How the incidental emotions of anger and guilt influence our judgments

Abstract

Incidental emotions, unrelated to the decision at hand, can significantly influence judgments and choices. The aim of the current study is to investigate the role of the trait and state emotions—specifically anger and guilt—in affecting evaluative judgments. A sample of 204 male participants was randomly assigned to one of the three induced-state emotion conditions (anger, guilt, or neutral). Participants read a forensic report describing an argument that ended with the offender’s decision to violently strike the opponent. The participants were asked to identify with the offender as much as possible and to judge the immoral and risky choice in terms of causality, responsibility, predictability, intentionality, severity, counterfactual thoughts, reversibility willingness, and anticipated emotions. Results showed that trait anger was linked to external causal attribution, judgment of others’ responsibility, and anticipated negative and positive emotions, while state anger was associated with less internal attribution and fewer upward counterfactual thoughts. Trait guilt was linked to a desire to reverse one’s decision and anticipated moral emotions, while state guilt correlated with internal attribution and judgments of severity. No interaction between trait and state emotions was found. These findings deepen our understanding of the role of incidental emotions in judgment and decision-making.

Introduction

There is no doubt that emotions significantly influence our daily judgments and choices across a wide range of life domains^1,2. Their impact on judgment and decision-making (JDM) can vary depending on when and how they arise during the decision process³.

In particular, emotions experienced at the moment of decision—referred to as immediate, decision-time⁴, or current emotions⁵—have been shown to focus on the act of deciding itself (option-focused), exerting their influence in two principal ways: by providing information and by directing attention^6,7. Individuals may use their current emotional state as a cue when evaluating people, objects, events, or situations, while also having their attention guided toward information perceived as relevant to explaining their emotional experience. A substantial body of research on emotion and risky decision-making is grounded in this framework^8–11.

Within this category of immediate emotions, a key distinction is drawn between integral and incidental emotions. Integral emotions are directly elicited by the decision dilemma itself and are often described as the “emotional cost of choosing” (e.g., anxiety, uncertainty, fear of loss)¹². Because they are an inherent part of the decision process, they are expected and typically taken into account during evaluation¹³.

In contrast, incidental emotions—also referred to as background emotions¹⁴—are affective states experienced at the time of judgment but unrelated to the decision or situation being evaluated^5,12. These emotions may be either dispositional, reflecting a stable emotional tendency, or situational, arising from a specific, unrelated event. Although both types of incidental emotions are irrelevant to the decision object, they can nonetheless significantly influence how individuals judge or choose.

Beyond a dimensional approach: the role of discrete emotions in JDM

Traditional models of emotion often rely on general dimensions such as valence (positive vs. negative) and arousal (low vs. high) to explain how emotions influence JDM. However, several studies have shown that a dimensional approach is insufficient to fully capture the complexity of emotional experience and its impact on decision outcomes (e.g.^7,15). “…A two-dimensional model seems inadequate for describing emotional experiences. Anger, sadness, and disgust are all forms of negative affect, and arousal does not capture all of the differences among them …. A more detailed approach is required to understand relationships between emotions and decisions” (¹⁶, p. 454).

In response to these limitations, recent research has shifted toward the study of discrete emotions, focusing on how specific affective states—despite sharing the same valence—can exert qualitatively different effects on cognitive processing and decision-making. For instance, both anger and fear are negatively valenced emotions, yet they have been shown to lead to contrasting behavioral tendencies: while anger typically fosters approach and action, fear promotes caution and avoidance.

Building on earlier theories such as Schwarz and Clore’s affect-as-information model⁷—which proposed that people rely on their momentary emotional state as a source of information when forming judgments—a particularly influential framework in this line of research is the Appraisal Tendency Framework (ATF), developed by Lerner and Keltner^9,12. According to the ATF, each emotion is characterized by a distinct pattern of cognitive appraisal dimensions—such as certainty, pleasantness, attentional activity, control, anticipated effort, and responsibility—that guide how individuals evaluate, judge, and decide, beyond general emotional valence^5,12. The ATF is grounded in three core assumptions: (a) discrete emotions differ across specific cognitive appraisal dimensions (e.g.¹⁷); (b) emotions trigger automatic physiological, behavioural, and experiential responses that help individuals respond to challenges (e.g.¹⁸); and (c) emotions act as motivational forces, based on their underlying appraisal themes—such as the “action tendency” described by Frijda¹⁹, or the “core-relational theme” by Lazarus²⁰.

Anger, guilt, and JDM

Following the affect-as-information theory and the ATF, the present study investigates the impact of two specific, discrete incidental emotions—anger and guilt—on JDM. Although both emotions share a negative valence, they differ markedly in their cognitive appraisal patterns and behavioral outcomes, especially in contexts involving risky or norm-violating behaviors, such as criminal actions.

Anger is traditionally classified as one of the six basic emotions²¹, but it can also be understood as a moral emotion, typically triggered by perceived violations of social norms or injustice²². It usually arises when an obstacle prevents goal attainment, acquiring moral relevance when it motivates individuals to restore justice, for instance by retaliating against those perceived as acting immorally. According to Lerner and Keltner¹², anger is associated with appraisals of high certainty and strong individual control, which foster optimistic evaluations of risk and personal success. This often leads to an underestimation of potential negative outcomes and a tendency to prefer risky options, even in contexts unrelated to the initial cause of anger (e.g.^23,24). Anger has been shown to foster defensive optimism and is characterized by a heuristic and rapid processing style²⁵, which may potentially reduce deliberative reflection on alternative courses of action compared to other emotions typically involved in moral situations²⁶, also because anger leads to external causal attributions, with individuals assigning blame to others for negative events²⁷. Beyond situational effects, evidence suggests that individuals with high trait anger are chronically biased toward interpreting situations in ways that promote external blame and moral condemnation, reflecting stable appraisal patterns that shape their social evaluations across contexts²⁸.

Guilt, in contrast, is a self-conscious emotion that arises when individuals recognize that they have violated internalized social or cultural standards. This realization often implies that one should have acted in a more correct, altruistic, or socially acceptable way^20,29. Guilt is rooted in internal appraisals of responsibility²⁷ and typically entails counterfactual thinking—that is, contemplating alternative possibilities to past events: what might have been³⁰. It motivates reparative behaviors aimed at re-establishing justice, even at personal cost. From an interpersonal perspective, guilt drives individuals to rectify errors, prevent future harm, and atone for transgressions³¹. In decision-making contexts, guilt heightens risk perception, increasing both the perceived probability and severity of negative outcomes, while reducing risk-taking tendencies and redirecting attention toward specific outcomes rather than others^32,33. Individuals induced to feel guilt tend to avoid risk and prefer safer options³⁴, and guilt-prone individuals are less likely to engage in risky or harmful behaviors, such as substance abuse³⁵. Beyond momentary effects, individuals high in trait guilt show a consistent propensity for reparative behaviors, partly driven by heightened other-oriented empathy and a motivation to alleviate both their own and others’ distress, shaping their moral responses across diverse contexts³⁶.

