Understanding passenger hostility in transit: a systematic review

Passenger hostility is commonplace in the transit industry. However, research investigating this job stressor is limited in both scope and quality. This systematic review examines three questions regarding passenger hostility: (a) what is its occurrence (incidence and prevalence); (b) what are its consequences; and (c) what factors are related to its occurrence, intensity, or consequences? Six databases were searched for studies published between 2010 and 2020 reporting the occurrence, consequences, or correlates of passenger hostility directed toward transit workers. Risk of bias was categorized using the Critical Appraisal Skills Programme. Results were summarized in tables and narratively synthesized. Nineteen eligible studies (10 quantitative and nine qualitative or mixed methods) were identified. Evidence from 14 studies indicated that passenger hostility occurs frequently, and is often underreported, especially in its less severe forms. Sixteen studies reported correlates of the occurrence, intensity, or consequences of hostility, with these grouped into worker, passenger, organizational, environmental, and event categories. Immediate and long-term impacts were identified in ten studies. This review contributes to understanding the likely antecedents, correlates and consequences of passenger hostility directed at transit industry workers. It also highlights some common methodological limitations in past research and proposes solutions to them. As such, the paper provides suggestions for future explorations and intervention targets aimed at reducing this common workplace stressor.

serious physical violence (i.e., behaviors intended to physically harm).
Transit workers experience many job stressors (Greiner & Krause, 2006).These include those relating to the physical environment (e.g., poor ergonomics and traffic congestion) and those relating to the job (e.g., time pressure, rotating shift patterns, reduced rest breaks, and social isolation), as well as organizational stressors (e.g., reduced decision-making authority; Tse et al., 2006).In terms of contemporary theories of job stress and burnout, such as Demerouti et al. (2001) job demands-resources model, transit workers typically have few resources (e.g., autonomy, information, support) with which to manage the demands of their job.This combination of high demands and limited resources may have significant consequences, with a review (Tse et al., 2006) highlighting the prevalence of physical health repercussions among transit workers including cardiovascular disease, gastrointestinal problems, musculoskeletal disorders, and fatigue, as well as psychosocial consequences such as post-traumatic stress disorder, alcohol, tobacco, and drug use.Beyond these, there are numerous organizational consequences of transit worker stress such as absenteeism, employee turnover, and traffic accidents (Tse et al., 2006).
Passenger hostility represents a further hazard that may add to these generic stressors and their consequences.Alarmingly, transit workers have shown to be up to four times more likely to face hostility at work relative to the typical worker ("Violence at Work", 2003as cited in Villotti et al., 2020), and they experience higher rates of verbal abuse than other vulnerable occupations, including law enforcement personnel (Geoffrion et al., 2017).Reasons for this may include the susceptibility of the service they provide to uncontrollable circumstances and events (e.g., traffic congestion, mechanical failures) and the likelihood that passengers outnumber transit employees and are emboldened by this numerical advantage.Some evidence suggests this problem is increasing, with a need for further intervention to support transit workers facing rising rates of passenger hostility (Deloitte, 2017).
In other service settings, hostility is linked with poor worker well-being (e.g., emotional dissonance, stress, and cognitive impairment) and job outcomes (e.g., poor service quality and high turnover; Sommovigo et al., 2019).Given the potentially severe consequences of experiencing passenger hostility for transit workers, and its seemingly increasing incidence, there is a need to understand passenger hostility more fully.This knowledge can be used to minimize the incidence and intensity of passenger hostility as well as its impact on the health and well-being of workers and their associates.Although preliminary efforts have been made to understand passenger hostility, factors contributing to its occurrence, intensity, or consequences remain underexplored.This is problematic given that identifying malleable factors that exacerbate or attenuate hostility constitutes an important first step in developing effective, well-targeted interventions.The current paper reports a systematic review of research investigating passenger hostility toward transit workers.The main objectives of this research are to provide a detailed and generalizable understanding of likely antecedents, correlates, and consequences of passenger hostility so as to inform the development of interventions that reduce its occurrence, intensity, and consequences.Three review questions were examined: 1. What is the occurrence (incidence; rate of exposure and prevalence; proportion of population with prior exposure) of passenger hostility directed toward transit workers? 2. What are the consequences of passenger hostility for transit workers? 3. What factors are related to the occurrence, intensity, and/ or consequences of this hostility?
In order to address these research questions, a systematic review of empirical literature was conducted by searching the extant literature about transit hostility and transit workers.This was performed following the PRISMA guidelines.

