A Just and Happy Ending: Exploring Viewers’ Judgment of Justice on Enjoyment of Crime Drama in Interactive Movie Events

ABSTRACT

This study examined viewers’ reasoning about justice and enjoyment of crime drama in interactive movie events. Based on the integrated model of enjoyment for crime drama, we examined the role of viewers’ judgment of justice for enjoyment and theoretically discussed the role of virtually present co-viewers. We conducted a lab experiment with N= 148 participants watching a full interactive movie manipulating the viewers’ notion of what is just with what is presented in screen. Results confirm that viewers’ enjoyment is a function of the viewers’ judgment of justice and that virtually present co-viewers can serve as a justice compass.

Viewers’ enjoyment of crime drama, one of the most popular genres in the entertainment industry, has attracted the interest of many scholars in media psychology (Bartsch et al., 2020; Grizzard et al., 2019; Tamborini et al., 2018; Waddell et al., 2019). Enjoyment has been examined as a function of affective as well as cognitive responses toward the crime and the characters. Affective responses include recipients’ suspense while watching and emotional reactions such as sympathy for the victim. The viewers’ enjoyment increases, for example, with the level of suspense during the reception process (excitation transfer, Zillmann, 1971; Zillmann et al., 1975) or when a hoped-for positive outcome occurs for liked characters and a negative outcome for disliked characters (affective disposition, Zillmann, 2000). While the majority of empirical studies focus on the impact of affective responses on viewer enjoyment (e.g., Eden et al., 2011; Tamborini et al., 2018; Waddell et al., 2019; Zillmann & Cantor, 1977), research on more controlled cognitive responses is increasingly gaining the interest of scholars (Bartsch et al., 2020; Grizzard et al., 2019; Hofer et al., 2020; Raney, 2002). Cognitive responses include moral judgment of crime and characters (e.g., Eden et al., 2017; Hofer et al., 2020) and reasoning about justice (Raney & Bryant, 2002; Schmitt & Maes, 2013). Viewers’ enjoyment increases, for example, the stronger the moral core of a story corresponds with the viewers’ moral foundations (MIME, Tamborini, 2011) or when the punishment a criminal receives matches the viewers’ perception of what is deserved for the crime committed (integrated model of enjoyment for crime drama, Raney & Bryant, 2002).

This study aims to extend the evidence on viewers’ reasoning about justice and enjoyment of crime drama and explores such cognitive responses against the backdrop of technological advancements in the film and television industry. The continuous growth of broadband Internet access (International Telecommunication Union, 2020) and spread of smart TVs in western industrialized homes fostered the development of new attractive entertainment formats with interactive options. Interactive movies have existed since the late 1960s but have increased in popularity through these technological developments. The creation of plots for interactive movies (i.e. viewers control the plot for their individual reception experience, e.g., Shafer et al., 2016; Vorderer et al., 2001) or interactive movie events (i.e. Films or series whose plot is directed by all viewers through a majority vote in online polls, e.g., Hofer et al., 2020) have already begun. Theoretical reflection and empirical investigation, however, are lagging behind.

We are particularly interested in the latter mentioned interactive movie events, which change the individual viewer’s reception situation by extending it into a co-viewing situation – that is, a reception situation in which media consumption takes place with others (Tal-Or, 2019; Tal-Or & Tsfati, 2018). Tal-Or (2019) distinguishes three broad categories of how co-viewers affect individual viewing processes: (1) Reactions to the co-viewers’ mere presence (e.g., through divided attention), (2) reactions to the co-viewers’ reactions (e.g., to their facial expressions and verbal or non-verbal behavior), and (3) co-viewing with outgroups (e.g., increased empathy for the outgroup co-viewer). In this study, we suggest the increasing number of interactive formats as belonging to a fourth category of virtual co-viewing. Interactive movie events represent viewing situations where co-viewers are not physically present and most likely unknown to the recipient, but nonetheless affect the individual viewing experience through their behavior. In such a reception situation, reasoning about justice might not be a purely individual process but affected by the judgment of virtually present co-viewers. So far, identified factors influencing viewers’ reasoning about justice are related to characteristics of the crime drama’s content, such as the severity of the punishment (Raney, 2005), or to the individual viewer’s characteristics, such as punitiveness and vigilantism (Raney, 2002). We suggest the voting behavior of virtually present co-viewers as an additional factor influencing viewers’ judgment of justice. We move beyond existing research and theoretically discuss and empirically test context effects of virtually present co-viewers on individual viewer’s enjoyment and their potential role as a viewer’s guiding compass for reasoning about justice.