Beyond their individual effects, there is growing evidence that trait emotional dispositions shape state affective experiences in daily life. In line with integrative and dynamic models of affective functioning, trait emotions influence not only the general propensity to experience certain affective states, but also modulate their intensity, variability, and regulation across real-world contexts^37–39. Rather than functioning as static predispositions, emotional traits dynamically interact with situational triggers, shaping affective responses as they unfold. Based on this framework, higher levels of specific emotional traits could amplify the cognitive and motivational tendencies typically associated with their corresponding state⁴⁰.

Currently, there is no comprehensive understanding of how emotions influence decision-making in moral situations, although existing evidence shows that biases caused by both emotional states and traits are associated with the likelihood of making risky decisions⁴¹. While many studies have examined state or trait emotions, often in isolation and in non-moral contexts, clarifying whether these emotional components act independently or synergistically is critical for refining theoretical models of emotional influence on moral judgment, particularly regarding how individuals evaluate responsibility, predictability, and intentionality. This knowledge could have a substantial impact on the analysis of moral scenarios, with potential implications for legal decision-making (e.g., in the evaluation of norm-violating behavior), and the development of interventions aimed at fostering fairer and more reflective judgments in morally controversial situations.

Building on this theoretical background, the present study focuses specifically on anger and guilt, two emotions characterized by distinct cognitive appraisal patterns and motivational tendencies. Despite the growing body of research on incidental emotions and cognition, no study to date has directly compared anger and guilt across both trait and state levels within the context of moral judgment, nor explored their potential interaction. By examining their effects at both the trait and state levels, we aim to clarify whether these emotional influences converge or diverge in shaping moral judgments. In doing so, we contribute novel insights into the psychological mechanisms underlying evaluations moral scenarios.

Aims and hypotheses

The present study aims to explore the role of the trait and state emotions of anger and guilt in impacting judgments. Incidental emotions of anger and guilt were manipulated by asking people to provide anger-related or guilt-related autobiographical memory, respectively (vs. a neutral-control condition).

The participants were then asked to read a story about a quarrel that ends with an offender’s decision to violently strike an opponent. The story is taken from a real forensic case of personal injury defense⁴². The story clearly operationalizes the immoral choice made by the offender, Marco, to violently assault his classmate, thereby committing a criminal act. At the same time, however, the criminal event appears to be the consequence of a series of previous discussions and provocations between the two boys. This history of previous conflict makes the judgment of responsibility and intentionality of the act unclear, as well as the emotions involved in the events, which surround the moral frame of anger (linked to aggressive behavior) and guilt (linked to criminal choice).

The participants were asked to identify with the offender as much as possible and to judge the risky and immoral choice made by Marco to punch his classmate in terms of causality attribution and responsibility, predictability, intentionality, severity, counterfactual thoughts, reversibility willingness, and anticipated emotions. The participants’ trait anger and trait guilt were also assessed as incidental emotions, along with the manipulated emotional states. Drawing upon the above-cited literature, we formulate the following hypotheses.

H1

Compared with guilt and neutral conditions, the incidental emotion of anger (both trait and state) is expected to lead individuals to (i) perform more external attributions as causes of the crime; (ii) attribute greater responsibility for the crime to others; (iii) provide higher judgments of predictability and intentionality, and lower judgments of severity of the outcomes; (iv) generate more downward and fewer upward counterfactual thoughts; (v) show less willingness to reverse their choice; and (vi) anticipate fewer moral and negative emotions, and more positive emotions.

H2

Compared with anger and neutral conditions, the incidental emotion of guilt (both trait and state) is expected to lead individuals to (i) perform more internal attributions as causes of the crime; (ii) attribute greater responsibility for the crime to oneself; (iii) provide lower judgments of predictability and intentionality, and higher judgments of severity of the outcomes; (iv) generate more upward and fewer downward counterfactual thoughts (v) show greater willingness to reverse their own choice; and (vi) anticipate more moral and negative emotions, and fewer positive emotions.

H3

Trait and state emotions (both anger and guilt) are expected to interact in a congruent manner, such that the presence of a high trait level (anger or guilt) will amplify the influence of the corresponding state emotion on moral judgments. This congruence between trait and state emotions is expected to influence moral judgments in line with the cognitive and behavioral tendencies typically associated with each specific discrete emotion.

Methods

Design and participants

The current study adopts a one-way design with State Emotion (Anger vs. Guilt vs. Neutral) as a between-subjects variable. The dependent variables are indices of moral judgments: causality and responsibility attributions; predictability; intentionality; severity; upward and downward counterfactual thoughts; reversibility; and anticipated moral, negative, and positive emotions. The study was given ethical approval by the Ethics Committee of the Department of Educational Science, Psychology, and Communication and was executed according to the Declaration of Helsinki (No. ET-19-13). The participants provided signed informed consent before they participated in the experiment.

We ran a priori power analysis via G*Power⁴³ for one-way ANCOVA between-subjects with three groups (IV₁ state emotions), two covariates (IV₂ and IV₃ trait emotions), a power of 0.80, a medium given effect size of f = 0.25, an α = 0.05, and number df = 5 (one categorical IV₁, two continuous IV₂ and IV₃, two two-way interactions IV₁*IV₂ and IV₁*IV₃). This analysis included 211 participants. Because some participants might not comply with the instructions, we initially oversampled to 246 participants. After an initial screening, forty-two participants were excluded from the study since they did not correctly understand the instructions for state emotion induction. A final sample of 204 participants was recruited and randomly assigned to one of the three conditions (anger condition, n = 68; guilt condition, n = 68; neutral condition, n = 68). The average age of the sample was 27.7 years (SD = 10.09, range = 18–62), with an average level of education of 14.90 years (SD = 3.07; range = 8–21). The sample was entirely composed of males to guarantee accurate identification with the protagonist of the crime story; being a woman would certainly have made the identification and consequently the formulation of the judgment less plausible.

Measure and procedure

Trait emotion assessment phase

To measure participants’ dispositional tendency to experience anger, we administered the Italian version of the State Trait Anger Expression Inventory (STAXI⁴⁴). The STAXI comprises forty-four 4-point items (1 = “not at all”; 4 = “very much”) measuring self-reported anger experience and expression. Although the inventory provides eight different scores, we considered only three trait-related subscales relevant to our aims: (a) Trait Anger (T_Ax)—measuring the general tendency to experience anger (Cronbach’s alpha = 0.80); (b) Trait-Anger-Temperament (T_Ax_T)—reflecting a general disposition to feel unprovoked anger (Cronbach’s alpha = 0.80); (c) Trait-Anger-Response (T_Ax_R)—the tendency to experience anger in response to provocation or unfair treatment (Cronbach’s alpha = 0.69). The other subscales (state anger, anger expression, anger-in, anger-out, and anger control) were excluded as they were not directly relevant to the study’s focus on dispositional anger.