Registration, protocol, and eligibility criteria
The review protocol was registered on Prospero (CRD42021253368).A full protocol is available from the authors upon request.A modified PICO strategy (Richardson et al., 1995) was used to determine inclusion criteria.Quantitative, qualitative, and mixed method empirical studies were eligible for inclusion, whereas non-empirical papers, reviews and systematic reviews, conference abstracts/papers, and book chapters were excluded.

Participants, outcomes, and exposure
Studies were included when the sample or a sub-sample constituted workers (e.g., driver, conductor) employed in the transit industry (e.g., bus, rail, tram, taxi, air, ferry).For outcome inclusion, the study must have reported data pertaining to: (a) the occurrence of passenger hostility; (b) the relationship between passenger hostility and a purported outcome variable; and/or (c) correlates of the occurrence, intensity, and consequences of passenger hostility.Exposure inclusion criteria specified that the study must have reported on passenger hostility of some form (e.g., incivility, abuse, aggression, or physical violence).Due to the dominance of studies conceptualizing passenger hostility as a subset of workplace hostility in general (i.e., that perpetrated by coworkers, workers, managers, or other parties), these studies were also eligible for inclusion, granted that passengers were also identified as a source of hostility.This is important because evidence suggests that, although acts of aggression may emanate from multiple sources, passengers greatly outweigh other parties as the perpetrators of workplace hostility in the transit industry (e.g., 87% of incidents; Assunção & Medeiros, 2015).Studies that did not separate passenger hostility directed at other passengers from that directed at transit workers, but did include the latter, were also eligible for inclusion.Studies were excluded if the passenger hostility was not directed toward transit workers.

Information sources, search strategy, and study selection
Six electronic databases (EBSCO, ProQuest, PubMed, Psy-cInfo [Ovid], Scopus, and Web of Science) were searched in late 2020.Three broad categories of search terms were used: (a) perpetrator (e.g., passenger or customer), (b) misbehavior (e.g., incivility, abuse, aggression), and (c) transit industry (e.g., bus, train).Search terms were combined with the 'OR' Boolean operator, with terms from the three categories combined with an 'AND' operator.To maximize the relevance of results: (a) the search strategy was applied, where possible, to title, abstract, and keywords, and (b) the keywords "chemistry", "biology" and "health care" were excluded.These exclusion criteria were based on preliminary searches identifying studies using our key terms in ways unrelated to human passengers in transit (e.g., studies relating to transporters of a biological kind such as peptides and lipids).Search terms were modified for each database to accommodate differences in search operators and rules.Details are available upon request.
Additionally, three supplementary searches were conducted using the search engine Google Scholar, with simplified search terms (i.e., "passenger aggression" AND buses; "workplace aggression" AND "bus driver"; and [aggression OR abuse OR incivility] AND "bus driver").All searches were restricted to full text studies, published in peer-reviewed English language journals, and date restricted to publication within the preceding decade (January 1, 2010, to July 15, 2020).This date restriction was applied given significant changes to the transport industry in the last decade, including a greater focus on reducing anti-social behavior broadly (e.g., Moore, 2011).Reference lists of eligible studies were screened for additional publications meeting the inclusion criteria.After compiling the search results, duplicates were removed prior to independent screening by two reviewers.These reviewers sequentially examined the article titles, abstracts, and full texts to determine eligibility, prior to data extraction.Disagreements between the two reviewers were resolved using a third reviewer.The PRISMA flow diagram (Page et al., 2021) in Fig. 1 provides further detail.

Data collection process and items
The full text of eligible studies was used for data extraction.Data extracted included the following details: source (author, publication year); sample (size, age, sex, country); study design; and the primary aim or focus of the study.The following data were further extracted for specific review questions.For review question one: the passenger hostility construct measured; the manner in which the construct is measured; the period over which data were collected; and the substantive results.For review questions two and three, the following were additionally extracted (where relevant): the aspect of passenger hostility measured (occurrence, intensity, or consequences); the follow-up study period; and the type of analysis performed (including variables controlled for).The accuracy of this extracted information was confirmed by the research team examining randomly selecting and checking four of the included studies.

Risk of bias and synthesis methods
The included studies were evaluated for risk of bias and methodological quality using the Critical Appraisal Skills Programme's cohort and qualitative checklists (CASP, 2020) to assess the (a) validity, (b) reliability, and (c) generalizability of results.Two independent reviewers assessed risk of bias, with disagreements resolved by a third reviewer and consensus from the broader research team.As reducing the risk of bias to a single numerical indicator both within and between studies can omit valuable information (CASP, 2020), the full methodological components assessed for risk of bias are presented in Table 1.Components that could not be appraised due to reporting or design limitations were marked as not satisfactory.Studies were rated for quality based on the number of areas of potential risk.Due to the heterogeneity within the included studies, a narrative approach was utilized to synthesize the findings from all included studies in text, with presentation of data, where appropriate, in tables.The narrative synthesis for each finding was framed around the number of studies and their risk of bias, representing the level of certainty in the body of evidence.