The Integrated Model of Enjoyment for Crime Drama

We build our assumptions on the integrated model of enjoyment for crime drama (Raney & Bryant, 2002) and apply it to the context of interactive movie events. The model presents a theoretical framework conceptualizing the cognitive response of moral reasoning about justice as a distinct process from affective responses during the consumption of crime drama. At the heart of the model is the viewers’ moral evaluation of a justice sequence that consists of one or more scenes in which a potentially violent and harmful crime is committed, and the offender is punished retributively. Both actions are causally related, with the latter being a consequence of the first. Thus, justice is restored by punishing injustice.

Two types of interacting variables affect how the justice sequence is judged by a viewer – audience inputs and message inputs. While watching, a viewer evaluates the actions of a character, such as the circumstances and severity of the crime, and arrives at a moral judgment about what punishment is appropriate to restore justice. To conceptualize this initial process (please see our research model in Figure 1), we build on research aiming to explain the enjoyment of morally ambiguous characters (MACs such as Dexter or Walter White in “Breaking Bads”; Janicke & Raney, 2015; Krakowiak & Oliver, 2012; Krakowiak & Tsay-Vogel, 2013). Rooted in the work of Bandura (2002), the mechanism of moral disengagement describes a set of cognitive strategies to evaluate bad or at least morally questionable behavior (Janicke & Raney, 2018). By using strategies such as justification (e.g., “he only defended himself”), displacement of responsibility (e.g., “he had no choice, he was forced by the circumstances”), or distortion of consequences (e.g., “lying doesn’t hurt anyone”), a viewer attributes external forces as responsible for a character’s behavior (Tamborini et al., 2018) and then re-interprets the character’s actions as justified in the specific case. In our case of reasoning about justice in a crime drama, for example, the recipient might justify a criminal act by saying that it saved someone else’s life or by accusing the victim of being a bad person. We thus assume that..

H1: Moral disengagement negatively predicts the level of punishment considered appropriate for a character’s criminal action.

Figure 1. Research model.

Next, the viewer compares his/her judgment of what is deserved with what is presented on screen. According to the integrated model of enjoyment for crime drama, a viewer’s enjoyment increases when justice is restored in line with the viewer’s expectations. That is, a character’s portrayed punishment is perceived as deserved for the crime committed. Vice versa, enjoyment should decrease if a plot develops in a different direction and the judgment of the justice sequence remains unbalanced.

Overall, empirical tests have predominantly supported the general proposition of the integrated model on the effects of cognitive responses to crime drama (Raney, 2002, 2005; Raney & Bryant, 2002). This was the case as long as (1) the crime was not too violent and harmful to cause ceiling effects (Raney, 2002) and (2) the crime sticked to the genre-based characteristics of crime dramas. For example, Raney (2005) found no relationship between deservedness and enjoyment when a crime (even though it was a mild one) was not punished at all. No study thus far manipulated the congruency of the recipients’ notion of justice for a committed crime and the punishment subsequently presented in the drama for the assessment of enjoyment. While Raney (2002) varied the crime committed while holding the punishment constant, Raney (2005) tested the effect of different punishments for the same crime. In line with the integrated model of enjoyment for crime drama, we assume that

H2:

Congruence between a viewer’s judgment of justice and the portrayed punishment on screen increases the viewer’s enjoyment, compared to incongruence between a viewer’s judgment of justice and the portrayed punishment.