To measure participants’ dispositional proneness to guilt, we administered the Test of Self-Conscious Affect—Version 3 (TOSCA-3^45,46). The TOSCA-3 consists of 16 everyday scenarios (11 negative and 5 positive), each followed by responses participants rate on a 5-point Likert scale (1 = “very unlikely”; 5 = “very likely”). Although the TOSCA-3 provides six scores (proneness to shame, proneness to guilt, externalization of blame, detachment/unconcern, pride in oneself, and pride in one’s behavior), we used only the proneness to guilt subscale (16 items; Cronbach’s alpha = 0.71). The other subscales were excluded as they are not relevant to the study’s aim.

Pre-manipulation phase (test)

The emotions experienced before participating in the study were assessed (emotions at Test). The participants were asked to indicate, on six 11-point Likert scales (0 = “not at all”; 10 = “completely”) how intensely they felt in that moment: (a) guilt; (b) anger; (c) frustration; (d) helplessness; (e) unfairness; and (f) shame.

State emotion manipulation phase

Following the procedure adopted by Schwarz and Clore⁷, Gangemi and Mancini⁴⁷, Gangemi, Mancini and van den Hout⁴⁰, incidental state emotion was manipulated by having participants provide anger-related (anger condition), guilt-related (guilt condition), or neutral (neutral-control condition) autobiographical memory. The specific instructions were as follows: ‘Please think of a specific event from your personal past, a specific, datable episode in which you experienced very strong guilt/anger. This memory can refer to any type of experience in your life, which may have happened from early childhood to yesterday. What is essential is that it is the memory of an event that caused much guilt/anger both immediately and in the following days, months or even years. It could be an event that still gives you guilt/anger today. Think about the memory long enough for you to have a sense of completeness, remembering the event in its entirety and in its full intensity of guilt/anger. Please describe the memory in as much detail as possible, what happened (e.g., suffering an injustice/wrong/betrayal or stealing something/offending someone) when (e.g., January 2019), who you were with (if there was anyone), how you heard and reacted’.

Instead, in the control-neutral condition, the instructions were as follows: ‘Please think about the last time you went to the supermarket to do shopping, think about the day and time you went, who you were with and what you purchased. Even if it is an everyday, routine event, try to focus your memory on this experience. Please describe the memory in as much detail as possible, what happened and when (specifying the date, e.g., January 2019), who you were with (if there was anyone), how you felt and reacted’. The participants were randomly assigned to one of the three experimental conditions.

Post-manipulation phase (retest)

Immediately after the state emotion induction phase, the participants were asked to assess the retrospective emotions they experienced during the event (emotions at encoding) and their emotions experienced by rethinking and recalling the specific event (emotions at recall—Retest). For both assessments, participants had to indicate on six 11-point Likert scales (0 = “not at all”; 10 = “completely”) how intensely emotions were felt: (a) guilt; (b) anger; (c) frustration; (d) helplessness; (e) unfairness; and (f) shame.

Story reading phase

In this phase, the participants were asked to read a short story taken from a real forensic case of personal injury defense and to identify as much as possible with the protagonist Marco (To guarantee anonymity, the name is fictitious)⁴².

‘Your name is Marco; you are 18 years old; you are 1.80 m tall, and you weigh 98 kg. This morning, you had a verbal dispute with one of your classmates, who provoked you greatly. As a reaction to this provocation, you offend him just as badly. On leaving the school, on the way home, your classmate chases after you and attacks you from behind. You try to break free and start fighting. After falling to the ground, you jump upon him and choose to punch him with all you might.’

Moral judgments phase

After reading the story, the participants were asked to judge what happened by imagining that they were Marco having chosen the risky criminal option (punch the classmate). The participants were assessed from 0 (“not at all”) to 100 (“completely”) to what extent they judged: (1) their own offense as the cause of what happened (internal causal attribution judgment), (2) their classmate’s provocation as the cause of what happened (external causal attribution judgment), (3) themselves as responsible for the choice (self-responsibility judgment), (4) their classmate as responsible for their own choice (other-responsibility judgment), (5) the choice to punch the classmate as predictable (predictability judgment), (6) as intentional (intentionality judgment), and (7) the outcomes of their own choice as severe both for themselves and for their classmate (severity judgment¸ Cronbach’s alpha = 0.63). Additionally, participants were asked to estimate from 0 (“not at all”) to 100 (“completely”) to what extent they hadn’t chosen to punch their classmate: (8) things would have been better (upward counterfactual thought; thought generated when people imagine better alternative states) or (9) things would have been worse (downward counterfactual thought; thought generated when people imagine worse alternative states). Moreover, each participant had to estimate from 0 (“not at all”) to 100 (“completely”) how much: (10) if it were possible to return, they would reverse their own choice (reversibility willingness judgment). Finally, each participant had to estimate from 0 (“not at all”) to 100 (“completely”) the extent to which they anticipated (11) moral emotions (guilt, shame, regret, disappointment; anticipated moral emotions; Cronbach’s alpha = 0.88), (12) negative emotions (anger, anxiety, fear, sadness; anticipated negative emotions; Cronbach’s alpha = 0.68), and (13) positive emotions (pleasure, satisfaction, pride; anticipated positive emotions; Cronbach’s alpha = 0.83). While the main manuscript primarily presents analyses using emotion clusters (moral, negative, and positive) for clarity and parsimony, we also conducted separate analyses for each individual anticipated emotion for the core findings (results are reported in the Supplementary Materials).

The battery was created ad hoc and subsequently administered online via Google Modules. Each participant was individually tested and completed all phases of the experiment in approximately 50 min. Following the experiment, the participants were fully debriefed.

Results

Manipulation check

To exclude any a priori differences between the three conditions (Anger vs. Guilt vs. Neutral) in the dispositional tendencies to anger and guilt, a one-way MANOVA was conducted with State Emotion as the between-subjects factor (Anger vs. Guilt vs. Neutral) and Trait Anger, Trait-anger-temperament, Trait-anger-response, and Trait Guilt as the dependent variables. No effect was found to be statistically significant (F_s < 1.50, p_s > 0.05, ηp²_s < 0.02; Table 1).

Table 1.

Manipulation check on trait and state emotions before and after manipulation.