Study selection, characteristics, and bias
As seen in Fig. 1, search and screening processes resulted in 10 quantitative, and nine qualitative and mixed method studies for inclusion.All quantitative studies meeting inclusion criteria used questionnaires.A cross-sectional design was used in six of these studies and a longitudinal design was used in two, the remaining two were descriptive in design.
Of the qualitative and mixed method studies, six used interviews, one used observational methods, and two used a combination of approaches including observations and questionnaires.Content analysis and critical incident technique were each used in three qualitative studies, whilst constant comparative method analysis, hermeneutic analysis, and grounded theory methods were used in one study each.One additional study presented results descriptively.Studies were conducted in a range of jurisdictions including: six in North America, four in Europe, three in Africa, two in Oceania, one in South America, one in Asia, and one globally.Samples included transit workers, transport operators, and archival records (e.g., police records) of violent incidents.Samples sizes ranged from 19 to 1,774 in the quantitative studies, and from 16 to 100 in the qualitative or mixed-method studies.The most frequently studied occupation was that of bus drivers.Detailed characteristics are presented in Table 1.
Overall, the risk of bias was most commonly low (i.e., high quality) in the quantitative studies, and most commonly moderate or high (i.e., low to medium quality) in the qualitative/mixed-method studies.For quantitative studies, the area with greatest risk of bias was inaccurate measurement of outcomes.For qualitative and mixed-method studies, the area with greatest risk of bias was inadequate consideration of the relationship between researcher and participants.The risk of bias of individual studies is an important indicator to consider when determining the strength of evidence obtained in the literature.The risks of bias for all studies are presented in Table 2.

Results of individual studies
Results from individual studies are reported, according to the review questions, in Tables 3 (review question one), 4 (review question two), and 5 (review question three).These tables present brief result overviews, with detailed results available in supplementary Tables S6, S7, and S8, respectively.

Review question 1: the occurrence of passenger hostility directed toward transit workers
Of the 19 identified studies, ten quantitative and five qualitative or mixed methods reported the occurrence of passenger hostility toward transit workers.Incivility (i.e., rude, or unsociable behavior) was not reported in any study, abuse (i.e., personally offensive verbal behavior) was indicated in four, aggression (threatening or intimidating hostile behavior) in eight, and physical violence in 10 studies.Five studies were judged to be of low quality, four were medium quality, and six were high quality.

Review question 2: the consequences of passenger hostility for transit workers
Nine studies reported detrimental correlates of passenger hostility for transit workers.Four were quantitative and five were qualitative or mixed methods.One of the studies was low quality, four were medium quality, and four were high quality.

Longer term consequences
An array of self-reported psychological correlates were identified across five studies (Bance et al., 2014;Byun et al., 2016;Couto & Lawoko, 2011;Geoffrion et al., 2017;Zhou et al., 2018).After accounting for demographic and occupational factors, both suicidal ideation (Byun et al., 2016) and burnout (Couto & Lawoko, 2011) were more likely reported by workers who had experienced hostility than by workers who had not.Flashbacks, nightmares, avoidance, loss of interest in pleasurable or important activities, sleeping problems, hypervigilance, and concentration problems were all endorsed as consequences of workplace hostility (Geoffrion et al., 2017).Moreover, acute stress disorder, post-traumatic stress symptoms (Zhou et al., 2018), psychological reactions (e.g., sleep problems), and depression were also identified as relatively common (Bance et al., 2014).Zhou et al. (2018) found that, after controlling for gender and post-traumatic stress symptoms, confidence in dealing with passenger aggression was negatively associated with symptoms of acute stress, 12 months after experiencing workplace hostility.Likewise, Gilbert (2011) reported that taxi drivers were deeply affected by the fatal shooting of another driver a year prior, and felt vulnerable after hearing of physical altercations experienced by colleagues.Regarding general health, transit workers reported increased driverpassenger conflict as contributing to a perceived lack of control, and high job demands leading to stress and negative health consequences (Bowles et al., 2017).