Impact of Virtual Co-Viewers on the Judgment of Justice

Individuals have different notions about what is just and fair (Schmitt & Maes, 2013; Zillmann, 2000). Thus, a justice sequence might be evaluated as justified by some and unjustified by other viewers. In the context of interactive movie events with virtually present co-viewers (Hofer et al., 2020), we suggest that divergent judgments motivate a cognitive process of reflecting on one’s own judgment and can lead to its adjustment toward the majority opinion. The notion that people tend to conform to and follow what they believe to be the majority opinion is well-known as the bandwagon effect (Barnfield, 2019; Lee et al., 2020; Mutz, 1998; Schmitt‐Beck, 2005). Such an impersonal influence on people’s opinions has been extensively examined in political communication with regard to the effect of opinion polls, but also extended to news coverage, social media, and health contexts (Bartels, 1988; Hahn et al., 2018; Mutz, 1992; Neo, 2020; Stolwijk et al., 2016; Van der Meer et al., 2016; Zerback et al., 2015). Three explanations for the occurrence of a bandwagon effect have been offered that can be applied to the judgment of justice in crime drama. First, the persuasive effect of polls might be traced back to its function as a social cue, based on the assumption that others have more knowledge to judge a certain situation or character (Bartels, 1988; Zerback et al., 2015). For our case, the judgment of virtual co-viewers can affect recipients’ judgment of justice if they believe the aggregated decision of others is more sophisticated than their own decision. Thus, the opinion of the majority works as a proxy for the supposed most intelligent choice, such that if most people think something is the right choice, it cannot be wrong (Mutz, 1998). Second, the judgment of others might trigger cognitive processes to further deal with the issue, prompting the individual to think about reasons for the differences in judgment between themselves and the public (Mutz, 1992). This reflection potentially results in a form of self-persuasion. Finally, Fein et al. (2007) argue that the majority preference reflects a social norm, making individuals feel obliged to comply with. Thus far, empirical research mainly gained support for an impersonal influence, showing that people adjust their opinion to at least some degree toward the majority public opinion (Fein et al., 2007; Hahn et al., 2018; Moy & Rinke, 2012; Mutz, 1998; Schmitt‐Beck, 2005; Zerback et al., 2015). Adopting the assumptions of the bandwagon effect to the case of judgment of justice, we assume that viewers are more likely to adjust their judgment of justice if it differs from the majority opinion of the virtually present co-viewers.

H3:

Viewers will adjust their judgment of justice if their judgment differs from the judgment of virtual co-viewers.

Adjusting one’s initial moral judgment of what is deserved to what is portrayed on screen is a cognitive process of post-hoc reasoning and reflecting. In order to re-balance injustice and justice in line with the majorities’ evaluation of an appropriate punishment, the crime has to be reevaluated. If a viewer initially perceived a crime as particularly violent or harmful and, thus, expected a hard punishment of the offender but others’ judgment is less severe, the viewer needs to downgrade his/her initial harsh evaluation of the crime in order to conform with others’ judgment and balance his/her judgment of justice. Vice versa, if a viewer initially perceived a crime as hardly violent or harmful and, thus, expected only mild punishment, but others’ judgment is more severe, the viewer needs to upgrade his/her initial mild evaluation of the crime. Based on previous research on the enjoyment of morally ambiguous characters, we assume that the individual viewer copes with the incongruence of his/her judgment of justice and the majorities’ judgment by adapting the level of moral disengagement to justify an adaption of their initial judgment of justice (see H1) and to achieve congruence with the judgment of others.

H4:

The viewer’s adjustment of their judgment of justice is mediated by moral disengagement such that viewers who perceive no congruency with the judgment of the majority compared to viewers who perceive congruence with the judgment of the majority adapt their level of moral disengagement (H4a). The extent of adaptation predicts adjusting one’s judgment of justice (H4b).

Method

We conducted a lab experiment with N = 148 participants in a one-factorial (judgment of justice: a viewer’s notion of what is deserved is congruent or non-congruent with what is portrayed on screen) between-subjects design with an online follow-up survey after one week. Questionnaires, data, and R scripts are fully available on the Open Science Framework (https://osf.io/8b9tk).

Recruitment and Participants

Study participants were students from a German university who attended for course credit. They were recruited by fellow students on campus, through Facebook groups, and personal networks. In total, N= 151 students completed the lab session and the follow-up questionnaire. We removed three participants from the analysis due to excessive amounts of missing data and/or rapid questionnaire completion based on quality criteria indicators embedded in the survey tool used for data collection (Leiner, 2019b). The final sample consisted of N= 148 students aged 18 to 52 years (M = 22.40; SD = 4.34). The majority of the participants were females (78%).