	Anger (n = 68) M (SD)		Guilt (n = 68) M (SD)		Neutral (n = 68) M (SD)		State emotion F2,201 (ηp2)
Trait anger	19.78 (4.39)		20.57 (5.33)		19.51 (5.42)		0.80 (.01)
Trait anger temperament	6.56 (2.16)		6.94 (2.51)		6.24 (2.46)		1.50 (.02)
Trait anger response	9.78 (2.57)		10.09 (2.80)		9.60 (2.95)		0.53 (.01)
Trait guilt	63.68 (8.03)		64.37 (7.17)		64.49 (7.53)		0.23 (.00)
Anger at encoding	8.99 (1.23)a		4.60 (3.63)b		.00 (.00)c		280.06*** (.74)
Guilt at encoding	2.56 (3.23)a		8.81 (1.78)b		.00 (.00)c		307.31*** (.75)
Frustration at encoding	7.04 (3.28)a		6.44 (3,80)a		.35 (1.21)b		104.78*** (.51)
Helplessness at encoding	6.81 (3.19)a		5.72 (3.92)a		.56 (1.74)b		79.65*** (.44)
Unfairness at encoding	8.59 (2.44)a		4.59 (3.74)b		.38 (1.80)c		148.19*** (.60)
Shame at encoding	2.93 (3.45)a		7.81 (3.11)b		.18 (.99)c		135.16*** (.57)

	Anger (n = 68)		Guilt (n = 68)		Neutral (n = 68)		State emotion F2, 201 (ηp2)	Time F1, 201 (ηp2)	State emotion*time F2, 201 (ηp2)
	Time 1 M (SD)	Time 2 M (SD)	Time 1 M (SD)	Time 2 M (SD)	Time 1 M (SD)	Time 2 M (SD)
Anger	1.90 (2.57)	5.69 (3.22)	2.71 (2.78)	3.22 (3.43)	1.50 (2.46)	.00 (.00)	35.23*** (.26)	16.36*** (.09)	52.55*** (.34)
Guilt	1.96 (3.02)	.94 (1.73)	2.44 (3.19)	6.09 (3.52)	1.50 (2.37)	.00 (.00)	54.57*** (.35)	3.02 (.02)	57.04*** (.36)
Frustration	2.85 (3.44)	3.81 (3.57)	3.66 (3.64)	4.04 (3.61)	2.63 (3.07)	.26 (1.13)	15.32*** (.13)	1.90 (.01)	16.99*** (.15)
Helplessness	2.74 (3.39)	3.41 (3.79)	3.69 (3.56)	3.13 (3.60)	2.74 (3.35)	.28 (1.05)	9.95*** (.09)	8.68** (.04)	11.86*** (.11)
Unfairness	3.19 (3.45)	6.00 (3.87)	4.21 (3.44)	3.50 (3.70)	3.38 (3.60)	.13 (1.09)	20.23*** (.17)	1.96 (.01)	41.30*** (.29)
Shame	1.32 (2.41)	1.10 (2.00)	2.12 (2.95)	5.66 (3.60)	1.15 (2.35)	.16 (1.03)	54.67*** (.35)	11.80** (.06)	38.05*** (.28)

M = mean; SD = standard deviation.

*p < .05; **p < .01; *** = p < .001.

Subscripts (_a,b,c) refer to Bonferroni-corrected post hoc comparisons: Means that do not share a subscript differ significantly at p < .001; means sharing the same subscript do not differ.

η_p² = partial eta squared.

To check the efficacy of the state emotion manipulation, a set of 3 × 2 repeated-measures ANOVAs was conducted with State Emotion as the between-subjects factor (Anger vs. Guilt vs. Neutral), Time (pre-manipulation [Test] vs. post-manipulation [Retest]) as the within-subjects factor, and pre- and post-emotion ratings as the dependent variables (Table 1).

The main effect of State Emotion on how participants rated all their emotional state was statistically significant (F_s (2, 201) > 9.95, p_s < 0.001, ηp²_s > 0.09). Bonferroni-corrected pairwise comparisons revealed that participants felt more anger for the anger event (M = 3.79; SD = 0.27) than for the guilt event (M = 2.96; SD = 0.27; t(67) = 4.74, p < 0.001) and the neutral event (M = 0.75; SD = 0.27; t(67) = 12.64, p < 0.001), and felt more guilt for the guilt event (M = 4.27; SD = 0.25) than for the anger event (M = 1.45; SD = 0.25; t(67) = 9.86, p < 0.001) and the neutral event (M = 0.75; SD = 0.25; t(67) = 11.88, p < 0.001). For the State Emotions effect on the other participants’ emotions (i.e., frustration, helplessness, unfairness, and shame), see Table 1.

The main effect of Time was statistically significant for the emotions of anger, helplessness, and shame (F_s (1, 201) > 8.68, p_s < 0.001, ηp²_s > 0.04). Specifically, participants reported higher levels of anger at Retest (M = 2.97, SD = 3.57) compared to Test (M = 2.03, SD = 2.64), as well as higher levels of shame at Retest (M = 2.31, SD = 3.43) than at Test (M = 1.53, SD = 2.61). In contrast, feelings of helplessness decreased from Test (M = 3.05, SD = 3.45) to Retest (M = 2.27, SD = 3.37). For the other emotions the effect of Time was not statistically significant (F_s (1, 201) < 3.05, p_s > 0.05, ηp²_s < 0.03).

The interaction effect State Emotion*Time was found to be significant for how participants rated all their emotional states (F_s (2, 201) > 11.86, p_s < 0.001, ηp²_s > 0.11) (Table 1). An analysis of simple effects revealed that the level of anger experienced during the recall of the anger event was greater in the post-manipulation (Retest) than in the pre-manipulation (Test) (M_pre = 1.90, SD = 2.57; M_post = 5.69, SD = 3.22; t(67) =  − 9.58, p < 0.001), whereas the anger score for the guilt event did not significantly change (M_pre = 2.71, SD = 2.78; M_post = 3.22, SD = 3.43; t(67) =  − 1.28, p = 0.21), and the level of anger experienced during the recall of the neutral event was lower at Retest than at Test (M_pre = 1.50, SD = 2.46; M_post = 0.00, SD = 0.00; t(67) = 5.03, p < 0.001). Examining differences between state emotion conditions at each time point, a significant difference emerged at Test only between guilt and neutral conditions (F (2, 201) = 3.78, p < 0.05, ηp² = 0.04), while at Retest, the anger condition reported significantly higher anger than both the guilt and neutral conditions, F(2, 201) = 75.00, p < 0.001, ηp² = 0.43).

