgment of characters

ABSTRACT

While affective disposition theory (ADT) claims that moral judgment of a character is the key mechanism underlying enjoyment, we often enjoy stories where characters violate our moral standards (morally ambiguous characters, or MACs). Initial research suggests that similarity between MACs and viewers (e.g. sharing ethnicity) may increase moral approval. However, the effect of internal similarity (similarity in higher-order human qualities) on moral judgment of MACs is underexplored. Humans share three basic psychological needs but are motivated by these needs to different degrees. Thus, similarity in degree of sensitivity to a given need is an internal quality that participants can share with MACs. The effect of similarity in basic needs on moral judgment of MACs is examined in a posttest-only experiment with 185 participants. Similarity in basic needs had a small effect on moral judgment of MACs as people, although it had no effect on judgment of a MAC’s behavior.

KEYWORDS:

• Affective disposition theory
• morally ambiguous characters
• basic psychological need theory
• similarity
• entertainment

The leading theory of narrative enjoyment (affective disposition theory, or ADT) claims that moral judgment of a character (approving or disapproving of their motives and actions) is the key mechanism that explains enjoyment. However, sometimes audiences enjoy stories where characters violate their moral standards. Identifying the mechanisms that allow us to approve of these morally ambiguous characters (MACs) in spite of their moral violations is thus a key goal in the ADT literature.

Initial research suggests that similarity between MACs and viewers may increase moral approval of characters (Tsay & Krakowiak, 2011). Similarity is thought to increase identification with the character (Cohen & Tal-Or, 2017): the experience of becoming absorbed in a character’s cognitions, emotions, and goals (Cohen, 2001). Identification should then increase moral approval of the character and their questionable actions (Janicke & Raney, 2018). However, Cohen and Tal-Or (2017) have noted that most studies on the relationship between similarity and identification have focused on external similarity (similarity in age, ethnicity, income level, etc.). Little extant research has explored the effects of internal similarity on identification (i.e., similarity in higher-order human qualities, such as personality; Cohen & Hershman-Shitrit, 2015). Moreover, no research to date has explored the effects of internal similarity on moral judgment of MACs (via the mediator of identification)—despite the demonstrated impact of higher-order factors such as motivation and values on moral judgment (e.g., Tamborini et al., 2021). Thus, the present research examines the effects of internal similarity on moral judgment of MACs (via identification).

There are, of course, many possible dimensions of internal similarity. For example, Cohen and Hershman-Shitrit (2015) operationalized internal similarity in terms of personality, studying the effects of similarity in Big Five traits (Costa & McCrae, 1992) on identification. Similarity in personality has also been found to influence identification and empathy toward MACs specifically; for example, Kjeldgaard-Christiansen et al. (2021) found that participants’ levels of the “dark triad” of personality traits (narcissism, Machiavellianism, psychopathy) positively correlated with villain sympathy, identification, and empathy (see also Greenwood et al., 2021). However, in the context of personality as a predictor of identification, Cohen and Hershman-Shitrit (2015) findings were not robust, suggesting that other dimensions of internal similarity may be relatively more important.

Perceptions of a character’s basic psychological needs have been shown to influence moral judgment of that character (e.g., Tamborini et al., 2021). According to self-determination theory (SDT, Ryan & Deci, 2000), humans are driven to pursue autonomy (feeling that we are the source of our own actions), competence (feeling that we have mastery or efficacy in the goals we pursue), and relatedness (feeling that we have meaningful connections with others). These are basic psychological needs because the absence of these factors is characterized by suffering, whereas their presence is characterized by thriving. Therefore, basic needs exist in all humans, but importantly, the relative importance of each need varies by individual. Thus, we may be more or less similar to others on this basic dimension. Building on this research, similarity in basic needs is chosen as a dimension of internal similarity for examination in the present study.

Specifically, we predict that the closer the level of a participant’s need for autonomy, competence, or relatedness is to the perceived level of the character’s need for autonomy, competence, or relatedness, the more positively the participant will judge both the character and the character’s behavior (i.e., they will think better of the character as a person, and they will be more likely to excuse the character’s morally questionable actions). In other words:

H1(a-b).

