Abstract
From retail and transport to hospitality and live events, customer-to-customer (C2C) interactions are integral to many service experiences. However, C2C interactions are not always positive, and customers regularly become targets of other customers’ misbehavior. Although such C2C misbehavior has always been a challenge for firms, it has escalated since the COVID-19 pandemic, with incidents becoming not only more frequent but also more intense and severe. However, it remains unknown how customers assess the magnitude of C2C misbehavior. Four studies, including 62 repertory grid interviews using scenario-based and diary-based approaches with international participants from Germany and the United Kingdom, reveal six dimensions that underlie customers’ magnitude perceptions: three intensity-related dimensions that assess the magnitude of the misbehavior itself and three severity-related dimensions that examine the magnitude of the misbehavior’s impact on other customers. Several sorting studies with 99 U.S. participants support the robustness, generalizability, distinctiveness, and hierarchical structure of our identified magnitude dimensions. By uncovering the facets that underlie customers’ evaluations of an incident’s perceived magnitude, our study provides an empirically grounded conceptualization and typology of C2C misbehavior. Managerially, this research enables firms, communities, and governments to monitor and categorize misbehavior incidents more effectively and to develop more targeted mitigation strategies.
Keywords
Introduction
Driven by technological advances, the rise of the sharing economy, and enhanced customer connectivity, service encounters today are increasingly characterized by customer-to-customer (C2C) interactions (Furrer, Landry, and Baillod 2024). These C2C interactions, whether online or offline, are integral to the consumption experience across various service settings, such as retail, hospitality, leisure, transportation, online gaming, and social networks (Libai et al. 2010; Nicholls 2010). Extant research highlights the importance of C2C interactions for business performance, given their potential to shape customers’ experiences (e.g., Colm, Ordanini, and Parasuraman 2017; Heinonen, Jaakkola, and Neganova 2018), emotions (e.g., Baker and Kim 2018; Miao and Mattila 2013), satisfaction (e.g., Martin 1996; Martin and Pranter 1989), and loyalty (e.g., Moore, Moore, and Capella 2005).
However, C2C interactions are not always positive. Negative service experiences often arise when customers become the targets of other customers’ misbehavior. Defined as “any behavioral act by a customer directed against other customers personally or against shared resources that violates prevalent social norms in a service setting” (Danatzis and Möller-Herm 2023, p. 459), such C2C misbehavior can take many forms. Examples range from littering (Schaefers et al. 2016) and territorial behaviors (Griffiths and Gilly 2012) to verbal or physical harassment (Grove et al. 2012). Although C2C misbehavior is nothing new, it has notably escalated since the COVID-19 pandemic (Khazan 2022; Soares et al. 2023). From unruly tourists in Spain and Bali (Hall 2024) and incidents of customers contaminating food in Japanese sushi restaurants (Khalil 2023) to aggressive and disruptive theater audiences on Broadway (Thorp 2023) and antisocial public behavior in the United Kingdom (UK Government 2025), firms, communities, and governments worldwide are grappling not only with a heightened frequency of C2C misbehavior incidents but also with their increased intensity and severity.
Misbehavior intensity (the perceived strength of the C2C misbehavior itself) and impact severity (the perceived strength of its consequences on other customers) are two crucial, yet distinct and underexplored facets of C2C misbehavior. Together, they shape customers’ overall magnitude perceptions of C2C misbehavior, yet they do not necessarily align. For example, low-intensity C2C misbehavior, such as eating smelly food on trains, passive-aggressive remarks in a checkout line, or subtle microaggressions against minority customers can severely impact those affected, often causing significant disruption or lasting emotional harm (Martin 1996; Sue et al. 2007; Yin and Poon 2016). Conversely, high-intensity C2C misbehavior, such as singing or dancing on public transport, talking loudly in a quiet zone, or smoking excessively in non-smoking areas (Danatzis and Möller-Herm 2023; Huang 2008), may not be perceived as severe if it causes minimal harm or disruption to others.
Given the traditional focus on customer misbehavior directed at firms or employees (e.g., Bitner, Booms, and Mohr 1994), prior research has paid less attention to misbehavior among customers and the magnitude of such behavior. Yet emerging evidence suggests that magnitude perceptions of misbehavior intensity can negatively affect customers’ conformity to norms or rules (Keizer, Lindenberg, and Steg 2008; Schaefers et al. 2016). Similarly, magnitude perceptions of impact severity have been found to decrease satisfaction (Huang 2008) while exacerbating its contagiousness—that is, the degree to which misbehavior spreads among customers (Danatzis and Möller-Herm 2023). However, it remains unclear what underpins these magnitude perceptions. Clearly distinguishing and accounting for both intensity- and severity-related magnitude perceptions is therefore of utmost managerial importance for enhancing customer experiences and designing mitigation strategies.
Compounding this gap is a prevailing dichotomy in definitional perspectives, which primarily frame customer misbehavior either as a violation of norms or as harm to others (Fisk et al. 2010). While analytically useful, this binary lens has contributed to fragmented conceptualizations of customer misbehavior over time and risks obscuring other relevant dimensions that may shape its perceived magnitude. What is more, prior work tends to focus on identifying misbehavior incidents per se rather than examining the constituent elements that shape their perceived magnitude. Hence, this paper seeks to answer the following research question: Which facets underlie the perceived magnitude of C2C misbehavior?
To address this question, we conducted four studies to explore the facets underlying C2C misbehavior magnitude perceptions, investigate their hierarchical structure, and derive a magnitude-based typology of C2C misbehavior. Using Kelly’s (1955) repertory grid technique (RGT), we carried out several scenario-based and diary-based studies, including 62 in-depth interviews with international participants from Germany and the United Kingdom. Our findings reveal six magnitude dimensions of C2C misbehavior. Intensity-related magnitude dimensions include (1) rule and norm deviation (i.e., the extent to which the misbehavior itself is perceived to violate formal rules or social norms), (2) deliberate advantage-taking (i.e., the extent to which the misbehavior itself is perceived as a conscious attempt by the perpetrator to gain personal advantage), and (3) attitudinal influences (i.e., the extent to which judgments of the misbehavior itself are shaped by personal attitudes). Severity-related magnitude dimensions, in turn, entail (4) disruption of the service experience (i.e., the extent and nature of disruption caused by the misbehavior to the service experience), (5) harm caused (i.e., the extent, type, and timing of harm caused by the misbehavior), and (6) loss of control (i.e., the extent to which the misbehavior causes one to feel unable to manage, influence, or escape its consequences). Several open- and closed-card sorting tasks with 99 independent U.S. judges provide cumulative support for the robustness, generalizability, distinctiveness, and hierarchical structure of the six magnitude dimensions and their two higher-order categories of misbehavior intensity and impact severity. Based on these insights, we discern four key types of C2C misbehavior rooted in magnitude perceptions.
Theoretically, this research advances prior work on C2C misbehavior (e.g., Bernritter et al. 2025; Danatzis and Möller-Herm 2023; Schaefers et al. 2016) and customer misbehavior more broadly (e.g., Fisk et al. 2010; Fullerton and Punj 2004) by providing the first magnitude-based conceptualization and typology of C2C misbehavior. Notably, our findings are not limited to a specific industry or country and go beyond prevailing norm- and harm-based perspectives in the extant misbehavior literature (e.g., Bitner, Booms, and Mohr 1994; Fombelle et al. 2020; Huang 2008). Instead, we uncover a comprehensive portfolio of intensity- and severity-related magnitude dimensions of C2C misbehavior, examine their dimensionality, and derive a corresponding typology. In doing so, we reconcile previously fragmented conceptual aspects of customer misbehavior (Fisk et al. 2010), thus providing a robust theoretical foundation for future empirical research in this area. Managerially, our findings equip firms, communities, and governments to monitor and categorize misbehavior incidents more effectively and to develop more targeted mitigation strategies.
Conceptual Background
C2C Misbehavior
C2C misbehavior is a form of C2C interaction (CCI) that refers to “customers influencing other customers through service interactions occurring in the firm and customer domains” (Heinonen, Jaakkola, and Neganova 2018, p. 722). Research on CCI has received substantial attention (e.g., Furrer, Landry, and Baillod 2024), although it has primarily focused on positive outcomes. Fewer studies address negative CCI (Fombelle et al. 2020; Nicholls and Mohsen 2019) resulting in negative outcomes for other customers, such as embarrassment (Kim and Yi 2017), anxiety (Johnson and Grier 2013), or dissatisfaction (Martin 1996).