Moreover, guilt felt during the recall of the guilt event was greater at Retest than at Test (M_pre = 2.44, SD = 3.19; M_post = 6.09, SD = 3.52; t(67) =  − 8.08, p < 0.001), whereas guilt scores for the anger event (M_pre = 1.96, SD = 3.02; M_post = 0.94, SD = 1.73; t(67) = 2.73, p < 0.01) and for the neutral event (M_pre = 1.50, SD = 2.46; M_post = 0.00, SD = 0.00; t(67) = 5.21, p < 0.001) decreased during the post-manipulation compared with the pre-manipulation. Examining differences between state emotion conditions at each time point, there were no significant differences among conditions at Test (F(2, 201) = 1.81, p = 0.17, ηp² = 0.02), while at Retest, the guilt condition showed significantly higher guilt scores compared to both anger and neutral conditions (F(2, 201) = 142,64, p < 0.001, ηp² = 0.59).

In addition, to check that the participants actually followed and understood the instructions and correctly recalled the correct event in each condition, a one-way MANOVA was conducted with State Emotion as the between-subjects factor (Anger vs. Guilt vs. Neutral) and emotions at encoding as the dependent variables (Table 1). The main effect of State Emotion on how participants rated the emotionality of the recalled events was statistically significant (F_s (2, 201) > 79.65, p_s < 0.001, ηp²_s > 0.44). Post hoc with Bonferroni correction indicated that participants rated the anger event (M = 8.99; SD = 1.23) as more angry than the guilt event (M = 4.60; SD = 3.63; t(201) = 11.5, p < 0.001) and the neutral event (M = 0.00; SD = 0.00; t(201) = 23.7, p < 0.001), and rated the guilt event (M = 8.81; SD = 1.78) as more guilty than the anger event (M = 2.56; SD = 3.23; t(201) = 17.10, p < 0.001) and the neutral event (M = 0.00; SD = 0.00; t(201) = 7.00, p < 0.001). For the other encoding emotions (i.e., frustration, helplessness, unfairness, and shame), see Table 1.

In summary, the experimental manipulation of participants’ emotion states was found to be effective: When participants recalled angry and guilty events, their angry and guilty emotional states increased in the post-manipulation at Retest, respectively.

The effect of incidental state emotions on judgments

To test the effects of state anger and state guilt emotions on moral judgments (H1 and H2), a one-way MANOVA was conducted with State Emotion as the between-subjects factor (Anger vs. Guilt vs. Neutral) and evaluative judgments as the dependent variables (Table 2). The results did not provide overall support for our hypotheses, as no significant main effects were found for most of the dependent variables. The only significant effect was observed for the self-responsibility judgment (F(2, 201) = 3.30, p < 0.05, ηp² = 0.03), where participants in the state anger condition attributed more responsibility to themselves compared to those in the state guilt condition (t = 2.48, p = 0.04).

Table 2.

One-way MANOVA on moral judgments.

Judgments	Anger (n = 68)	Guilt (n = 68)	Neutral (n = 68)	State Emotion F2. 201 (ηp2)
	M (SD)	M (SD)	M (SD)
Internal causal attribution	40.29 (30.77)	41.34 (31.81)	34.74 (29.58)	.91 (.01)
External causal attribution	70.88 (28.38)	71.18 (29.59)	71.56 (31.78)	.01 (.00)
Self-responsibility	61.99 (33.29) a	47.85 (33.27) b	57.85 (32.39) b	3.30* (.03)
Others-responsibility	65.46 (27.87)	63.90 (30.39)	62.41 (34.45)	.16 (.00)
Predictability	43.22 (33.80)	45.81 (32.37)	44.91 (35.84)	.10 (.00)
Intentionality	38.18 (32.72)	45.28 (32.09)	44.37 (37.87)	.86 (.01)
Severity	71.23 (23.86)	77.60 (20.02)	69.98 (23.78)	2.22 (.02)
Upward counterfactual thought	69.21 (30.37)	68.04 (27.36)	65.15 (35.57)	.30 (.00)
Downward counterfactual thought	42.57 (31.81)	41.60 (33.18)	40.65 (36.35)	.06 (.00)
Reversibility willingness	66.25 (34.53)	65.59 (34.39)	70.69 (34.11)	.44 (.00)
Anticipated moral emotions	59.35 (28.61)	51.91 (28.03)	49.76 (30.19)	2.06 (.02)
Anticipated negative emotions	59.35 (21.97)	56.66 (26.04)	53.11 (23.64)	1.16 (.01)
Anticipated positive emotions	28.51 (23.73)	35.35 (26.37)	28.97 (27.76)	1.47 (.01)

M = mean; SD = standard deviation.

*p < .05.

Subscripts (_{a, b, c}) refer to Bonferroni-corrected post hoc comparisons: means that do not share a subscript differ significantly at p < .05; means sharing the same subscript do not differ significantly.

η_p² = partial eta squared.

Given the limited and unexpected findings from the MANOVA, we conducted additional exploratory analyses to further investigate the relationship between participants’ experienced emotional states and their evaluative judgments. Specifically, we ran Pearson’s correlations among judgments and the anger and guilt emotions experienced at post-manipulation (Retest) within each state emotion condition (Table 3) (In the neutral condition, we did not analyze the correlations because the average levels of anger and guilt emotions at Retest were equal to zero). The results revealed that, in the state anger condition, anger experienced after recalling an angry event was positively associated with judgments of predictability and intentionality and negatively associated with internal causal attribution and upward counterfactual thought. No correlation with guilt at Retest was found. Instead, in the state guilt condition, guilt experienced after recalling a guilty event was positively associated with internal causal attribution and judgments of severity. No correlation with anger at Retest was found.

Table 3.

Pearson’s correlations between moral judgments and state emotions at Time 2 (Retest) recall.

Judgments	State anger condition (n = 68)		State guilt condition (n = 68)
	Anger at retest	Guilt at retest	Anger at retest	Guilt at retest
Internal causal attribution	− .24*	− .02	.13	.25*
External causal attribution	− .11	− .04	− .01	.20
Self-responsibility	.03	− .03	.04	.05
Others-responsibility	.13	− .17	− .09	.14
Predictability	.29*	.053	.07	− .00
Intentionality	.26*	.13	− .10	.13
Severity	− .16	− .15	.01	.33**
Upward counterfactual thought	− .34**	− .09	.06	.03
Downward counterfactual thought	.06	.15	− .03	− .09
Reversibility willingness	− .20	− .13	− .20	.09
Anticipated moral emotions	− .19	.04	.12	.22
Anticipated negative emotions	− .12	.17	.22	.18
Anticipated positive emotions	.11	.03	.04	.06

*p < .05; **p < .01.

The effect of incidental trait emotions on judgments

To test the effects of trait anger and trait guilt on judgments (H1 and H2), we ran Pearson’s correlations among the trait emotions of anger and guilt and judgments (Table 4). The results revealed a differentiated pattern of associations across the trait anger components. Specifically, internal causal attribution was positively associated with trait anger temperament, whereas external causal attribution was significantly correlated with total trait anger. All three trait anger components showed positive associations with judgments of others’ responsibility, predictability, and intentionality. In addition, trait anger (total and response) was positively related to the anticipation of both negative and positive emotions, whereas trait anger (total and temperament) was negatively associated with the willingness to reverse the choice. Trait guilt was positively associated with a willingness to reverse decisions and with anticipated moral emotions. Overall, these findings suggest that specific facets of trait anger and guilt may differentially contribute to moral evaluation processes.