Review question 3: factors related to occurrence, intensity, or consequences of passenger hostility on transit workers
Sixteen studies (eight quantitative and eight qualitative or mixed methods) reported factors related to the occurrence, intensity, or consequences of passenger hostility on transit workers.This set of 16 studies comprised six low quality, six medium quality, and four high quality studies.These factors are presented below grouped into six categories: worker, passenger, organizational, social support, environmental, and event factors.

Worker factors
Several worker factors were identified as relevant to hostility occurrence (i.e., gender, age, exercise, tenure, immigration status, and specific driver actions), intensity (i.e., appearance and tenure), and consequences (i.e., gender and coping).Mixed findings were obtained for the relationship between employee gender and occurrence of hostility.A low quality study identified that female staff more frequently experience hostility (Salomonson & Fellesson, 2014), whilst high (Friis et al., 2020) and medium quality studies (Assunção & Medeiros, 2015) reported no relationship, after controlling for inspector interpersonal actions and passenger gender.Evidence also linked gender to the consequences of hostility, indicating male bus drivers with a recent experience of workplace hostility were more likely to experience moderate post-traumatic stress disorder problems at one-month follow-up, relative to female bus drivers (Zhou et al., 2018).In contrast, females reported less confidence in coping with aggression at 12-month follow-up after controlling for acute stress and post-traumatic stress disorders.
Evidence was identified for an association between worker age and hostility occurrence such that, after controlling for health and occupational factors, older individuals less often experience workplace hostility (Assunção & Medeiros, 2015).Other factors including worker demographics (e.g., country of birth, physical appearance) were reported to be a focus of abuse, whilst specific features (e.g., body size and tattoos) were identified as discouraging hostility (Salomonson & Fellesson, 2014).Weekly worker exercise was also linked to reduced hostility occurrence (Assunção & Medeiros, 2015).Furthermore, self-rated illhealth, previously diagnosed chronic diseases, and sickness absenteeism were associated with experience of workplace hostility, with the latter two effects holding after accounting for age, health, occupational, and environmental factors (Assunção & Medeiros, 2015).
Evidence was inconsistent regarding links between job tenure and hostility occurrence, with a medium quality study reporting a positive relationship among bus drivers and fare collectors (Assunção & Medeiros, 2015), and a low quality study descriptively reporting the same among flight attendants (Krasnoshchechenko et al., 2019).However, an inverse tenure-hostility relationship was reported for train conductors and bus drivers in a low quality study (Salomonson & Fellesson, 2014).Descriptive results suggested greater tenure among flight attendants was related to more frequent conflict resolution with passengers (Krasnoshchechenko et al., 2019).However, contradictorily, those with greater tenure scored lower in the cognitive, behavioral, and motivational components of conflict resolution (Krasnoshchechenko et al., 2019).Evidence also showed that workers were encouraged to seek treatment because of the adverse impacts of hostility such as coping with trauma-related symptoms; changes in personality, emotions and relationships; and the financial burden resulting from time off work (Bance et al., 2014).Whilst various worker factors were identified, findings indicate that the contribution of diverse worker factors (e.g., risk preference, confidence, job knowledge) to the likelihood of abuse, physical assault, or false allegations may differ between native and foreign-born drivers (Schwer et al., 2010).Specific worker behaviors during interactions were associated with passenger hostility.After controlling for other interpersonal actions (e.g., authority, dominance, accommodation) and gender, accommodating and authoritative or dominating behavior from the worker were identified as discouraging or encouraging hostility, respectively (Friis et al., 2020).Other worker behaviors including confrontation (Gilbert, 2011), passenger ejection (Salomonson & Fellesson, 2014) and rigid policy enforcement (Fellesson et al., 2013), as well as driver attitudes (Lincoln & Gregory, 2015) were also identified as eliciting passenger hostility, whilst interpersonal skills (i.e., verbal skills, balanced approach, and flexibility) were reported to resolve hostility (Salomonson & Fellesson, 2014).

Passenger factors
Three studies identified the role of passenger characteristics, including boredom and showing off, the presence of groups (Lincoln & Gregory, 2015), intoxication (Salomonson & Fellesson, 2014), passengers disrespecting drivers, unmet expectations, and crime (e.g., robbery; Couto et al., 2011), as possible causes of hostility.Regarding passenger gender, contradictory evidence emerged.Whilst one high- (Friis et al., 2020) and one low quality study (Salomonson & Fellesson, 2014) reported no link between gender and hostility (with the former controlling for inspector gender and interpersonal actions), analysis of CCTV footage and incident records in an otherwise low quality study indicated young males were most often the perpetrators of hostile incidents (Lincoln & Gregory, 2015).