Beyond C2C interactions, where customers directly impact each other, research on customer copresence (e.g., Argo, Dahl, and Manchanda 2005) explores the indirect influence that customers may exert through their mere presence. These studies examine instances of observing other customers misbehave (Benkenstein and Rummelhagen 2020) or suffering unintended consequences due to customer density and crowding (Hui and Bateson 1991).
In turn, research on C2C misbehavior—and customer misbehavior more broadly—focuses exclusively on negative, often intentional, customer behavior. Individuals exhibiting such behavior are often also referred to as dysfunctional customers (Harris and Reynolds 2003), jaycustomers (Lovelock 1994), misbehaving consumers (Fullerton and Punj 2004), or problem customers (Bitner, Booms, and Mohr 1994). Recognized as a costly, everyday problem for service firms (Fisk et al. 2010), various studies have explored its determinants (e.g., Wirtz and Kum 2004), types (e.g., Bitner, Booms, and Mohr 1994; Fullerton and Punj 2004), and negative consequences (e.g., Harris and Reynolds 2003; van Kleef et al. 2015).
Prior definitions of customer misbehavior typically adopt one of two perspectives: they focus either (1) on the violation of social norms or (2) on the harm caused by such behaviors (see Web Appendix 1 for an overview). The norm-based perspective centers on the behavior of the perpetrator and the extent to which it violates socially constructed standards of exchange in a given service setting. In contrast, the harm-based perspective adopts a victim-oriented lens, stressing the impact the misbehavior has on others. The tensions between these two different definitional perspectives and their implications for studying customer misbehavior have long been highlighted in prior service research (e.g., Fisk et al. 2010).
The same definitional dichotomy is evident in studies focusing explicitly on C2C misbehavior (see Web Appendix 1). Most studies adopt a harm-based perspective, stressing the damage or disruption the misbehavior causes to other customers. For instance, Huang (2008, p. 522) discusses other-customer failures, which occur “when any action by another customer has a negative impact on one’s service experience.” Similarly, Gursoy, Cai, and Anaya (2017, p. 2341) describe disruptive service behavior as “an act by a customer that negatively affects the service experience of other customers.” In contrast, other studies adopt a norm-based perspective focusing on the perpetrator’s actions. For instance, Schaefers et al. (2016) draw on Fullerton and Punj’s (2004) norm-based view and introduce the notion of “indirect customer misbehavior,” referring to norm-violating behavior “directed at the accessed product and occur[ring] in the absence of others” (p. 3; see also Srivastava, Jayasimha, and Sivakumar 2022). Encompassing both direct and indirect C2C misbehavior, Danatzis and Möller-Herm (2023) expand this norm-based perspective and define C2C misbehavior as any norm-violating behavior targeting other customers personally or their shared resources (see also Bernritter et al. 2025; Danatzis, Möller-Herm, and Herm 2024).
While prior research offers valuable insights, the prevailing definitional dichotomy has led to fragmented conceptualizations of C2C misbehavior. By focusing on either the nature of the misbehavior itself (i.e., the norm-based perspective) or its impact on others (i.e., the harm-based perspective), existing studies often provide only a partial understanding of C2C misbehavior. Furthermore, the focus on either norms or harm risks overlooking other potential dimensions of C2C misbehavior. Prior research also tends to focus on identifying instances of C2C misbehavior, with less attention given to the underlying facets that shape perceptions of its magnitude. What makes a C2C misbehavior appear more “minor” or “major” to customers, however, offers valuable insights both for enhancing customer experiences and for guiding more informed and targeted mitigation strategies (Danatzis and Möller-Herm 2023).
C2C Misbehavior Magnitude
C2C misbehavior can generally be experienced at different levels of magnitude. These magnitude perceptions are inherently subjective and influenced either by the nature of the misbehavior itself or by its impact on others. Drawing on parallels from research on natural disturbances (Iwasaki and Noda 2018) and pain perceptions (Iliffe et al. 2009), the magnitude of C2C misbehavior can similarly be understood along different degrees of misbehavior intensity and impact severity. Intensity generally refers to the strength of a phenomenon or behavior (Iwasaki and Noda 2018)—for example, the magnitude of an earthquake, the wind speed of a hurricane, or the force of a given act of C2C misbehavior. Severity, in turn, generally refers to the subsequent impact of a phenomenon or behavior (Iwasaki and Noda 2018; Iliffe et al. 2009), such as the fatalities caused by an earthquake, the material damage from a hurricane, or the disruption caused by an act of C2C misbehavior. Perceptions of misbehavior intensity and impact severity can therefore be understood as two fundamental categories of C2C misbehavior magnitude perceptions, each ranging from low to high degrees. Misbehavior intensity pertains to the perceived magnitude of the misbehavior itself, while impact severity relates to the magnitude of its perceived consequences. The two are conceptually independent, allowing for both contrasting and co-occurring levels of misbehavior intensity and impact severity in a given incident. Consider, for example, territorial behavior in a café. The displayed behavior can range from subtly spreading one’s belongings to an adjacent seat (low-intensity misbehavior) to assertively claiming a shared space (high-intensity misbehavior). Either misbehavior may be perceived as causing minor inconvenience (low-severity impact) or major disruption (high-severity impact) to other customers.
It is crucial to differentiate between the perceived intensity of the misbehavior itself and the severity of its impact as both magnitude aspects do not necessarily align. For example, high-intensity misbehavior—such as singing, cheering, or dancing on public transport, talking during a movie in the cinema, or taking phone calls in quiet zones of trains or co-working spaces—may be perceived as irritating yet cause only minimal harm or disruption to others (Danatzis, Möller-Herm, and Herm 2024; Huang 2008). Low-intensity misbehavior, in contrast—such as eating smelly food on trains, making passive-aggressive remarks in a checkout line, or engaging in subtle racial microaggressions—can severely impact those affected, often causing significant disruption or lasting emotional harm that may extend well beyond the immediate encounter (Martin 1996; Sue et al. 2007; Yin and Poon 2016).
Notably, given their subjective nature, both perceptions of misbehavior intensity and impact severity can vary across customers (Benkenstein and Rummelhagen 2020; Gursoy, Cai, and Anaya 2017). Multiple customers may perceive the same misbehavior differently, judging its perceived magnitude along a “severity [and intensity] continuum” (Karelaia and Keck 2013, p. 784). Previous research has employed multiple scales to quantify these variations in magnitude perceptions. For example, to assess misbehavior intensity, studies have measured the perceived wrongness of the act (Vitell and Muncy 1992; Wilkes 1978), its acceptability (Neale and Fullerton 2010), or the degree of norm deviance (Reynolds and Harris 2009). To assess impact severity, prior research has measured the degree of perceived loss (Huang 2008; Reynolds and Harris 2009), harm (Danatzis and Möller-Herm 2023), or influence on other customers (Yin and Poon 2016; Zhang, Beatty, and Mothersbaugh 2010).
Despite these quantification measures, research that systematically explores the nature and structure of misbehavior magnitude perceptions remains sparse. Jones (1991) is among the few to explore intensity-related magnitude dimensions, identifying factors that shape how morally intense individuals perceive a misbehavior to be. Similarly, Vitell (2003) suggests that the degree to which a behavior is judged to be ethical depends on the perpetrator’s perceived role in committing the act (active vs. passive), its perceived illegality, and the extent of resulting harm. Likewise, Dootson et al. (2016, p. 750) introduce the notion of a deviance threshold, “a mental line in the sand” that separates “right” from “wrong” behavior.
Although these studies provide rich initial insights, prior research has primarily focused on binary judgments of whether other customers approve or disapprove of a behavior, rather than on the facets that shape perceptions of its magnitude. In other words, while extant studies offer valuable guidance on what leads customers to classify a given behavior as misbehavior or not, they have yet to identify the specific attributes and dimensions that determine how intense a misbehavior or how severe its impact is perceived to be. Hence, to understand how customers judge the magnitude of C2C misbehavior, it is essential to explore the various magnitude facets that underlie perceptions of misbehavior intensity and impact severity.