Table 4.

Pearson’s correlations among moral judgments and trait emotions of anger and guilt in total sample and across conditions.

Judgments	Total sample (N = 204)
	Trait anger			Trait anger temperament			Trait anger response			Trait guilt
Internal causal attribution	.07			.20**			− .03			− .02
External causal attribution	.21**			.11			.17*			− .02
Self-responsibility	.11			.11			.10			− .01
Others-responsibility	.25**			.14*			.21**			− .07
Predictability	.28**			.24**			.15*			− .10
Intentionality	.31**			.29**			.17*			− .11
Severity	.15*			.08			.18*			.01
Upward counterfactual thought	− .14			− .12			− .07			.13
Downward counterfactual thought	.07			.12			.05			− .06
Reversibility willingness	− .17*			− .15*			− .08			.14*
Anticipated moral emotions	− .01			− .08			.08			.16*
Anticipated negative emotions	.18**			.08			.24**			.03
Anticipated positive emotions	.25**			.20**			.17*			− .11

	State anger condition (n = 68)				State guilt condition (n = 68)				Neutral condition (n = 68)
	T_Ax	T_Ax_T	T_Ax_R	TG	T_Ax	T_Ax_T	T_Ax_R	TG	T_Ax	T_Ax_T	T_Ax_R	TG
Internal causal attribution	− .16	.03	− .27	.17	.14	.27*	.00	− .13	.19	.24*	.13	− .11
External causal attribution	.12	− .01	.16	− .14	.20	.20	.10	.07	.29*	.13	.25*	.02
Self-responsibility	− .02	.04	− .05	− .14	.19	.16	.19	.08	.18	.16	.18	.07
Others-responsibility	.25*	.10	.26*	− .25*	.18	.13	.08	− .09	.31**	.18	.29*	.11
Predictability	.25*	.14	.14	− .14	.22	.27*	.06	− .17	.35**	.31**	.24	.00
Intentionality	.40**	.32**	.24	− .26*	.32**	.36**	.14	− .15	.24*	.21	.13	.04
Severity	− .01	− .09	.12	.00	.06	.10	− .00	.09	.32**	.18	.36**	− .06
Upward counterfactual thought	− .23	− .15	− .18	.08	− .10	− .08	− .08	.08	− .11	− .13	.01	.23
Downward counterfactual thought	.06	.19	− .07	− .06	− .03	− .05	.05	.09	.17	.22	.14	− .19
Reversibility willingness	− .19	− .16	− .04	.20	− .15	− .14	− .12	− .04	− .17	− .14	− .06	.25*
Anticipated moral emotions	− .08	− .18	.06	.28*	.13	.17	.06	.09	− .09	− .23	.13	.12
Anticipated negative emotions	.15	.08	.21	− .11	.28*	.26*	.19	.03	.10	− .13	.32**	.17
Anticipated positive emotions	.22	.21	.10	− .34**	.32**	.28*	.23	− .09	.18	.09	.14	.07

Pearson’s correlation coefficients are reported for the total sample (N = 204) and separately for each experimental condition: State Anger (n = 68), State Guilt (n = 68), and Neutral (n = 68).

T_Ax = trait anger (total score); T_Ax_T = trait anger temperament; T_Ax_R = trait anger reaction; TG = trait guilt.

*p < .05; **p < .01.

To further explore these patterns, we examined correlations between trait emotions and judgments within each state emotion condition (anger and guilt; see Table 4). In the state anger condition, total trait anger was positively correlated with others’ responsibility, predictability, and intentionality. Trait anger temperament was positively associated with intentionality. Trait anger reaction was positively correlated with others responsability. Trait guilt was negatively associated with others’ responsibility, intentionality, and anticipated positive emotions, and positively associated with anticipated moral emotions. In the state guilt condition, total trait anger was positively associated with intentionality and both anticipated negative and positive emotions. Trait anger temperament was positively associated with internal causal attribution, predictability, intentionality, and both anticipated negative and positive emotions. 

To clarify the unique contribution of each trait anger facet, we also conducted partial correlation analyses. These analyses allowed us to isolate the unique associations between each trait anger component and moral judgment dimensions (results are reported in the Supplementary Materials).”

Interaction effect of state*trait incidental emotions on judgments

To test the interaction effect trait*state incidental emotions of anger and guilt, we conducted a series of hierarchical regression models (HRMs) with state and trait emotions as predictors and judgments as outcomes (Table 5). We first included in the model the trait emotions of anger and guilt, then added the state emotions in the second step, and finally added the interaction terms in the third step. The results revealed only the role of trait emotions in impact judgments. Specifically, trait anger predicted higher levels of external causal attribution, judgments of other-responsibility, predictability, intentionality, severity, and anticipated negative and positive emotions; and a lower willingness to reverse the choice made. Trait guilt predicted higher levels of upward counterfactual thought, reversibility willingness and anticipated moral emotions. Neither the effect of state emotions nor the interaction effect trait*state was found to be significant.

Table 5.

Hierarchical regression models.

Predictors	Internal causal attribution			External causal attribution			Self-responsibility
	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3
	β	β	β	β	β	β	β	β	β
Block 1
Trait anger	.07	.07	.16	.21**	.21**	.29*	.11	.13	.17
Trait guilt	− .02	− .01	− .10	− .01	− .01	.04	− .00	.00	.08
Block 2
State anger (neutral)		.18	.17		− .03	− .05		.12	.10
State guilt (neutral)		.20	.18		− .06	− .06		− .33	− .33
Block 3
State*trait anger (neutral)			− .31			− .20			− .24
State*trait guilt (neutral)			− .08			− .01			− .03
R2	.01	.01	.05	.05	.05	.05	.01	.05	.06
Adjusted R2	− .00	− .01	.01	.04	.03	.02	.00	.03	.03
F (df)	.58 (2, 201)	.69 (4, 199)	1.38 (8, 195)	4.74 (2, 201)*	2.38 (4, 199)	1.41 (8, 195)	1.24 (2, 201)	2.47 (4, 199)*	1.67 (8, 195)