Stimulus

The stimulus used was the German television movie “Terror – Your Verdict.” The movie was aired in October 2016 as an interactive public broadcasting television event in Germany, Austria, and Switzerland. The movie is based on a popular theater play by Ferdinand von Schirach and is inspired by the 9/11 terrorist attack: Major Lars Koch, pilot in the German armed forces, stands trial. Lufthansa-Airbus Flight LH 2047 was hijacked by terrorists on its way from Berlin to Munich with 164 passengers on board. The Airbus suddenly changed its course toward the Allianz-Arena in Munich where 70,000 people had gathered for a soccer game. In the very last minute, Lars Koch decided to shoot down the Airbus to prevent terrorists from crashing into the crowded soccer stadium. However, he did not have an order to do so and, thus, disregarded applicable law. The jury now needs to decide whether Lars Koch is guilty of murdering the 164 airplane passengers. At the beginning of the movie, the presiding judge directly addresses the audience to explain their task to serve as a juror in the trial. The judge asks the audience to listen carefully to the witnesses as at the end of the movie they will have to decide whether Lars Koch is guilty or not guilty of murder. The plot is situated solely in the court room and lasts for about 90 minutes. The legal foundation of the case is undisputed. As per the Aviation Security Act (Luftsicherheitsgesetz), it is illegal to order a shootdown or grant immunity for doing so. However, the events unfold around the moral philosophical dilemma whether it is acceptable to kill 164 people in order to save 70,000. After all witnesses are heard and all evidence is presented, the movie stops, and the presiding judge asks the audience for their verdict – guilty or not guilty. Which one is aired depends on the majority of the incoming votes from the audience. The final scene includes the reading of the verdict and is about 5 minutes long.

Procedure

Upon recruitment, the participants were asked whether they want to participate in a university research project on how audiences experience and respond to interactive entertainment formats on television. Subsequent to their consent, they were invited to watch a full-length movie in our university lab and asked to choose a two-hour lab slot that suited them best. We created a pleasant atmosphere in the lab with dimmed lights and free snack with a maximum of six participants per slot to reduce noise and unrest in the room. A trained student assistant was always present during the lab session to answer any questions or help in case of technical difficulties.

After a short introduction, the participants watched the movie on an individual laptop with headphones. The movie stopped automatically after about 90 minutes before the reading of the verdict. The participants answered a short online questionnaire on their laptop programmed with SoSciSurvey (Leiner, 2019a) measuring their moral disengagement and collecting their verdict. At the end of this questionnaire, the participants were informed that they will now watch the verdict that most of the participants had opted for in the study so far. Then they were randomly assigned to an experimental group that either watched a verdict that was congruent or one that was non-congruent with their choice. The lab session ended with a short questionnaire assessing the participants’ overall enjoyment of the movie and demographic data. We also collected the students’ e-mail address to contact them for the second part of the study. The e-mail was stored separately from the responses and held in trust by SoSciSurvey to ensure anonymity. The invitation e-mail to the online questionnaire was sent out automatically via the survey tool. One week after the initial lab session, we re-assessed the recipients’ moral disengagement as well as their verdict and asked for open comments on the study. The students were thanked and debriefed after completion of all follow-up questionnaires.

Measures

Enjoyment

Movie enjoyment was measured with four items adapted from previous studies (e.g., Janicke & Raney, 2018; Krakowiak & Oliver, 2012) on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree): (1) “Watching the movie was fun,” (2) “I enjoyed watching the movie,” (3) “Watching the movie was entertaining,” (4) “I’m glad the movie didn’t last any longer (reversed).” A Principal Axis Factoring (PAF) was conducted (rotation: oblimin) to assess the construct’s dimensionality. Item (4) was dropped due to low communalities. The remaining items formed a one-dimensional scale (KMO = .72, Bartlett test p< .001) explaining 63% of the total variance (Table 1). We assessed the reliability of the enjoyment scale with Cronbach’s alpha which showed a good fit for the sample (α = .82, M= 5.60, SD = 1.09).

Table 1. Results of the Principal Axis Factoring (PAF) for enjoyment: Factor loadings and explained variance.