Overview of Studies
We conducted four studies to explore the facets that underlie C2C misbehavior magnitude perceptions, investigate their hierarchical structure, and derive a magnitude-based typology of C2C misbehavior. Using the RGT across two qualitative studies, Study 1 employs two different research designs (Study 1A: scenario-based approach, and Study 1B: 4-week diary study), involving 49 in-depth interviews with German residents to uncover the attributes that differentiate C2C misbehaviors in terms of their perceived magnitude and to examine their dimensionality. To test the robustness and generalizability of Study 1’s findings, Study 2 uses a 6-week diary study with 13 international participants in the United Kingdom to explore potential cultural differences and an alternative RGT technique. By systematically combining different RGT approaches (scenario-based and diary-based designs) and elicitation techniques (classical triadic “minimal context” and “full context” forms), we reduce method-specific bias, triangulate findings, and increase the trustworthiness of our results. We further use sorting tasks to assess the conceptual distinctiveness and hierarchical structure of insights generated in RGT interviews. Specifically, Study 3 employs an open sorting task with 39 U.S. participants to examine whether the higher-order categories of misbehavior intensity and impact severity identified in Studies 1 and 2 naturally emerge inductively when customers assess the C2C misbehavior’s magnitude. Finally, Study 4 employs a complementary deductive approach, using a closed sorting task with 60 U.S. participants to corroborate the dimensionality and distinctiveness of C2C misbehavior magnitude perceptions. Based on the empirical insights of these four studies, we discern four key types of C2C misbehavior that are rooted in customers’ magnitude perceptions of misbehavior intensity and impact severity.
Study 1A/B
The goal of Study 1 was to empirically derive the facets underlying customers’ perceptions of C2C misbehavior magnitude. To this end, we conducted two qualitative studies involving 49 in-depth interviews with German residents who had previously encountered C2C misbehavior. Study 1A used a scenario-based approach to analyze 36 C2C misbehavior incidents across 13 service settings, while Study 1B employed a 4-week diary study to capture customers’ first-hand experiences of C2C misbehavior. Combining both approaches allowed for triangulation of insights across methods: while the scenario-based design of Study 1A systematically captures variation across service contexts, the diary-based Study 1B offers depth and ecological validity through rich, real-life experiences. Together, they mitigate method-specific biases and enhance the credibility and generalizability of the findings. Following best practices for multilingual research (Colm, Ordanini, and Parasuraman 2017), data collection and analysis were conducted in the interviewees’ native language (mostly German) before translating the results into English. This approach ensured qualitative rigor and preserved the inherent meaning of respondents’ statements (Twinn 1997).
Research Design
Study 1A
For Study 1A, we conducted 30 in-depth interviews with German residents who had previously encountered C2C misbehavior as a selection criterion. The sample includes a balanced gender split (60% female) and a wide range of ages (Mage = 39.9 years) and socioeconomic backgrounds (see Web Appendix 2a). Participants were recruited via convenience sampling, and interviews were conducted face-to-face in 2021, either in researchers’ offices or nearby public spaces. Interviews lasted between 37 and 96 minutes (M = 61.8 minutes), with interviewees recruited from two German metropolitan areas. A total of 31 hours of interview recordings were transcribed verbatim, resulting in 354 pages of single-spaced text.
The interviews were based on 36 hypothetical scenarios depicting distinct C2C misbehavior incidents adapted from previous research (see Web Appendix 3). To ensure generalizability, the scenarios were carefully selected to reflect a broad range of (a) service contexts and (b) misbehavior magnitudes. Specifically, the 36 scenarios spanned 13 service settings characterized by a high degree of customer copresence and CCI, grouped into six overarching service contexts: leisure activities (movie theater, sports event, bowling center, theme park), retail (supermarket, fashion store), mobility (bus, train, carsharing), library services, hotels, and restaurants. We further assessed each scenario for its magnitude (low, medium, or high). To avoid respondent fatigue, we created three sets of 12 scenarios, with each set used in 10 interviews. Each set included scenarios across all six service contexts and a balanced mix of low-, medium-, and high-magnitude C2C misbehaviors.
Study 1B
Study 1B adopted a diary study to capture customers’ first-hand experiences of C2C misbehavior in their daily lives. A total of 23 business communication students at a German university agreed to take part in the diary study, which was followed by an in-depth interview. To enhance completion rates, participants who completed both components received a €50 Amazon voucher as a monetary incentive. In total, 19 participants completed the diary study and the subsequent interview (73.68% female, Mage = 23.05 years; see Web Appendix 2b).
Following established guidelines (Janssens et al. 2018; Ohly et al. 2010), the diary study employed an event-sampling approach, where participants were asked to report C2C misbehavior incidents immediately after they occurred or, if impractical, as soon as possible thereafter. Over a 4-week period starting in November 2024, participants were asked to document around 10 C2C misbehavior incidents of varying magnitude across different service settings using an online form in Qualtrics. To facilitate submissions, participants were sent regular reminders and encouraged to bookmark the link to the online form in their mobile browsers for quick access. Submissions followed an open format (Ohly et al. 2010), where participants could record incidents either in writing or as voice messages. To facilitate recall and reflection, participants were asked to provide open-ended responses describing (a) what happened and who was involved or affected, (b) where the incident occurred, and (c) why they considered the behavior to constitute misbehavior. In total, the 19 respondents reported 191 incidents (average: 10.05), ranging from 7 to 14 incidents each.
These incidents formed the basis for the in-depth interviews, which were conducted face-to-face in December 2024, lasting between 40 and 70 minutes (average: 50.5 minutes). 16 hours of recordings were transcribed verbatim, resulting in 356 pages of single-spaced text.
Data Collection
What makes a misbehavior appear more “minor” or “major” to customers is inherently subjective, often difficult for interviewees to articulate, and challenging to interpret. For data collection, we therefore used Kelly’s (1955) RGT as it allows interviewees to articulate personal views of complex experiences, including conscious and unconscious facets, that are otherwise difficult to elicit through direct questioning (Zeithaml et al. 2020). The RGT has been used previously in marketing and management research to elicit subjective perceptions of customers, designers, or managers (e.g., Lemke, Clark, and Wilson 2011; Macdonald, Kleinaltenkamp, and Wilson 2016; Micheli et al. 2012).
All interviews in Studies 1A and 1B followed established RGT procedures (Fromm 2004; Tan and Hunter 2002), with only minor variations. In Study 1A, each interviewee was presented with a set of 12 C2C misbehavior scenarios, whereas in Study 1B, interviews drew on participants’ own C2C misbehavior incidents recorded during the diary study. These scenarios (Study 1A) or real-life incidents (Study 1B) constitute the “elements” in RGT terminology (Goffin, Lemke, and Szwejczewsk 2006) and were written on cards provided by the researchers at the start of each interview. To ensure familiarity with the scenarios in Study 1A, interviewees could exclude any they could not relate to at the start of the interview. However, none chose to do so, indicating that the supplied misbehavior scenarios were relevant to all participants.
Following a variant of Kelly’s (1955[1991]) full context version (see also Fromm 2004; Tan and Hunter 2002; Scheer 1993), interviewees were presented with the full set of cards (i.e., elements: scenarios in Study 1A and incidents in Study 1B) at the start of the interview and asked to rate them based on their perceived magnitude on a six-point Likert scale ranging from “not serious at all” (1; very low magnitude) to “very serious misbehavior” (6; very high magnitude). Corresponding numbers from 1 to 6 were laid out on the table, and interviewees assigned each card to one of these ratings, effectively sorting the cards into up to six distinct piles (see Web Appendix 4 for an example)—each pile representing a different level of perceived magnitude (e.g., all cards rated “1” formed pile 1, those rated “2” formed pile 2, and so on). Notably, participants were free to sort the cards into any number of piles between one and six, resulting in varying numbers of cards per pile across interviewees.
Next, interviewees were asked to compare each pair of adjacent piles in sequence—starting from the lowest available magnitude rating (e.g., comparing pile 1 with pile 2, then pile 2 with pile 3, and so on)—using the question: “Looking at these two piles of misbehavior situations, in what way are the displayed C2C misbehaviors in pile X similar, and at the same time, different from the displayed misbehaviors in pile Y in terms of their magnitude?” This comparative questioning elicited what are termed “personal constructs” in RGT—in this study, attributes of C2C misbehavior situations that differentiate these misbehaviors in terms of their perceived magnitude, as expressed in the interviewee’s own words. To preserve the exploratory nature of our study, we avoided using supplied constructs from prior misbehavior literature, thereby ensuring that all elicited personal constructs remained grounded in the interviewees’ own interpretations of misbehavior magnitude (Fransella 2003). The resulting discussion led to the elicitation of both the “construct pole” (e.g., intentional misbehavior) and the “contrast pole” (e.g., unintentional misbehavior) of each construct, which were recorded on a form called a “repertory grid” (see Web Appendix 5 for an exemplary grid table). This process was repeated until no new personal constructs emerged, before using the same question to elicit further constructs when comparing the next pair of piles.