Predictors	Others-responsibility			Predictability			Intentionality
	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3
	β	β	β	β	β	β	β	β	β
Block 1
Trait anger	.25***	.25***	.33**	.27***	.27***	.34**	.31***	.31***	.25*
Trait guilt	− .06	− .06	.15	− .09	− .09	.03	− .10	− .10	.07
Block 2
State anger (neutral)		.08	.07		− .07	− .07		− .21	− .20
State guilt (neutral)		− .00	.01		− .03	− .03		− .04	− .03
Block 3
State*trait anger (neutral)			− .13			− .08			.14
State*trait guilt (neutral)			− .29			− .25			− .28
R2	.07	.07	.09	.08	.08	.09	.11	.11	.13
Adjusted R2	.06	.05	.05	.07	.07	.06	.10	.10	.10
F (df)	7.01 (2, 201)***	3.56 (4, 199)**	2.43 (8, 195)*	9.14 (2, 201)***	4.58 (4, 199)**	2.62 (8, 195)*	11.80 (2, 201)***	6.35 (4, 199) ***	3.68 (8, 195) ***

Predictors	Severity			Upward counterfactual thought			Downward counterfactual thought
	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3
	β	β	β	β	β	β	β	β	β
Block 1
Trait anger	.14*	.13	.31**	− .13	− .13	− .10	.07	.07	.16
Trait guilt	.01	.01	− .04	.13	.13	.25*	− .06	− .06	− .19
Block 2
State anger (neutral)		.05	.03		.15	.14		.05	.03
State guilt (neutral)		.31	.30		.12	.12		.01	.10
Block 3
State*trait anger (neutral)			− .32			− .16			− .11
State*trait guilt (neutral)			.12			− .16			.30
R2	.02	.04	.06	.04	.04	.05	.01	.01	.03
Adjusted R2	.01	.02	.02	.03	.02	.01	− .00	− .01	− .01
F (df)	2.14 (2, 201)	2.02 (4, 199)	1.58 (8, 195)	3.63 (2, 201)*	2.03 (4, 199)	1.31 (8, 195)	.83 (2, 201)	.43 (4, 199)	.72 (8, 195)

Predictors	Reversibility willingness			Anticipated moral emotions			Anticipated negative emotions
	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3
	β	β	β	β	β	β	β	β	β
Block 1
Trait anger	− .16*	− .16*	− .14	− .00	− .00	− .08	.19**	.18**	.11
Trait guilt	.13	.13	.24*	.16*	.17*	.12	.03	.04	.17
Block 2
State anger (neutral)		− .11	− .10		.35*	.36		.25	.26
State guilt (neutral)		− .11	− .11		.08	.07		.11	.11
Block 3
State*trait anger (neutral)			− .03			.07			.04
State*trait guilt (neutral)			− .25			− .05			–.20
R2	.05	.05	.06	.03	.05	.06	.04	.05	.06
Adjusted R2	.04	.03	.02	.02	.03	.02	.03	.03	.02
F (df)	4.92 (2, 201)**	2.58 (4, 199)*	1.54 (8, 195)	2.63 (2, 201)	2.49(4, 199)*	1.54 (8, 195)	3.64 (2, 201)*	2.39 (4, 199)	1.63 (8, 195)

Predictors	Anticipated positive emotions
	Block 1	Block 2	Block 3
	β	β	β
Block 1
Trait anger	.25**	.24**	.19
Guilt anger	− .10	− .11	.08
Block 2
State anger (neutral)		− .04	− .04
State guilt (neutral)		.19	.19
Block 3
State*trait anger (neutral)			− .04
State*trait guilt (neutral)			− .25
R2	.07	.08	.11
Adjusted R2	.06	.06	.07
F (df)	7.87 (2, 201)***	4.51 (4, 199)**	2.98 (8, 195)**

*p < .05; **p < .01; ***p < .001

Discussion

The current study aimed to investigate how incidental emotions of anger and guilt—both trait and state—affect evaluative judgments in a forensic case of personal injury defense depicting a quarrel between two classmates.

As hypothesized in H1, state anger was associated with higher judgments of predictability and intentionality, along with lower internal causal attribution and fewer upward counterfactual thoughts. Additionally, trait anger showed significant associations with external causal attribution, judgments of other-responsibility, predictability, intentionality, severity, and anticipated emotions (both negative and positive), while also correlating negatively with the desire to reverse the decision made. These findings point to a strong and direct influence of trait anger rather than a transitory effect of state anger. Individuals with a general predisposition to frequently experience anger across various situations may be more inclined to perceive situations as unfair or frustrating, which could heighten their likelihood of making risky choices, such as committing a crime. This trait-based effect appears stronger than the influence of a temporary emotional state. A possible speculative explanation for this result is that while our manipulation successfully induced state anger, the activation might not have been sufficiently intense or directly linked to perceptions of injustice and offense depicted in the scenario. State anger, by definition, refers to a reactive form of anger triggered by specific events. According to state-trait anger theory⁴⁴, Spielberger described three different components of anger, hostility, and aggression⁴⁸: anger refers to arousal in response to certain events (state anger), hostility to a dispositional tendency toward frequent and intense anger (trait anger), and aggression to the expression of anger. Our results, based on emotional induction rather than real-life triggering events, suggest that participants’ dispositional trait anger (hostility) was more salient than the temporarily induced state anger.

Among the several judgments assessed, predictability and intentionality judgments emerged as the most consistent, showing associations with both state and trait anger across correlational analyses and between-group comparisons. However, it is important to note that the overall pattern of results was modest, and that the principal findings did not fully support the original hypotheses.

Our findings regarding the associations between anger and judgments of other-responsibility, predictability, and intentionality align with previous studies examining the carryover effects of anger on attributions of causality, blame, and evaluations⁴⁹. Incidental anger increases the tendency to hold other individuals responsible for subsequent events⁵⁰ and to make punitive judgments about other individuals, both related and unrelated to the original source of anger⁵¹. It seems that feeling anger may promote punitive attributions toward a defendant and harsher punishments, even when the defendant is unrelated to the initial cause of anger⁵². This pattern suggests that greater anger is associated with a stronger tendency to place blame on others. In two experiments, Subra⁵³ reported a significant link between anger and the proportion of intentional judgments when participants evaluated ambiguous sentences, with angry participants more likely to endorse intentional explanations than neutral participants. These findings suggest that anger increases the likelihood of making hostile inferences. Given that anger is an emotion closely linked to perceptions of control¹², it seems reasonable to propose that when individuals perceive an action as being under personal control, they are more likely to conclude that it was performed intentionally.

Our initial hypothesis regarding self-responsibility was not fully supported. Contrary to expectations, state anger was associated with a greater degree of self-responsibility. Several potential explanations may account for this unexpected outcome. One possible reason is methodologically based: It seems that it is not the induced emotional state itself that had the greatest impact on judgment, but rather the emotion recalled during the autobiographical memory retrieval. Indeed, the anger reported at post-manipulation (after the recall of the anger event) was not significantly associated with the self-responsibility judgment in the anger condition. Another explanation may lie in how anger influences perceptions of human agency. Anger might lead individuals to emphasize personal responsibility over situational factors when explaining events. Previous research^54,55 has shown that anger can cause people to focus more on individual agency rather than contextual explanations, which may explain why anger is linked to judgments of both self-responsibility and others’ responsibility. Finally, incidental anger may influence how individuals attribute causality and evaluate responsibility in different ways.