	Enjoyment (N = 148)
Enjoy1 Watching the movie was fun	.840
Enjoy2 I enjoyed watching the movie	.727
Enjoy3 Watching the movie was entertaining	.803
Explained variance (%)	63%

Moral Disengagement

We developed nine context-related items measuring moral disengagement for the specific plot on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree). Our proceeding was guided by the studies of Janicke and Raney (2018) as well as Krakowiak and Tsay-Vogel (2013). Sample items include “If Lars Koch hadn’t shot down the plane, 70,000 people would have died” and “He had no other choice, since no one else reacted except for him.” The participants’ moral disengagement was measured twice in the study. It was first measured in the lab session after the participants watched the movie but before watching a congruent or non-congruent verdict. We re-assessed the participants’ moral disengagement in the follow-up questionnaire. A Principal Axis Factoring (PAF) was conducted (rotation: oblimin) for both measurement times. Four items had to be dropped due to low communalities. It was expected that the proportion of some items’ variance explained by the factors would be low as we could not make use of a validated scale but had to develop items that were specifically suitable for our stimulus. A one-dimensional structure was consistently confirmed for the five remaining items for moral disengagement measured in the lab session (KMO = .72, Bartlett test p< .001) and follow-up (KMO = .82, Bartlett test p< .001) explaining 36% and 46% of the total variance (Table 2). We assessed reliability of both scales again with Cronbach’s alpha which resulted in a good fit for the first (α = .73, M= 5.07, SD = 1.06) and for the second measurement (α = .80, M= 5.22, SD = 1.11).

Table 2. Results of the Principal Axis Factoring (PAF) for moral disengagement: factor loadings and explained variance prior to watching a verdict and in the follow-up assessment.

	Moral Disengagement_pre (N = 148)	Moral Disengagement_post(N = 148)
MD1 If Lars Koch hadn’t shot down that plane, 70,000 more people would have died.	.701	.575
MD2 He had no other choice, since no one else reacted but him.	.600	.788
MD3 Lars Koch is the only person who took responsibility in this dilemma.	556	.608
MD4 If Lars Koch had acted differently; they would have surrendered to the terrorists.	.551	.711
MD5 If he had decided otherwise, it would have been failure to render assistance.	.584	.692
Explained variance (%)	36	46

Judgment of Justice

The participants’ judgment of justice was measured with a single question: “Now that you have seen the movie, we are interested in the verdict you would make.” The participants could choose between “Lars Koch is guilty, he should be convicted” or “Lars Koch is innocent, he should be acquitted” depending on what they believed is justified for shooting down the Airbus. The vast majority of the participants pleaded “not guilty” (86%). This was comparable to the TV vote when the movie event was aired in 2016. This unequal distribution does however not affect our experimental manipulation as random assignment was not based on the students’ actual vote choice but assigned the students an ending that was either congruent or non-congruent with their choice – regardless of their initial vote, resulting in balanced experimental groups (congruent, n = 72; non-congruent, n = 76). The participants’ verdict was re-assessed in the follow-up one week later with the question “Which judgment would you make today?” with the same response options. And again, a majority (85%) pleaded “not guilty.” No differences were observed between groups with regard to age (F(1, 146) = 1.52, p = .22) and gender (p= .80, Fisher’s exact test).

Results

Data analyses were conducted in R version 4.0.1 (R Core Team, 2020). Our first hypothesis addressed the viewers’ initial judgment of justice. H1 stated that moral disengagement negatively predicts the level of punishment considered appropriate for a character’s criminal action. A binary logistic regression was conducted with the level of moral disengagement measured prior to watching one of the two verdicts as independent variable and the participants’ initial judgment (guilty, not guilty) as dependent variable. Table 3 summarizes the results of the binary logistic regression. As predicted, moral disengagement had a significant impact on the recipients’ initial judgment of justice (p< .001). The greater the level of moral disengagement, the greater the likelihood of judging Lars Koch as “not guilty.”

Table 3. Impact of moral disengagement on the recipients’ judgment of justice.

	Estimate	SE	z	Exp(B)	p
(Intercept)	−5.26	1.55	−3.40	.005	<.001
Moral disengagement (pre)	1.52	.34	4.46	4.555	<.001

N = 148, moral disengagement measured prior to watching one of the two endings, AIC = 88.87.

H2 assumed that congruence between a viewer’s judgment of justice with the portrayed punishment on screen increases the viewer’s enjoyment, compared to incongruence between a viewer’s judgment of justice and the portrayed punishment. A one-way ANOVA was conducted to determine whether the enjoyment of viewers who watched the verdict they had voted for (e.g., voted Lars Koch as “guilty” and watched “guilty” verdict) differed from the enjoyment of viewers who watched the opposite verdict (e.g., voted Lars Koch as “guilty” and watched “not guilty” verdict). Results are in line with the theoretical assumptions of the integrated model of enjoyment for crime drama and confirm hypothesis H2 (F(1, 146) = 6.71, p= .011, η² = .04). Viewers who watched the verdict that was in line with their notion of what Major Koch deserved for shooting down the airbus enjoyed the movie more (n = 72, M= 5.83, SD = 0.94) than viewers who watched the opposite verdict (n= 76, M= 5.38, SD = 1.18).