At the end of the interview, each interviewee rated all cards (i.e., elements: scenarios in Study 1A and incidents in Study 1B) on a six-point scale for each of the personal constructs they had generated. We chose a six-point scale to avoid a neutral midpoint and prompt more decisive evaluations, in line with our aim to clearly differentiate between different levels of perceived C2C misbehavior magnitude. The use of six-point formats is considered effective for capturing nuanced construct-element relationships (Riemann 1991; Scheer 1993) and has been used in prior RGT-based marketing and psychology research (e.g., Elgeti, Danatzis, and Kleinaltenkamp 2020; Woehr et al. 1998). These ratings were recorded in the grid, while interviewees’ reflections during this task provided additional insights into the meaning of each personal construct (Macdonald, Kleinaltenkamp, and Wilson 2016). Given the complexity of the elements and the need to maintain a natural flow in the interview, we conducted the rating task at the end of the interview. Pretest interviews showed that interrupting participants to rate personal constructs immediately after elicitation disrupted their thought process and hindered deeper elaboration, especially as multiple constructs typically emerged from a single comparison. This decision aligns with Scheer’s (1993) recommendations to assess complex elements—such as C2C misbehavior situations—across all personal constructs only after their full elicitation. At the end of the interview, Study 1B participants were additionally provided with formal definitions of misbehavior intensity and impact severity and were asked to rate each incident on a six-point scale, enabling us to capture overall intensity- and severity-related magnitude perceptions.
Data Analysis
One major advantage of using RGT is that it allows for different types of data analysis. The data pool generated from the interviews includes not only qualitative data from the interview transcripts but also quantitative rating data from the grid tables. Data analysis was identical for Study 1A and Study 1B and followed a four-step process (see Table 1 for details). First, personal constructs from the transcripts and grid tables were analyzed and labeled, resulting in 633 first-order personal constructs across both studies. Second, these personal constructs were subsumed into 26 standardized personal constructs, complemented by intercoder reliability checks and Pareto analyses to assess theoretical saturation. Third, the standardized personal constructs were subsumed into 13 second-order themes, six aggregate dimensions, and two higher-order categories that shape perceptions of C2C misbehavior magnitude. Table 2 shows the final data structure (see also Web Appendix 6 for definitions of each magnitude dimension, theme, and standardized construct). In the final step, 13 of the 26 standardized personal constructs were identified as key constructs across both studies (see Table 2 for a breakdown). To check for potential demographic patterns regarding key construct frequencies, we performed subgroup analyses by age and gender within our most diverse sample (Study 1A) but did not observe any systematic differences.
Four-Step Data Analysis Procedures Employed in Study 1A and 1B.
Overview of C2C Misbehavior Magnitude Categories, Dimensions, Themes, and Standardized Constructs Identified in Studies 1A and 1B.
Results
The goal of Study 1 was to empirically derive the facets underlying C2C misbehavior magnitude perceptions. Findings from Studies 1A and 1B reveal 26 standardized constructs that fall into 13 conceptual themes and six magnitude dimensions (see Table 2 for an overview). These dimensions are (1) rule and norm deviation (i.e., the extent to which the misbehavior itself is perceived to violate formal rules or social norms), (2) deliberate advantage-taking (i.e., the extent to which the misbehavior itself is perceived as a conscious attempt by the perpetrator to gain personal advantage), (3) attitudinal influences (i.e., the extent to which judgments of the misbehavior itself are shaped by personal attitudes), (4) disruption of the service experience (i.e., the extent and nature of disruption caused by the misbehavior to the service experience), (5) harm caused (i.e., the extent, type, and timing of harm caused by the misbehavior), and (6) loss of control (i.e., the extent to which the misbehavior causes one to feel unable to manage, influence, or escape its consequences).
We subsumed the first three dimensions into the higher-order category of misbehavior intensity, defined as the perceived strength of the misbehavior itself, regardless of its consequences for others. The last three dimensions, in turn, were subsumed into the higher-order category of impact severity, defined as the perceived impact of the misbehavior on others, independent of the act itself. Notably, this higher-order categorization into intensity- and severity-related magnitude dimensions is theory-informed and emerged through iterative cycling between the data and our prior conceptualizations of intensity and severity, consistent with our abductive research approach (see Table 1; Gioia, Corley, and Hamilton 2013). Together, these facets shape customers’ magnitude perceptions of C2C misbehavior, which we define as the perceived intensity of any misbehavior by a customer toward other customers or shared resources and the perceived severity of its negative impact on those affected.
As shown in Table 2, our findings enrich previous misbehavior literature, which has primarily focused on norm-based or harm-based constructs when assessing the magnitude of C2C misbehavior. While our study affirms the relevance of many of these constructs—and their associated dimensions, rule and norm deviation and harm caused, as key indicators of misbehavior intensity and impact severity, respectively—it also reveals a broader set of magnitude constructs and dimensions that have been previously overlooked. Notably, 14 of the 26 standardized constructs identified in Studies 1A and 1B—seven of which were found to be particularly important for distinguishing between varying degrees of C2C misbehavior—are absent from prior misbehavior research. Eight of these additional constructs form two entirely new dimensions: attitudinal influences (an intensity-related dimension) and loss of control (a severity-related dimension), both of which highlight novel aspects of how customers assess the magnitude of C2C misbehavior across different service settings.
Intensity-Related Magnitude Dimensions
Examining the 13 standardized constructs within the three magnitude dimensions of misbehavior intensity yields the following insights:
Within the first intensity-related dimension, rule and norm deviation, the construct violation of injunctive norms—frequently discussed in prior literature (e.g., Fullerton and Punj 2004; Srivastava et al. 2022) and mentioned by 73% of interviewees—shows only limited ability to differentiate among degrees of C2C misbehavior, given its low ANV value (6.41), below the threshold of 7.74. Often described as the extent to which a behavior violates “unwritten” or “implicit” rules of society (interviewees GS-2, GS-22, GD-6, GD-12), as well as personal beliefs about what constitutes “appropriate,” “respectful,” “moral,” or “decent” behavior in a given service context (GS-4, GS-14, GS-3, GS-30), this construct seems to serve primarily as a baseline, top-of-mind judgment for customers in determining whether a behavior qualifies as misbehavior at all, rather than as an indicator of its perceived magnitude. In contrast, the constructs violations of formal laws and violations of descriptive norms are mentioned often (51% and 47% of interviewees, respectively) with high ANV values (8.69 and 8.06), indicating that they are more important in differentiating among C2C misbehavior magnitude levels. Our findings show that C2C misbehaviors are judged as more major when they are considered a “criminal offense” or a “violation of the law or regulations” (GS-1, GS-3, GD-10). Similarly, the more a behavior is seen as inconsistent with what is expected and viewed as “typical,” “normal,” or “common practice,” the higher its perceived magnitude (GS-16, GS-17).
Within the second intensity-related dimension, deliberate advantage-taking, the well-established construct intentionality (e.g., Berry and Seiders 2008; Colm, Ordanini, and Parasuraman 2017), along with the newly identified construct recklessness, are seen as most important in evaluating different magnitude levels (mentioned by 65% and 51% of interviewees, with high ANV values of 7.84 and 7.78, respectively). Our findings suggest that C2C misbehaviors are perceived as more major when they are viewed as deliberately initiated (e.g., intentionally eating smelly food on trains, GS-10) or as carried out in a way that appears “careless,” “inattentive,” or “indifferent” (GS-18, GD-19) (e.g., not offering an elderly woman on crutches a seat in the underground, GD-16). In contrast, while the construct selfishness is frequently mentioned by 43% of interviewees, it shows only limited ability to distinguish between varying levels of misbehavior (ANV = 7.09).