Similarity in basic needs to the MAC will positively predict positive moral judgment of (a) the MAC as a person and (b) the MAC’s behavior.

The proposed mechanism underlying this effect is identification. In line with Cohen and Hershman-Shitrit (2015), the greater the similarity between the character and the participant in terms of the relative importance of the three basic needs, the more the participant will identify with the character:

H2.

Similarity in basic needs will positively predict identification.

Following the predictions of Janicke and Raney (2018), greater identification with the MAC should in turn predict higher levels of moral approval:

H3(a-b).

Identification will positively predict moral judgment of (a) the MAC as a person and (b) the MAC’s behavior.

If these predictions hold true, this would indicate that:

H4(a-b).

Identification will mediate the effect of similarity in basic needs on moral judgment of (a) the MAC as a person and (b) the MAC’s behavior.

In short, we predict that similarity in basic needs should increase moral approval of MACs via the mechanism of identification.

Method

Participants

Both undergraduates at a private southeastern university in the U.S. (n = 141), and adults of various ages across the U.S. (n = 126) were recruited. Eighty-four participants were dropped for partial completion or attention check failure, leaving 185 participants ranging in age from 18 to 81 (see Table S3 in Appendix C of the online supplemental materials for demographic information; https://osf.io/zxu8v/?view_only=d16a77b8087d417fa859c5f5fd0791f9).

Undergraduates participated for extra credit. Adult participants were recruited via snowball sampling through social media and professional networks in the southeastern and northwestern U.S. Because differences in judgment between conditions—rather than participants’ actual moral judgments—were the outcome of interest, sample homogeneity was considered acceptable.

Experimental design

Similarity in basic needs was operationally defined as the difference between the importance of each basic need to the participant and the participant’s perception of the importance of each basic need to the MAC. In other words, if a participant sees the MAC as having a particularly strong need for relatedness, and the participant also has a strong need for relatedness, then the character and the participant are highly similar in that need.

To manipulate similarity in basic needs, then, a method of varying participants’ perceptions of the importance of each need to the MAC was required. Conceptually, the fact that we all share basic needs does not entail that we notice these needs in others; thus, if the shared drives underlying a character’s actions are not brought to our attention, the underlying similarity may not influence our judgment of the character. Presumably, then, if there is less information about the MAC’s basic needs available to the participant, then the participant will see those needs as less important to the MAC. Likewise, if there is more information about the MAC’s needs available to the participant, then the participant will see the needs as more important to the MAC. Based on this logic, to manipulate emphasis or deemphasis on the MAC’s basic needs, a short story with two characters (a MAC and a victim of the MAC’s morally questionable act) was written for the project. Information about the basic psychological needs of the MAC was either presented or withheld, resulting in either emphasis or deemphasis of the MAC’s needs.

This study is part of a larger program of research for which information about the victim’s basic needs was also presented or withheld, resulting in a 2 (MAC needs [present or absent]) × 2 (victim needs [present or absent]) research design. For the present study, only the main effects of the presence or absence of the MAC’s basic needs were examined. Both conditions where the needs of the MAC were present were collapsed into one, and both conditions where those needs were absent were collapsed into one, resulting in two conditions (MAC needs [present or absent]).

Procedure

Participants received a link to the consent form, stimulus material, and instrument via e-mail. They were randomly assigned to read 1 of the 4 versions of the story, after which they filled out the questionnaire. For number of participants in each condition and average duration, see Table S4 in Appendix C of the OSF.

Stimulus development

The story features a MAC harming a victim. For the stimuli and manipulations, see Appendix A of the online supplemental materials. Three pilot studies were conducted to ensure that the four versions successfully manipulated the salience of each character’s basic needs to readers, thereby validating the logic of manipulating similarity in basic needs by including or excluding information about the basic needs of the MAC. To ensure that the three needs were manipulated individually, the manipulation was adjusted and further pilot studies conducted until the manipulation predicted significant differences in perceived importance for relatedness and competence, and a trend in the predicted direction for autonomy (see results for Pilot Study 3 in Appendix B of the OSF). In the main study sample, similarity in competence and relatedness were significantly higher when information about protagonist basic needs was included than when it was not (see Table S5 in Appendix C of the OSF).