Caution is also warranted when interpreting the observed positive association between trait anger and severity judgments of outcomes, as this link was modest and emerged only in correlational analyses, suggesting that other variables may mediate or moderate this relationship. Notably, participants in our study were asked to retrospectively judge a choice already made, rather than make an immediate contextual decision. Research suggests that feelings of anger may lead individuals to underestimate risks and consequences during decision-making dilemmas. In particular, compared with those who feel fear, angry individuals are more inclined to engage in risk-seeking behavior⁹. Judging a decision retrospectively rather than choosing between two options could yield subtly different results regarding severity evaluation. Moreover, individuals experiencing anger tended to generate fewer upward counterfactual thoughts, whereas those high in trait anger demonstrated less desire to reverse their decisions and anticipated more intense emotional outcomes, both negative and positive. These patterns are consistent with the antecedent appraisals of anger in terms of perceived control and responsibility. Mental simulations of alternative outcomes are known as counterfactual thoughts⁵⁶, and complex emotions such as shame, remorse, or guilt are closely tied to counterfactual reasoning^57,58.

Confirming H2, state guilt was associated with greater internal causal attributions and higher judgments of severity. Additionally, trait guilt was positively linked to a stronger desire to reverse the immoral choice made and to greater anticipation of moral emotions. Studies investigating counterfactual thinking connected to shame and guilt have shown that upward counterfactual thoughts, such as ‘If I hadn’t…,’ are typically linked to guilt, as they attribute the undesirable outcome to a specific behavior^59,60. Because moral judgments encourage individuals to take responsibility for their mistakes and pursue reparative actions, self-responsibility and internal causal attribution are crucial to the social and moral functions of guilt⁶¹. Individuals high in guilt or experiencing state guilt are more likely to advocate for reparative behaviors, including going back and changing their decision. The motivation to repair harm caused to others has been consistently associated with guilt across a wide range of studies^61–63. Regarding choices, experiencing incidental guilt has been shown to heighten perceptions of the severity of negative consequences⁴⁰.

In contrast to H3, no significant interaction effects between trait and state emotions were found. This result suggests that trait emotions, such as anger and guilt, may exert a more stable and pervasive influence on moral judgments than transient emotional states. It appears that individuals’ general predisposition to experience certain emotions plays a stronger role in how they evaluate situations as immoral or risky. Based on the current findings, trait and state emotions seem to operate independently, each contributing separately to the evaluative process during moral judgment.

Conclusions

The present study set out to explore how incidental emotions—specifically anger and guilt, both as stable traits and transient states—shape the way individuals evaluate morally complex situations. By focusing on a realistic forensic case, we aimed to understand whether and how these emotions influence attributions of responsibility, perceptions of intent and severity, and the inclination to mentally revise past choices.

Our findings provide insights into the distinct roles of anger and guilt in evaluative processes, while also underscoring the inherent complexity of emotional influences on moral reasoning.

In summary, the results suggest that emotions—even those sharing a similar valence—can influence evaluative judgments in qualitatively different ways. Understanding the specific effects of anger and guilt on how individuals assess responsibility, predictability, and intent is crucial not only for advancing emotion theory, but also for informing domains such as legal decision-making, conflict resolution, and the development of interventions aimed at fostering more reflective and equitable moral reasoning. In our study, trait emotions appeared to exert a more pervasive and stable influence on evaluative judgments, whereas the impact of state emotions was more limited and inconsistent. Although our manipulations effectively altered emotional states, these transient changes did not consistently produce the hypothesized effects on moral judgment. However, the limited impact of state emotions does not diminish the significance of our findings; rather, it offers a more nuanced understanding of when and how emotions—particularly dispositional ones—shape evaluative processes, suggesting that trait-based influences may exert a stronger and more consistent effect compared to situational emotional states in certain contexts. This also highlights the scientific value of reporting null or unexpected results, not least to mitigate publication bias⁶⁴.

Naturally, the current study has several limitations. First, we assessed retrospective judgments, which can be considered indicative of future decisions, but participants were not asked to make real-time forced choices, such as in moral dilemmas (e.g.^59,60); thus, the influence of incidental emotions on decision-making processes may differ slightly. Second, while participants were instructed to identify with the agent of the choice, it is possible that some responded from a judge’s perspective rather than an actor’s perspective, potentially affecting how emotions influenced their evaluations. Third, we did not consider individual differences such as emotion regulation strategies, which have been strongly implicated in mitigating emotion-related biases⁶⁵. Last, although regression analyses were conducted to examine the unique contribution of trait and state emotions, as well as their possible interaction, in predicting moral judgments, the overall explanatory power of these models was limited. This suggests that other unmeasured variables may also play a substantial role in shaping such evaluations. For example, individual differences in moral disengagement⁶⁶, cognitive reflection⁶⁷, cognitive abilities or difficulties^68,69, empathy⁷⁰, personality traits⁷¹, or specific characteristics of the scenarios⁷² have all been shown to influence moral judgment and decision-making and may also have impacted the effects observed here.

Future research would benefit from exploring these dynamics further, for example by using more immersive and vivid emotion induction techniques or employing real-time moral decision-making paradigms that provoke deeper emotional engagement. Additionally, examining incidental emotions across a broader range of moral contexts—such as sacrificial dilemmas or everyday ethical conflicts—may provide further insight into the interplay between emotional states, traits, and moral evaluations. Finally, given that moral judgments can be sensitive to the nature of the scenario—including the type of transgression, the social roles involved, and the perceived gravity of the situation—future studies should aim to replicate and extend these findings across different morally relevant contexts to better assess the generalizability of the results.

Author contributions

TL: Conceptualization, Methodology, Data Curation, Formal Analysis, Investigation, Writing—Original Draft. FA: Formal Analysis, Writing—Review & Editing. AC: Resources, Writing—Review & Editing. All authors contributed to the critical revision of the manuscript, approved the final version for publication, and agree to be accountable for the content of the work.

Funding

This research was funded by the Italian Ministry of University and Research under the 2022 PRIN (Research Projects of National Interest) Program, Research Grant No. 20225ECXPP.

Data availability

According to integrity and transparency principles in research, all study materials and data are available on the OSF platform (https://osf.io/zkvrq/).

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study was approved by the Ethics Committee of the University of Bari ‘Aldo Moro’, Department of Education, Psychology and Communication (No. ET-19-13).

Informed consent

Informed consent was obtained from all individual participants included in the study.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-16277-x.