Finally, within the newly identified third intensity-related dimension, attitudinal influences, our findings suggest that customers’ magnitude perceptions are shaped more strongly by their attitude toward the behavior itself than by their attitude toward the person committing it (e.g., sympathy or antipathy for the perpetrator). Both associated constructs personal dislike and personal understanding were frequently mentioned by 35% and 55% of interviewees, respectively, and were shown to effectively differentiate between varying levels of misbehavior magnitude (ANV = 8.17 and 7.84, respectively). Personal dislike refers to one’s aversion to a specific behavior, such as strong negative reactions to smoking in non-smoking areas (GS-10). In contrast, personal understanding reflects the extent to which customers can empathize with the misbehavior, particularly when they can relate to the underlying motivation (e.g., hiding books in the library due to academic pressure, GS-5).
Severity-Related Magnitude Dimensions
Analyzing the 13 standardized constructs within the three magnitude dimensions of impact severity reveals the following insights:
Within the first severity-related dimension, disruption of the service experience, our findings show that the well-established construct disruption of one’s own service experience (e.g., Fullerton and Punj 2004; Harris and Reynolds 2003) is most important in distinguishing between different levels of C2C misbehavior magnitude. Mentioned by 49% of interviewees with an ANV value of 8.19, it refers to how much a misbehavior is seen to directly disrupt one’s own service experience by disadvantaging, targeting, or interfering with one’s ability to participate in the service as intended (e.g., when someone takes one’s vegetarian meal at a conference (GS-10), steals one’s parking space (GS-7), or physically encroaches on one’s space to the point of preventing work during a train ride (GD-12)). Surprisingly, the other well-established construct in the literature, emotional disruption—describing the extent to which the misbehavior causes emotional strain during the service experience—fails to qualify as a key construct due to its low ANV (7.57), despite being mentioned by 57% of respondents. Instead, our findings reveal two additional key constructs: disruption of other customers’ service experience, referring to how much a misbehavior negatively affects bystanders (e.g., when hotel guests complain loudly, thereby imposing their issue on other guests, GS-13), and functional disruption, which captures how much the misbehavior interferes with the service flow, access to resources, or the achievement of consumption goals (e.g., when someone takes a phone call in a movie theater, hindering one’s goal of watching the movie attentively, GS-3). Both constructs are mentioned by 27% and 59% of interviewees, respectively, with high ANV values of 9.15 and 8.21.
The second severity-related dimension is harm caused by the misbehavior. Unlike disruption of the service experience, which refers to a temporary interference with one’s own or others’ service experience during a specific encounter—often forcing unwanted adjustments in how customers participate in the service—harm caused by the misbehavior captures whether the misbehavior is perceived to inflict actual damage on other customers, potentially lasting and extending well beyond the immediate service encounter. Three key constructs emerged, two of which have been identified in previous literature: physical harm and non-physical harm (e.g., Huang 2008; Jones 1991; Muncy and Vitell 1992). Physical harm (31% of interviewees, ANV = 7.84) captures a “threat to health” (GS-22), “physical violence” (GS-5), or a danger to one’s “physical integrity” (GS-8), such as the impact of smoking on nonsmokers (GS-21). Non-physical harm (47% of interviewees, ANV = 7.81) refers to the extent of financial, moral, or psychological damage caused by the misbehavior (e.g., the time, cost, and emotional effort involved in replacing a stolen phone, GS-17). Notably, our findings reveal a third key construct previously overlooked in the literature: duration of harm (31% of interviewees, ANV = 8.41), which refers to the lasting harm a customer experiences even after the misbehavior has ended (e.g., the lasting impact that a racially insulted person might endure, GS-12). However, despite being top-of-mind for 43% of interviewees, the harm customers experience due to repeated or prolonged exposure to misbehavior is less able to distinguish between different magnitude levels (ANV = 7.18).
Finally, within the third, newly identified severity-related dimension, loss of control, one of the four constructs is of particular importance in judging different levels of C2C misbehavior: unpredictability and risk of escalation caused by the misbehavior. Mentioned by 31% of interviewees and with an ANV value of 8.74, this newly identified construct captures how much a misbehavior creates an unpredictable or aggressive atmosphere that increases the perceived risk of escalation (e.g., nervousness around chanting or hostile soccer fans (GS-29) or a passenger shadowboxing in a train carriage (GD-19)). Of the remaining three constructs, two—perpetrator dominance and need for personal intervention—demonstrate high ability to differentiate among magnitude levels (ANVs of 9.39 and 10.28) despite low mention rates (18% and 6%, respectively). In contrast, the construct helplessness, referring to the sense of helplessness caused when a misbehavior traps someone in a situation they cannot avoid or escape, is mentioned by 65% of interviewees but is less effective in distinguishing between varying magnitude levels of C2C misbehavior (ANV = 6.76).
Study 2
To test the robustness of Study 1’s findings, the purpose of Study 2 was twofold. First, to explore potential cultural differences in assessing the magnitude of C2C misbehavior by using an international sample, as prior research suggests that such judgments may vary by cultural background (Mattila 1999). Second, to triangulate the findings with an alternative elicitation technique by applying the classical triadic RGT “minimal context form” (Tan and Hunter 2002) to further reduce method-specific bias and enhance the trustworthiness of our results.
Research Design and Data Collection
A total of 26 international students enrolled in a research methods module at a British university agreed to take part in a 6-week diary study, which was followed by an in-depth interview. Participants who completed both components received a £50 Amazon voucher as a monetary incentive. In total, 13 participants completed the diary study and the subsequent interview (92.31% female, Mage = 24.9 years; see Web Appendix 8). The diary study procedures closely mirrored Study 1B, except for an extension of about 2 weeks due to the Christmas and New Year holidays. Like Study 1B, respondents reported 10.0 incidents on average, ranging from 9 to 11 incidents each (130 in total). Interviews were conducted face-to-face in January/February 2025, lasting between 27 and 58 minutes (average: 45.35 minutes). A total of 9.8 hours of audio recordings were transcribed verbatim, resulting in 192 pages of single-spaced text. Data collection procedures matched those of Study 1B for the first six interviews, which again used the full context form to elicit constructs, while the final seven interviews used the triadic minimal context form where interviewees were presented with different triads of elements with the question: “Looking at the three C2C misbehavior situations, how are two of them similar and different from the third in terms of their magnitude?” (Micheli et al. 2012). To ensure all elements were covered, triads were presented sequentially by systematically substituting elements (Tan and Hunter 2002). For example, the first triad included elements 1, 2, 3; the second, 4, 5, 6; the third, 7, 8, 9; the fourth, 7, 10, 2; and so on, until all elements had been covered (see Web Appendix 9 for the full set of triads used). Unlike Study 1B, respondents also rated all C2C misbehavior incidents (i.e., elements) against each elicited personal construct immediately after being presented with each triad.
Analysis and Results
Data analysis procedures closely mirrored those of Study 1. Similar to Study 1B, an average of 16.15 constructs was elicited per interview, totaling 210 constructs across all interviews, with theoretical saturation being achieved after nine interviews (see Web Appendix 7D). All 26 standardized constructs from Study 1 also emerged in Study 2, and 11 of the 13 key constructs identified in Study 2 matched those from Study 1, suggesting only minor cultural differences across international and German respondents (see Web Appendix 10 for details). Finally, comparing the two elicitation methods revealed only modest differences in the number of standardized constructs elicited per interview (Mlong form = 9.33; Mtriadic = 12.14). Taken together, these findings support the robustness and cross-cultural applicability of the constructs and dimensions identified in Study 1. While the triadic method showed a slight advantage in the number of constructs elicited, the overall consistency across methods and samples reinforces the validity and generalizability of our findings.
Study 3
The purpose of Study 3 was to inductively explore the dimensionality of C2C misbehavior magnitude perceptions. Specifically, we examined (a) whether the two higher-order categories of misbehavior intensity and impact severity naturally emerge when people assess the six C2C misbehavior magnitude dimensions identified in Studies 1 and 2, thus assessing the categories’ conceptual distinctiveness and potential overlap (higher-order dimensionality); and (b) whether the six magnitude dimensions meaningfully map onto their proposed higher-order categories (first-order dimensionality), thereby assessing potential interpretative flexibility in our proposed allocations. To this end, we employed an open sorting task in which participants freely grouped the six magnitude dimensions into as many groups as they deemed fit without any imposition of a predefined hierarchical structure. Frequently used in marketing and consumer psychology (e.g., Blanchard et al. 2017; Eigenraam et al. 2018; Halkias and Diamantopoulos 2020), open sorting tasks are particularly suited for uncovering how consumers naturally perceive and categorize stimuli based on the implicit associations they hold about how these stimuli “go together” (Blanchard and Banerji 2016).