In the stimulus, we chose to sacrifice control for higher ecological validity. The manipulated sentences attempted to convey basic needs and threats to those needs in a way that could conceivably appear in a real short story. While the manipulation would have been cleaner if we had used more explicit, on-the-nose statements (e.g., “she needed to feel connection to other people, and was afraid that she would not”), the ecological validity of the study would have dropped significantly.

To avoid contaminating the manipulation, no filler information was added. It should be noted that this imbalance may have influenced other predictors of identification (Cohen & Tal-Or, 2017), introducing potential confounds.

Measures

Internal Similarity

Following Cohen and Hershman-Shitrit’s (2015) precedent, similarity in basic needs was measured by comparing participants’ evaluations of the importance of basic needs for themselves to the perceived importance of the basic needs to the MAC. Importance of basic needs to either the self or the MAC were measured using La Guardia and colleagues’ (2000) nine-item scale of basic psychological needs. Three items assess the strength of the needs for autonomy, competence, and relatedness, respectively, and were adapted to read as statements of importance rather than of fact. Participants filled out the scale for both themselves and for the MAC. The order was randomized; self-evaluation and character evaluation had an equal chance of appearing first. The difference between each participant’s own score and the scores attributed to the MAC was computed. To change the score into a measure of similarity (rather than a measure of difference), scores were reverse coded. Each subscale was moderately reliable: α = .67 for autonomy (M = 5.09, SD = 1.20), α = .68 for competence (M = 5.82, SD = .95), and α = .59 for relatedness (M = 5.51, SD = .97), but reliabilities were higher when the items were computed as differences rather than similarities: α = .71 for autonomy, α = .68 for competence, and α = .71 for relatedness. This is likely because the difference score accounted for the direction of the difference—whether the perceived importance of a need to the character was higher or lower than the importance of that need to the participant. Similarity was computed as the absolute value of the difference, independent of direction.

Identification

Identification was measured using Cohen’s (2001) ten-item scale (see also Janicke & Raney, 2018; Tsay & Krakowiak, 2011). The items, measured on a five-point Likert scale, assess viewers or readers’ absorption in the goals, cognitions, and emotions of the MAC (α = .83 M = 3.25; SD = .73).

Moral Judgment of Behavior

Moral judgment of behavior was measured using the seven items from Raney et al (2009; see also Sanders & Tsay-Vogel, 2016, p. 240). One item was removed because it was designed to reflect attitudes toward vigilantism, which was not relevant to the current stimuli (α = .84 M = 2.15, SD = .79).

Moral Judgment of Character

To measure moral judgment of character, 16 items were adapted from Goodwin et al. (2014). Participants were asked to rate the MAC with respect to how much she possessed each of the high morality traits on the list on a Likert scale from 1 (Not at all) to 9 (Extremely). The scale was reliable (α = .77, M = 5.48; SD = .81).

Perceived Altruism

Prior studies indicate that the altruism or egoistism of the MAC’s motives influences audience moral judgments, and that the basic needs from self-determination theory are egoistic motives (Tamborini et al., 2021; Krakowiack & Tsay-Vogel, 2013). Thus, perceived altruism was assessed as a control variable. Two items were created (e.g., “How much do you think that Jane cares about the following people? (i.e., how much does she want the following people to be safe and happy)?”). On a scale from 1 (Ambivalent [does not care one way or the other]) to 5 (Cares a lot), participants rated the degree to which the MAC cared about her father and the victim. Ratings for her father and the victim were averaged together to create a score for altruism (M = 3.73, SD = 1.19).¹

Results

Hypothesis 1 (a-b)

Hypotheses 1—that similarity in basic needs to the MAC would positively predict positive moral judgment of (a) the MAC as a person and (b) the MAC’s behavior—was tested using three hierarchical regression analyses. Perceived altruism was controlled for on the first step. Similarity in basic needs was added on the second step (one need per analysis), and either moral judgment of person or moral judgment of behavior was added as the dependent variable (see Tables S6–8 in Appendix D of the OSF).