Our final Hypotheses 3 and 4 set out to explore the impact of the judgment of virtually present co-viewers on viewers’ judgment of justice. Following the band wagon heuristic, H3 assumed that viewers will adjust their judgment of justice if their judgment differs from the judgment of virtual co-viewers. Whether the viewers adjusted their judgment of justice was determined by comparing their initial judgment during the lab session before watching one of the two verdicts and their judgment one week later in the online follow-up questionnaire. The resulting difference score is a binary variable where “1” represents a change and “0” represents no change compared to the viewers’ initial judgment of justice. Overall, the viewers hardly changed their initial judgment. Only five participants changed their verdict to the opposite. However, they all belonged to the experimental group that watched a verdict opposite to their initial notion of what was deserved (Table 4).

Table 4. Cross reference table displaying the number of participants changing their verdict in T2 one week after the reception of the movie compared to T1 according to their experimental condition.

		Congruency of Verdict with own Choice
		Non-congruent	Congruent
Adjustment of participants’ judgment of justice	No adjustment	71	72
	Adjustment	5	0

N = 148.

The small proportion of participants who changed their judgment is best described as a “rare event” in the data, such as a rare disease that affects only a few people in the entire population (King & Zeng, 2017). The problem of such a strong asymmetry in the outcome variable is that the maximum likelihood estimation of the logistic model that would usually be applied to such data produces a bias in the predicted probabilities. The Firth method (Firth, 1993) uses a penalized likelihood to reduce small-sample bias in maximum likelihood estimation and was proposed as solution for the rare event problem (Firth, 1993; Heinze & Schemper, 2002). We used the “logistf” R package (Heinze et al., 2020) to perform the adjusted logistic regression with the Firth method to test our assumptions. The watched judgment (0 = “non-congruent,” 1 = ’congruent’ [with own judgment]) was introduced as independent variable, and the created binary variable for the adjusted judgment (0 = ’no change’, 1 = ’change’) as dependent variable. The results confirm the descriptive findings (Table 5). Respondents who watched a verdict that differed from the majority judgment of the virtual co-viewers were 11 times more likely to adjust their judgment toward the perceived majority judgment than respondents whose judgment was in line with the majority judgment (p= .029). Thus, the findings are in line with the prediction made by the bandwagon effect and confirm H3. However, results should be interpreted with care and require replication due to the small number of viewers who changed their initial verdict.

Table 5. Impact of the virtual co-viewers majority judgment on the adjustment of the individual viewer’s judgment of justice.

	Estimate	SE	95% Conf. interval	Chisq	Exp(B)	p
(Intercept)	−2.56	0.45	[−3.56, −1.79]	67.40	.077	<.001
judgment of virtual co-viewers (0 = non-congruent)	−2.41	1.50	[−7.30, −0.21]	4.78	.090	.029

N = 148, Model fitted by penalized ML, Likelihood ratio test = 4.78 (p= .0289), Wald test = 2.6 (p= 0.107).