Method and Materials
A total of 46 native English speakers from the U.S., recruited through Prolific.com, completed an open sorting task. To increase data quality, we used screeners: two respondents sped through the survey, while five failed to create any distinct groups, resulting in a final sample of 39 respondents (41.3% female, Mage = 43.8 years). Following best practices for open sorting tasks (Blanchard and Banerji 2016), participants were presented with six cards (one per C2C misbehavior magnitude dimension), each containing a definition of the respective magnitude dimension, along with two illustrative interview quotes from Study 1A. Participants were asked to read each dimension card carefully and form distinct groups by putting together the cards they felt “belonged together” under a higher-order category. To ensure that participants were naïve to the framework, no mention of the two proposed categories of misbehavior intensity and impact severity was made at any point in the sorting task. Instead, participants were instructed to form as many categories as made sense to them using an interactive online interface (Optimal Workshop) that allowed drag-and-drop sorting of the cards into a self-determined number of category groups. To reduce ambiguity, each card had to be placed into one category, and cards could not be assigned to multiple categories. Participants could rearrange cards between categories throughout the sorting task. After sorting the cards into distinct categories, participants were asked to provide a label for each category that captured its common theme. They were also encouraged to provide an optional comment explaining their sorting strategy in more detail.
Analysis and Results
In total, the 39 respondents formed 114 categories, with the number of categories that emerged ranging from two to five per respondent (M = 2.92, SD = .74). To analyze the data, we employed multiple complementary qualitative and quantitative analyses.
Following Halkias and Diamantopoulos (2020), we first conducted a qualitative analysis of the sorting data by investigating the dimensions used to form categories and the descriptive labels participants assigned to them. Specifically, we qualitatively coded the 114 inductively derived categories according to the two proposed categories of misbehavior intensity and impact severity. Categories containing only intensity- or severity-related dimensions were coded as INTENSITY or SEVERITY, respectively. In turn, categories consisting mainly of intensity- or severity-related dimensions were coded as INTENSITYtainted or SEVERITYtainted, while those with an equal mix of both were coded as MIXED. As shown in Web Appendix 11, out of the 114 derived categories, 81 (71.05%) contained exclusively intensity- or severity-related dimensions. Of the remaining 33 categories, nine were identified as predominantly intensity- or severity-related (7.89%), while only 24 (21.05%) could not be clearly assigned to either of the two. Follow-up analyses of these 24 MIXED categories revealed no systematic alternative grouping pattern. Furthermore, the vast majority of respondents formed at least one intensity-focused (37 out of 39) or severity-focused (36 out of 39) category, showing that 94.9% of respondents intuitively organize C2C misbehavior magnitude perceptions along our two proposed categories of misbehavior intensity and impact severity. The labels participants assigned to the categories further support their conceptual distinctiveness. For instance, categories coded as INTENSITY included labels like aggressive opportunism, nature of the misbehavior itself, or deviation in [the] misbehavior, stressing characteristics intrinsic to the act itself. In contrast, categories coded as SEVERITY were labeled as consequences of misbehavior, fallout from bad behavior, or impact of the misbehavior, highlighting the outcomes or effects of the act. These participant-generated labels closely mirror the conceptual distinctions outlined in our prior definitions of misbehavior intensity and impact severity based on Studies 1 and 2 findings. Together, these results offer qualitative support for our proposed higher-order structure, demonstrating clear conceptual distinctiveness and minimal overlap between the two higher-order categories.
To provide further quantitative evidence of our proposed higher-order structure, we conducted an agglomerative hierarchical cluster analysis to identify natural groupings among the six magnitude dimensions. Hierarchical clustering is often used to analyze sorting data because it considers all co-occurrences among the sorted data simultaneously while yielding a dendrogram that indicates how many clusters best represent the data (Eigenraam et al. 2018; Paea et al. 2021). To this end, we first calculated the co-occurrences among the sorted dimensions across all respondents. The resulting co-occurrence matrix shows how many participants placed a given pair of dimensions together in the same category. For example, the dimension disruption of the service experience has been grouped together with the dimension harm caused by 21 participants (see Web Appendix 12 for details). The co-occurrence matrix was then transformed into a distance matrix and used as input for cluster analysis, employing Ward’s (1963) minimum variance method. As shown in Web Appendix 13, the results reveal a clear two-cluster solution: The first cluster entails the three severity-related magnitude dimensions harm caused, disruption of the service experience, and loss of control, while the second cluster comprises the three intensity-related dimensions rule and norm deviation, attitudinal influences, and deliberate advantage-taking, matching the coding structure of Studies 1 and 2. Robustness checks using average-linkage, centroid, and k-means clustering produced similar results, thus supporting the data-driven emergence of misbehavior intensity and impact severity as the primary higher-order categories of C2C misbehavior magnitude.
Finally, we calculated the coefficient of relative substantive agreement (crsa) to quantify the degree to which the six magnitude dimensions map onto the two higher-order categories of misbehavior intensity and impact severity (Halkias and Diamantopoulos 2020). Ranging from 0 to 1, this measure reflects the proportion of co-occurrences a given dimension shares with other dimensions that theoretically belong together (e.g., the severity-related magnitude dimension harm caused with disruption of the service experience and loss of control), relative to the total number of both correct and incorrect co-occurrences. With crsa values ranging from .67 to .80, all six dimensions well exceed the recommended threshold of .50 (Halkias and Diamantopoulos 2020; see Web Appendix 14 for results), thus offering confidence in the first-order dimensionality of our proposed two-category structure.
Study 4
The purpose of Study 4 was to further examine the dimensionality of C2C misbehavior magnitude perceptions using a deductive approach. Unlike Study 3, Study 4 adopts a theory-driven method to investigate the extent to which respondents allocate a given dimension or construct to its intended higher-order category. To this end, we employed a closed sorting task in which participants were first provided with a definition of misbehavior intensity and impact severity and then asked to allocate each dimension or construct to either of the two categories. This approach allowed us to assess how strongly each dimension or construct is associated with its predetermined higher-order category, thus assessing potential interpretative flexibility at both the construct and dimensional level (see Web Appendix 15 for method details).
To increase confidence in our hierarchical structure, we conducted two closed sorting tasks: one at the dimensional level (where respondents allocated the six magnitude dimensions to either misbehavior intensity or impact severity; Study 4A) and one at the construct level (where respondents allocated the 26 constructs to either category; Study 4B). To quantify the reliability with which the constructs or dimensions were allocated to their intended category, we calculated the proportional reduction in loss (PRL) for both categories. Often used in marketing research for closed sorting tasks (e.g., Macdonald, Kleinaltenkamp, and Wilson 2016), the PRL measure assesses how closely participants’ allocations align with each other and with the intended classification to quantify the reduction in expected loss from classification errors (Rust and Cooil 1994). The obtained PRL statistics for misbehavior intensity (Study 4A: .85, Study 4B: .80) and impact severity (Study 4A: .99, Study 4B: .96) well exceed the recommended threshold of .70, thus providing confidence in the reliability of our proposed allocation of the six magnitude dimensions and 26 constructs to their respective category.
Typology
Based on the empirical insights from Studies 1 to 4, we propose a magnitude-based typology of C2C misbehavior grounded in customers’ perceptions of misbehavior intensity and impact severity. Crossing both magnitude categories yields a 2 × 2 typology, revealing four distinct types of C2C misbehavior, each characterized by a unique configuration of high or low misbehavior intensity and impact severity evaluations (see Figure 1). Unlike many previous customer misbehavior typologies (see Web Appendix 16 for an overview), such a magnitude-based typology is rooted in customers’ evaluations of the incident’s perceived magnitude rather than in the misbehavior’s objective features or the perpetrator’s observable traits.
Types of C2C misbehavior with typical examples.