Only similarity in need for relatedness predicted moral judgment of the antihero as a person. Perceived altruism was significant on Step 1 (β = .31, p < .001), and similarity in relatedness was significant on Step 2 (β = .17, p < .05; see Table S8). Similarity in none of the basic needs predicted moral judgment of the antihero’s behavior, although similarity in relatedness approached significance (β = .16, p = .06; see Table S8) with perceived altruism significant on Step 1 (β = .20, p = .02). Thus, while similarity in relatedness influenced judgment of the MACs as a person, similarity in basic needs did not affect condemnation of the questionable behavior.

Hypothesis 2

To test the second hypothesis—that similarity in basic needs would positively predict identification—three linear regression analyses were conducted with similarity in each of the basic needs (respectively) as the predictor variable and identification as the outcome variable (see Tables S9–11 in Appendix D of the OSF). The effect of similarity in need for competence on identification was significant at p = .004 (β = .24; see Table S10). Thus, similarity to a character in one of the basic needs—but not the others—predicted identification.

Hypotheses 3 (a-b)

The third hypothesis was that identification with the MAC would positively predict (a) moral judgment of the MAC as a person, and (b) moral judgment of the MAC’s behavior. Perceived altruism was controlled for on the first step of the hierarchical regression, identification was added on the second step, and either moral judgment of person or moral judgment of behavior was added as the dependent variable (see Table S12). While Hypothesis 3a was not supported, the data were in the predicted direction and approached significance. On Step 1, perceived altruism was significant (β = .30, p < .001). On Step 2, identification approached significance (β = .16, p = .056), even after controlling for perceived altruism. This suggests that, with some qualifications, the more we identify with a morally ambiguous character, the more positive our judgment of that character as a person.

The relationship between identification with the MAC and moral judgment of the MAC’s behavior (H2b) was nonsignificant. While perceived altruism was significant on Step 1 (β = .20, p = .02), identification was not significant on Step 2 (β = .11, p = .22; see Table S12).

Hypothesis 4(a-b)

The fourth hypothesis predicted that identification would mediate the effect of similarity in basic needs on both types of moral judgment. Because similarity in relatedness also did not predict identification, and similarity in competence did not also predict moral judgment, no mediation effects were possible (see Figure 1). Thus, Hypotheses 4a and 4b were not supported.

Figure 1. Statistical model of the effect of similarity in basic needs on moral judgment.

Discussion

While similarity between a MAC and a viewer has been found to increase moral approval of the MAC (via identification), extant research has focused on external types of similarity (e.g., demographic similarity) rather than internal similarity (i.e., similarity in basic needs). In the present data, both identification and internal similarity in need for relatedness significantly predicted moral judgment of the character as a person; and internal similarity in need for competence significantly predicted identification. No other relationships were significant (see Figure 1). These findings offer four insights to the conversation. First, they suggest that certain mechanisms may affect our moral judgment of a MAC as a person independent of our judgment of their behavior. Both similarity in basic needs and identification with the MAC affected moral judgment of the MAC as a person, but not moral judgment of the MAC’s behavior. Perhaps, then, we can understand and care about a person, even if we disagree with what they do—and similarity in basic needs and identification can help to facilitate such understanding and care. Thus, future ADT research on internal similarity might focus on judgments of person over judgments of behavior.

Second, it is possible that the aspects of a character’s psychological makeup that facilitate identification may differ from the aspects that influence moral judgment. Only similarity in relatedness predicted moral judgment of the MAC as a person, and only similarity in competence predicted identification. On the one hand, this finding likely reflects the greater importance of competence and relatedness in the story itself—means for autonomy were not significantly different between conditions, while means for relatedness and competence were (see Table S5 in the supplemental materials). However, the complete lack of overlap in the needs that predicted identification and the needs that predicted moral judgment suggests that even through identification predicts moral judgment, the aspects of a character that influence identification are not necessarily the aspects of a character that influence moral judgment. This complicates the predicted relationship between identification and judgment (Janicke & Raney, 2018).