And finally, we assumed that the viewers’ adjustment of their judgment of justice is mediated by moral disengagement such that viewers who perceive no congruence with the judgment of the majority compared to viewers who perceive congruence with the judgment of the majority adapt their level of moral disengagement (H4a). The extent of adaptation in turn predicts adjusting one’s judgment of justice (H4b). We created a variable recording the changes in the level of moral disengagement by computing the difference of the level of moral disengagement prior to making the initial judgment and the level of moral disengagement one week after the lab session. We expected that respondents who watched a different verdict than they had voted for would adapt their level of moral disengagement. Whether the respondents strengthen or weaken their initial moral disengagement depends on their initial judgment. In this study, we expected them to strengthen their level of moral disengagement if they pleaded “guilty” but watched the “not guilty” verdict. Vice versa, we expected the respondents to weaken their moral disengagement when they pleaded “not guilty” and watched the “guilty” verdict. We did not expect viewers who watched the verdict they had voted for to adjust their level of moral disengagement. Hence, we obtained the absolute value of the changes in the level of moral disengagement as the magnitude of the adjustment – not its direction – affects the changing judgment. We tested a mediation model with the congruency of the viewers’ own judgment and the majority judgment of co-viewers as independent variable (X), the absolute value of the changes in moral disengagement as mediator (M), and the adjusted judgment as dependent variable (Y). Table 6 summarizes the results of the mediation model. Results show that respondents who watched a verdict non-congruent with their own choice adapted their level of moral disengagement to a slightly greater extent (M= .70, SD = .69) than respondents in the congruent condition (M= .55, SD = .44). However, the difference did not reach statistical significance (a = −.147, p= .120). As well, the greater the range of the adjustment in moral disengagement, the more likely was the adjustment of judgment, but the effect again did not reach significance (b = .043, p= .264). While the total effect c of the model is significant (c = .066, p= .019), the adjusted level of moral disengagement was not confirmed as mediator in the relation of judgment of others and own adjusted judgment (a_xb = −.006, p= .372). Thus, hypotheses H4a and H4b are rejected.

Table 6. Adjusted moral disengagement as moderator (M) in the relation of virtual co-viewers majority judgment (X) and adjusted judgment of justice (Y) in a mediation model.

	Adjusted moral disengagement (M)				Adjusted judgment of justice (Y)
Antecedent		β	SE	p		β	SE	p
judgment of virtual co-viewers (X), (0 = non-congruent)	a	−.147	.095	.120	c‘	−.060	.026	.021
Adjusted moral disengagement (M)		– –	– –	– –	b	.043	.039	.264

Displayed are standardized regression coefficients. Total effect c = −.066, p= .021.

Discussion

The goal of this study was to shed light on viewers’ enjoyment and reasoning about justice against the backdrop of increased popularity of interactive plots and formats. We based our study on the integrated model of enjoyment for crime drama (Raney & Bryant, 2002) and suggest that virtually present co-viewers are a further factor influencing viewers’ enjoyment and reasoning about justice. In line with the integrated model of enjoyment for crime drama and previous studies on the enjoyment of MACs (Janicke & Raney, 2015; Krakowiak & Oliver, 2012), we show that viewers’ moral disengagement predicts viewers’ judgment of justice – that is the evaluation of a justice sequence – of what punishment is appropriate for a committed crime (Raney & Bryant, 2002). In addition, despite acknowledging small effect sizes (η² = .04), we find that the congruence between what the viewer believes to be a fair verdict and the verdict the person on trial actually receives affects the viewers’ enjoyment. Enjoyment dropped significantly when a viewer’s judgment conflicted with the actual verdict. Retributive action which does not correspond to one’s own idea of justice apparently cannot be ignored by recipients and thus has effects on their enjoyment of the movie. Moreover, if viewers were confronted with a majority judgment of other viewers that was not in line with their own judgment, they were more likely to change their judgment toward the majority judgment. Being confronted with an opposing majority opinion regarding the appropriate punishment prompted the recipients to reevaluate their initial decision and engage in post-hoc reasoning. In some cases this resulted in an adjustment of their original judgment due to cognitive self-persuasion or the belief that the majority cannot be wrong which is in line with theoretical assumptions of a bandwagon effect (Mutz, 1992, 1998; Zerback et al., 2015). However, it has to be noted that only a small portion of respondents indeed changed their verdict. Overall, results point to the fact that in sum the judgments of justice are rather stable. The participants’ interpretation of “majority” might also play a role. We only informed our participants that the verdict they were about to watch was voted for by a majority of the study participants to date. Changing their judgment one week after might have been affected by their perception of how large the majority was, for example, through follow-up conversations with fellow participants or even newly gained information about the voting turn-out of the original movie. Replication, especially with different stimuli, is needed to validate our findings.