To empirically validate this typology and identify real-life examples of C2C misbehavior incidents for each type, we conducted two follow-up cluster analyses using the misbehavior intensity and impact severity ratings of the 321 misbehavior incidents collected in diary Study 1B (191 incidents) and Study 2 (130 incidents; see above). Results from a hierarchical cluster analysis indicate a clear four-cluster solution, which was subsequently validated using k-means clustering. The resulting clusters mirror the theoretically expected 2 × 2 structure: a low-intensity misbehavior/low-severity impact cluster (Type 1: 101 incidents), a low-intensity misbehavior/high-severity impact cluster (Type 2: 47 incidents), a high-intensity misbehavior/low-severity impact cluster (Type 3: 37 incidents), and a high-intensity misbehavior/high-severity impact cluster (Type 4: 136 incidents). Although most incidents fall into the diagonal clusters (Types 1 and 4), more than a quarter (26.2%) appear in the off-diagonal (Types 2 and 3), indicating that misbehavior intensity and impact severity are conceptually independent despite their expected positive association—one does not inherently imply the other, consistent with our theoretical rationale. ANOVA results further show that both misbehavior intensity (F = 362.12, p < .001) and impact severity (F = 317.12, p < .001) significantly differentiate cluster membership, demonstrating that each category contributes unique explanatory power (see Web Appendix 17 and 18 for details). Next, we detail each type with exemplary incidents (see Web Appendix 19 for an overview of incidents per type).
Type 1: Minor Misbehavior with Minor Consequences
Type 1 C2C misbehavior is perceived as low in both the intensity of the misbehavior itself and the severity of its impact. Typical examples include everyday inconveniences in supermarkets such as disputes over checkout dividers (GD-2), blocked aisles (GD-7), or shopping cart mix-ups (GD-2), as well as ghosting incidents on online platforms (e.g., GD-9: “Someone ignored me on Vinted when I asked a question about the product”). Other instances involve public displays of affection (GD-17) or minor breaches of etiquette, such as placing feet on seats (ED-8-t) or burping in close proximity (ED-12-t). Overall, such misbehaviors involve minor norm violations that are not perceived as intentional and that rarely upset others as they are noticeable yet easily ignored. Our findings suggest that customers typically feel slightly irritated but ultimately unaffected, often resolving the situation simply by looking away or ignoring the behavior.
Type 2: Minor Misbehavior with Major Consequences
Type 2 C2C misbehavior is not perceived as intense in execution yet is experienced as highly severe in its impact on other customers. Typical examples include territorial behavior in confined spaces such as seat hogging (GD-2) or encroaching on others’ personal space on planes or trains (GD-8, ED-4), disruptive children in otherwise calm and premium environments like high-end restaurants (ED-7-t) or incidents of blocking people’s way in public transport, causing others to miss trains (GD-14, GD-18) or struggle to navigate (ED-13-t). Other examples include noise or odor disturbances, such as loud chewing in restaurants (GD-13), strong body odor (GD-10), or smelly food consumption (GD-8) in public spaces. Overall, though not necessarily seen as deliberate or reckless, our findings suggest that these behaviors can seriously disrupt other customers’ emotional experience or interfere with their consumption goals.
Type 3: Major Misbehavior with Minor Consequences
Type 3 C2C misbehavior is perceived as highly intense in execution but low in negative impact. Typical examples include cutting in line at coat checks, bars, or supermarkets (GD-12, ED-5, ED-7-t), as well as verbal confrontations, ranging from sharp complaints (GD-19) to sudden shouting in public settings like trains (ED-5, ED-11-t). Other examples include the inappropriate use of cell phones in quiet environments, such as phones ringing during movies in the cinema (GD-6) or loud calls in designated quiet zones (GD-19). Overall, such misbehaviors are usually perceived as highly irritating; however, the harm they cause is minor, and disruptions are usually short-lived and can be quickly resolved, making them intense but ultimately inconsequential.
Type 4: Major Misbehavior with Major Consequences
Type 4 C2C misbehavior is perceived as both highly intense in execution and highly severe in its impact on others. Typical examples include using or damaging products in stores (GD-3, GD-5, ED-8-t) and physical aggression or threats of violence, ranging from forceful pushing on public transport (e.g., GD-14: “The man grabbed me with both hands and pushed me aside so he could get on the bus”) to full-scale altercations in nightlife settings (GD-13; GD-14). Other typical examples include verbal and non-verbal harassment such as unwanted advances (GD-1), explicit verbal comments (GD-5), or invasive staring in queues, public transport, or spas (e.g., GD-7: “I was at a textile-free spa and a man stared at me in a very disgusting way”). Overall, such misbehaviors typically involve serious and often deliberate violations of social norms, official rules, or laws and are experienced as threatening, harmful, or deeply disrupting. Our findings suggest that affected customers often feel helpless and face long-term consequences.
Discussion
Theoretical Contributions
Theoretically, this research makes four key contributions. First, our research advances the growing body of research on C2C misbehavior (e.g., Bernritter et al. 2025; Danatzis and Möller-Herm 2023; Schaefers et al. 2016) by being the first to provide an empirically grounded conceptualization of C2C misbehavior rooted in magnitude perceptions. Extant research traditionally focuses on customer misbehavior directed at firms (e.g., Wirtz and Kum 2004) or employees (e.g., Bitner, Booms, and Mohr 1994), while C2C misbehavior has still “received the least attention in the deviancy literature,” as highlighted by Fombelle et al. (2020, p. 392). Importantly, our conceptualization is not limited to a specific industry or country and goes beyond prevailing norm- or harm-based perspectives of extant misbehavior conceptualizations (Fisk et al. 2010). Instead, we uncover a comprehensive portfolio of 26 constructs that fall into 13 themes, six dimensions, and two higher-order categories, offering a nuanced understanding of how customers perceive the magnitude of C2C misbehavior. In so doing, we reconcile previously fragmented conceptual aspects of C2C misbehavior, thus responding to calls to not only shift “the perspective from the provider to the customer’s lifeworld” (Heinonen, Jaakkola, and Neganova 2018, p. 725) when investigating (negative) CCI but also to resolve “tensions between norms- and harm-based definitions of deviance” (Fisk et al. 2010, p. 421) while uncovering “under which circumstances customer misbehavior is perceived differently by other customers” (Benkenstein and Rummelhagen 2020, p. 233).
Second, our proposed conceptualization is the first to clearly distinguish between intensity-related and severity-related magnitude dimensions of C2C misbehavior. Perceptions of misbehavior intensity and impact severity are two distinct magnitude categories that do not necessarily align. Specifically, we find that intensity-related magnitude dimensions include the extent to which the misbehavior (1) deviates from norms and regulations, (2) is seen as a deliberate attempt by the perpetrator to gain personal advantage, and (3) is influenced by personal attitudes. Severity-related magnitude dimensions, in turn, entail the extent to which the misbehavior is perceived to (4) disrupt the service experience, (5) cause harm, and (6) instill a sense of control loss to others. A total of 26 standardized constructs underlie the six dimensions, 13 of which are found to be particularly important for distinguishing between varying degrees of C2C misbehavior. Notably, over 50% of the constructs—and key constructs—identified in our studies are absent from prior misbehavior research. Moreover, several constructs previously discussed in the literature, such as the violation of injunctive norms (e.g., Fullerton and Punj 2004; Srivastava et al. 2022), selfishness (e.g., Benkenstein and Rummelhagen 2020), or prolonged exposure to misbehavior (e.g., Martin and Pranter 1989), appear to function more as baseline, top-of-mind judgments for determining whether a behavior qualifies as misbehavior, rather than as indicators of its magnitude. Our findings thus advance prior misbehavior research by not only delineating the constituent elements that shape perceptions of C2C misbehavior magnitude, but also by determining their relative importance.
Third, our research is the first to investigate the dimensionality of C2C misbehavior magnitude perceptions. Using several sorting tasks involving 99 independent judges from the United States, we provide qualitative and quantitative evidence that our two higher-order categories of misbehavior intensity and impact severity naturally emerge when people assess the six C2C misbehavior magnitude dimensions, and that both the 26 constructs and six identified dimensions meaningfully map onto their proposed higher-order categories. Together, these results provide confidence not only in the conceptual distinctiveness of the identified six magnitude dimensions and the two categories of misbehavior intensity and impact severity but also in their hierarchical structure, offering a robust foundation for future empirical work.