One explanation of these particular findings is that the importance of altruism in moral judgment extends to relatedness (Tamborini et al., 2021): altruism is the desire for the wellbeing of another, and relatedness is the need for someone to desire our wellbeing. The need for competence, by contrast, may be uniquely relevant to identification because the storyline prominently featured a work setting and absorption in the experience of the character tracks with the types of situations that are emphasized.² Either way, more research triangulating the type of character qualities that lead to identification, the type that influence judgment, and the relationship between identification and moral judgment is warranted. Future research could explore potential conceptual connections between altruism and relatedness, and could experimentally examine the extent to which the basic need sharing that predicts identification is dependent on story setting.

Third, if internal similarity influences moral judgment, it is likely not via identification as a mediator. Although internal similarity was conceptualized as a predictor of identification (e.g., Cohen & Tal-Or, 2017), identification did not mediate the effects of similarity in any of the basic needs on moral judgment of the MAC as a person. Because similarity in competence predicted identification and similarity in relatedness predicted judgment of person, however, it is possible that internal similarity in basic needs may predict both identification and moral judgment of person—only, independent of one another. This improves our understanding of morally ambiguous characters by highlighting a potential mechanism that has received relatively little attention. While identification is well established as a mediator in the ADT literature on morally ambiguous characters, similarity has been explored only minimally. This construct may be worthy of study, not just as a predictor of identification, but as a predictor of moral judgment in its own right.

An important qualification, however, is that while similarity is theorized to predict identification, the influence may be bidirectional. This issue is common in research on similarity and identification, and can be addressed by establishing similarity prior to narrative exposure (see Cohen & Tal-Or, 2017, p. 141). The present research prioritized ecological validity (which would have been disrupted by learning about the characters ahead of time); however, this choice makes it difficult to assess potential bidirectional influence.

Fourth, while internal similarity in certain basic needs may influence moral judgment of characters, it is not likely the most powerful predictor. Compared with the effects of altruism, similarity in basic needs explained relatively little of the variance in moral judgment. While similarity in basic needs may be one of the factors that influence our judgment of MACs, there are likely other factors (e.g., altruism) that explain relatively more of the variance.

Explaining how narratives facilitate acceptance of people we normally disagree with is relevant for understanding entertainment processes, but it is also relevant for understanding how narrative media can help us bridge real-world divides (e.g., those on the other side of political or cultural lines). Even when we come from different backgrounds and perspectives, there are things we share as members of the human species. Narrative media can give us privileged access to the core, higher-order needs of others. The present research takes us one step closer to understanding how this process works.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08824096.2024.2358798.

Additional information

Notes on contributors

Catherine A. Marple

Catherine A. Marple (Wake Forest University) is a doctoral student in the College of Communication Arts & Sciences at Michigan State University. She studies the processes whereby popular entertainment narratives influence the memory structures that shape perceptions of situations and people. Applications include using narrative entertainment (1) as a mental health tool and (2) to facilitate empathy across social divides.

Marina Krcmar

Marina Krcmar (University of Wisconsin-Madison) is a Professor of Communication at Wake Forest University. Her research is focused on children, families and media use, with more specific attention on well-being. More recently, she has focused on moral reasoning in children and adolescents and how this is related to media use.

Notes

1. Perceived vulnerability of both the MAC and the victim was also measured for the larger program of research; however, these variables were not used in the current study.

2. While interpreting these results, two important limitations should be considered: perceived need for autonomy was not significant in either the pilot studies or the main study; and the reliabilities for each need sharing subscale were relatively low. Thus, autonomy may have featured in the relationship between identification and moral judgment if it had been emphasized more in the stimuli; likewise, the differences in how each need affected identification versus judgment may have been less dramatic with more robust reliability.