Theoretical and Conceptual Implications

This is the first study to actively manipulate the congruency of the viewers’ notion of what is deserved for a crime committed with the punishment shown on screen. The results fully support the integrated model of enjoyment for crime drama and extend it to interactive TV formats (Raney & Bryant, 2002). We also widen the view on potential influencing factors on viewers’ enjoyment and judgment of justice by considering virtually present co-viewers against the backdrop of technological advancement in home entertainment. While media effects research is predominantly interested in individual processes of watching television, context effects of co-viewing are increasingly gaining scholarly attention (e.g., Tal-Or, 2019; Tal-Or & Tsfati, 2018; Weber & Quiring, 2019). These studies focus on the effects of co-viewing as a social activity (Cohen, 2017). Tal-Or (2019) distinguishes three broad categories of how co-viewers affect individual viewing processes. (1) Reactions to the co-viewers’ mere presence, (2) reactions to the co-viewers’ reactions, and (3) co-viewing with out-groups. In this study, we suggest the increasing number of interactive formats as a fourth category of co-viewing. We find that the judgment of others serves as a justice compass for a viewer’s reasoning about justice. Hence, the bandwagon effect which has so far predominantly been analyzed in political contexts (Bartels, 1988; Mutz, 1998; Schmitt‐Beck, 2005) can also be applied to entertaining media.

Limitations and Future Studies

This study has some limitations that provide perspectives for future studies. Because we gain insights based on a lab experiment where participants watched a full movie instead of short video clips, we compensated this effort in data collection with a single stimulus movie. Therefore, our study requires replication to test the generalizability of our results. More specifically, the vast majority, namely 86% of the participants, voted for “not guilty” suggesting that the storyline might have led to a one-sided view, for example, by creating a very likeable character or casting a very popular actor. This leads to an asymmetry in the experimental groups. While most of the participants of the congruent group watched the not guilty verdict, most participants in the non-congruent group watched the guilty verdict. Future studies could look into stimuli that generate the opposite or a more balanced distribution of the perceived deserved retributive punishment indicating a different or even more difficult moral dilemma. Moreover, as the congruence of what is deserved and what is portrayed on screen is manipulated with only two options, future studies could use self-report measures to assess the evaluation of a justice scene in more detail. Further, this study used a rather simple measure of enjoyment as dependent variable which might be one factor that accounts for the small effect sizes. While the scale has been confirmed as reliable, it does not reflect the differentiation between hedonic and eudaimonic responses (Oliver & Bartsch, 2011; Oliver & Raney, 2011). While our study taps more on the hedonic aspect of entertainment, future studies should also measure eudaimonic aspects as greater levels of appreciation and meaningfulness have been found for narratives that require a greater amount of reasoning and reflection during the reception process (Lewis et al., 2014). Furthermore, the data do not lend support to moral disengagement as a relevant mediator of the relationship between judgment of virtual co-viewers and a viewer’s adjusted judgment. Given the small variance in the dependent variable and limited power this did not come by surprise. Nonetheless, further studies should also consider factors beyond moral disengagement that could account for an adjustment of the participants’ judgment of justice. The bandwagon effect points to others’ behavior as social cues and social norms that are perceived by others which have been found to be a strong predictor of individuals’ behavior (e.g., Theory of Normative Social Behavior, Rimal & Real, 2005). Future studies should also consider social desirability as an intervening factor. Possibly we were not able to provide significant evidence of a greater adjustment of the initial judgment as the participants might have been motivated to demonstrate consistency in their judgment.

Conclusion

The increasing popularity of interactive plots requires theoretical reflection and empirical examination. The focus of this study is on viewers’ enjoyment and reasoning about justice which are both substantially affected by the majority judgment of virtual co-viewers. More research is needed on identified mechanisms and potential influencing factors against the background of ongoing developments in the entertainment industry that produces new formats to attract and connect viewers.

Disclosure Statement

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

Additional information

Notes on contributors

Corinna Oschatz

Corinna Oschatz (Ph.D, University of Mainz) is an Assistant Professor of Political Communication & Journalism at the Amsterdam School of Communication Research, University of Amsterdam, Amsterdam, Netherlands. Her research interests include narrative persuasion, political communication, and methods of empirical social science.

Katharina Emde-Lachmund

Katharina Emde-Lachmund (M.A., Hanover University of Music, Drama and Media) is a Research Assistant at the Department of Journalism and Communication Research at the Hanover University of Music, Drama and Media, Germany. Her research interests include media psychology, political communication, and methods of empirical social science.

Marsha Helmstädter

Marsha Helmstädter (B.Sc., University of Koblenz-Landau) is a student of Developmental and Clinical Psychology at the University of Heidelberg. Her research interests include experimental psychopathology, narrative persuasion and substance use disorder treatment.