Finally, this paper extends previous typologies of C2C misbehavior (e.g., Baker and Kim 2018; Gursoy, Cai, and Anaya 2017) and customer misbehavior more broadly (e.g., Daunt and Harris 2012; Lovelock 1994) by introducing and empirically validating a magnitude-based typology. Rather than focusing on the misbehavior’s objective features or the perpetrator’s observable traits, our typology is rooted in customers’ subjective evaluations of the perceived intensity of the misbehavior itself and the severity of its perceived impact. Besides delineating four key types of C2C misbehavior, we also identify numerous typical misbehavior examples based on the empirical findings of our RGT studies.
Managerial Implications
Our findings equip firms, communities, and governments to discern different degrees of C2C misbehavior. Rather than identifying C2C misbehavior per se, our conceptualization provides managers with an actionable set of six dimensions (and 26 constructs) to gauge the magnitude of C2C misbehavior incidents. Moving beyond simple frequency measures, our findings enable firms to effectively monitor and categorize C2C misbehavior incidents across sectors and countries, regardless of whether they are dealing with unruly restaurant guests, rowdy theater audiences, or anti-social public behavior (Khalil 2023, Thorp 2023). This magnitude-based typology allows managers not only to classify misbehavior incidents but also to design and prioritize more targeted prevention, intervention, and compensation strategies.
Prevention Strategies
Previous research suggests that C2C misbehavior can be contagious, often starting with one seemingly isolated incident that may spread, triggering other customers to also engage in misbehaviors—some of which may be unrelated and even more intense or severe in their impact than the initial transgression (Danatzis and Möller-Herm 2023). As such, we recommend that managers proactively work to prevent C2C misbehavior across all four types, however low its perceived misbehavior intensity and impact severity.
For instance, service providers can employ strategies that seek to proactively manage social norms in service settings. Signage, in particular, has been found to be well-suited to discourage socially undesirable behaviors in service environments (Esmark and Noble 2016). Whether in the form of printed posters at nightclub entrances communicating zero-tolerance policies toward harassment or in-app reminders about service rules when accessing shared resources such as coworking spaces, bikes, or e-scooters, signage effectively sets social norms and reinforces behavioral expectations (Danatzis, Möller-Herm, and Herm 2024; Danatzis et al. 2025). Research shows that signage is most effective when paired with emotional appeals that signal trust or evoke moral emotions such as shame, guilt, or pride (Fombelle et al. 2020), or use fear appeals or community-oriented framing, especially for C2C misbehavior with low perceived impact severity (i.e., Type 1 and Type 3; Danatzis, Möller-Herm, and Herm 2024).
Besides influencing social norms, service providers can prevent C2C misbehavior by leveraging servicescape features through the design of ambient conditions. Examples include the use of soothing background music, relaxing scents, or calming color schemes in stores, planes, spas, and other enclosed environments, or the use of classical music at train stations or in public spaces, which has been found to deter antisocial behavior, loitering, and vandalism (BBC 2024; Fombelle et al. 2020). In service environments that rely on social cohesion or serve as protected safe spaces for marginalized communities such as nightclubs, consent-based adult venues, private member’s clubs, or peer-to-peer sharing platforms, service providers may also prevent C2C misbehavior through proactive customer selection. Specific measures may include entry screening, membership vetting, or peer rating systems to ensure that admitted customers align with community norms and values (Danatzis et al. 2025).
Intervention Strategies
When prevention fails and C2C misbehavior occurs, we recommend that service providers intervene actively and interrupt the perpetrator’s actions to (a) reduce the perceived intensity of the displayed misbehavior (especially for Type 3 and Type 4 incidents), (b) halt ongoing harm (particularly for Type 2 and Type 4 incidents), and (c) curb the potential for contagion, which applies to all four types. Crucially, this involves the enforcement of service rules. Firms should authorize both frontline employees and on-site managers—often best positioned to assess the situation—to reprimand misbehaving customers in person, issue warnings, or, where needed, involve security staff to remove the perpetrator. Type 1 incidents should warrant only minimal intervention, with a disapproving look or a general reminder about expected norms usually being sufficient. In contrast, both Type 2 and Type 3 incidents require staff to address the behavior directly, while adjusting tone to match the situation. Since Type 2 incidents are often unintentional, polite instructional guidance should usually suffice. More intense Type 3 incidents, by comparison, may require a polite yet more assertive tone (Henkel et al. 2017), while humor could potentially help defuse tension across both types (Beál et al. 2023). Zero-tolerance enforcement through security staff, in turn, appears particularly important in Type 4 incidents where both misbehavior intensity and impact severity are high (e.g., sexual harassment in nightclubs; Danatzis et al. 2025).
If security staff is unavailable, and in-person interventions are deemed unsafe or impractical, firms should instead rely on remote interventions to enforce service rules. For example, on-site managers or frontline employees could deliver personalized announcements remotely using loudspeakers (e.g., on planes or in train stations), electric megaphones (e.g., in parks), or microphones (e.g., during events). Recent research highlights the effectiveness of such remote interventions when enforcing unpopular service rules, such as prohibitions on smoking or alcohol consumption in restricted areas, mask mandates, or requests to lower loud music in public spaces, which often provoke hostile reactions when enforced in person. These interventions have been shown to work best when they clearly identify the misbehaving customer (e.g., via clothing or seat location) (Danatzis and Möller-Herm 2023). To support their staff, firms should establish clear guidelines and provide targeted trainings, including scenario-based exercises and role plays, to equip their employees with the necessary skills and resources to deal with misbehaving customers (Kamran-Disfani et al. 2023).
Compensation Strategies
For C2C misbehavior incidents with high impact severity (i.e., Type 2 and Type 4), service providers should focus not only on the perpetrator but also on the victims of misbehavior. Customers frequently expect some form of reparation to compensate for losses caused by C2C misbehavior (Huang 2008). Service providers should thus engage in service recovery by offering compensations such as refunds, vouchers, discounts, upgrades, free products, apologies, and acknowledgment of the problem to repair the damage caused (Hess, Ganesan, and Klein 2003). Notably, we caution the use of compensation for incidents with low impact severity (Type 1 and Type 3). Besides leading to unnecessary costs, research shows that when firm responses exceed customer expectations, such as providing a refund for misbehavior that is perceived to cause little to no damage, customers may become suspicious of the firm’s motives if such gestures appear disproportionate (Raithel and Hock 2021).
Limitations and Future Research
We acknowledge several limitations that may offer opportunities for future research.
First, this study used a qualitative research approach, namely RGT to elicit the facets that underlie C2C misbehavior magnitude based on 62 scenario-based (Study 1A) or diary-based in-depth interviews (Study 1B and Study 2). A valuable extension would be to quantify the proposed framework, particularly by developing an instrument to operationalize the six identified magnitude dimensions of C2C misbehavior. Being able to measure these magnitude perceptions quantitatively would allow researchers to better understand the interdependencies between them, as well as their relative importance in driving downstream consequences, such as C2C misbehavior contagion, customer satisfaction, and word-of-mouth behavior.
Second, subgroup analyses by gender and age in our most diverse sample (Study 1A) revealed no meaningful differences in key construct frequencies across the six aggregate dimensions or the two higher-order categories (misbehavior intensity and impact severity). To extend the generalizability of our findings, future work could investigate whether the observed lack of demographic patterns holds in other cultural or socio-economic settings.
Third, our proposed conceptualization and associated typology allow firms to more effectively classify and monitor C2C misbehavior instances based on its perceived magnitude. However, further empirical research is required to determine which specific prevention, intervention, and compensation strategies are most effective in addressing different types and degrees of C2C misbehavior. Future research could further examine how effective these strategies are in service failure contexts where C2C misbehavior arises from perceived injustice, as well as in environments that vary in the strength of social norms or the clarity of rules and regulations. Experimental research offers a particularly promising avenue forward.
Finally, future research could develop automated approaches to detect and categorize C2C misbehavior instantly. Respective machine learning and natural language processing models could analyze large volumes of real-time data of C2C interactions, from video surveillance to written or verbal complaints, to facilitate the seamless implementation of countermeasures.
Supplemental Material
sj-docx-1-jsr-10.1177_10946705261439674 – Supplemental material for Customer-to-Customer Misbehavior Magnitude: Dimensions and Typology
Supplemental material, sj-docx-1-jsr-10.1177_10946705261439674 for Customer-to-Customer Misbehavior Magnitude: Dimensions and Typology by Annelie Neetzow, Ilias Danatzis and Jana Möller-Herm in Journal of Service Research
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