Organizational factors
Evidence emerged linking organizational practices, policies, and norms with the occurrence of hostility and its consequences.More specifically, fare pricing and fare retrieval were identified as common precipitants of hostility within four studies (Couto et al., 2011;Gilbert, 2011;Lincoln & Gregory, 2015;Salomonson & Fellesson, 2014).Limitations in service delivery were explored in two studies, suggesting restricted service resources (Fellesson et al., 2013), unrealistic timetabling, and automated systems lead to passenger frustration (Lincoln & Gregory, 2015).Regarding organizational responses, two studies implicated conflict avoidance, greater system control (i.e., fare inspectors; Couto et al., 2011), police liaison, and security features (e.g., radio/duress alarms, and onsite security; Bonfanti & Wagenknecht, 2010) as efficient methods of reducing hostility.The importance of education for drivers and the community to reduce the  Gregory and Lincoln (2018) Cases of physical violence between drivers and passengers unfolds in time and space • After a violent interaction, instances of bus drivers treating subsequent passengers with irritability • Often resumed work immediately, where the psychological consequences of the incident likely affected their driving quality and customer service Lincoln and Gregory (2015) Violence and aggression against bus drivers Bus stakeholders described an overall sense of pessimism regarding violence and aggression, suggesting it will never be solved occurrence of hostility was also identified in two studies (Couto et al., 2011;Lincoln & Gregory, 2015).Several factors indicative of high job demands were associated with experiencing workplace violence during a 12-month period.For example, working overtime, without breaks, and on different buses were each associated with greater experience of workplace violence (Assunção & Medeiros, 2015).More broadly, bureaucratic organizational policies appear to worsen the consequences of passenger hostility, with transit workers suggesting the rigidity and number of work-related procedures (i.e., paperwork, low guidance, and financial demands) following an incident were significant stressors (Bance et al., 2014).One study descriptively reported that over a quarter of bus drivers perceive discussion of non-physical workplace aggression to be useless, whilst one in ten participants in this study viewed passenger aggression as a part of their job (Geoffrion et al., 2017).These sentiments were also found amongst bus drivers, although they often denied or minimized their own encounters with hostile passengers (Lincoln & Gregory, 2015).As such, participants in the study by Lincoln and Gregory (2015) indicated that they report fewer than half of all incidents, while Uzosike and Douglas (2017) suggested that rates of reporting may be as low as 29% of incidents of workplace psychological violence (i.e., verbal abuse, bullying, or tribal harassment).

Social support
Social support was negatively associated with several consequences of hostility.Specifically, two studies noted social support was associated with lower levels of burnout (Couto & Lawoko, 2011), while counter-supportive behaviors were associated with greater post-traumatic stress symptoms (Zhou et al., 2018).Regarding the source of support, evidence was mixed, with a high quality study indicating burnout occurrence did not vary with source of support (e.g., supervisor, co-worker, external; Couto & Lawoko, 2011), whilst transit workers in a medium quality study indicated peers to be a fundamental source (Bance et al., 2014).Perceived managerial support was positively associated with the likelihood of support-seeking following an event (Bance et al., 2014).However, Geoffrion et al. (2017) reported that fear of judgement from employers for complaining about workplace hostility is common (i.e., 50%), as are negative judgements from colleagues (i.e.26%).

Environmental factors
Environmental factors, both inside and outside of the transit vehicle, were associated with hostility occurrence and intensity in some studies.The rate of hostility was linked to within-vehicle factors including overcrowding in two studies Taxi drivers' experience of abuse or threat in the last 12-months, more likely if foreign-born vs native-born.p = 0.01 Taxi drivers' experience of physical assault in the last 12-months, more likely if: ○ Foreign-born vs native-born.p = 0.01 ○ Preference for risks.p = 0.05 for native-born.Non-significant for foreign-born ○ Greater knowledge.p = 0.05 for native-born.Non-significant for foreign-born ○ Greater confidence.p = 0.05 for native-born.p = 0.10 for foreign-born Taxi drivers' experience of robbery in the last 12-months, more likely if avoidance perception of risks.p = 0.01 for native-born, untested for foreign-born Taxi drivers' experience of fare evasion in the last 12-months, more likely if greater knowledge.p = 0.10 for native-born.Non-significant for foreign-born Taxi drivers' experience of false allegations in the last 12-months, more likely if: ○ Foreign-born vs native-born.p = 0.01 ○ Avoidance perception of risks.p = 0.01 for native-born.p = 0.10 for foreign-born Qualitative or Descriptive Findings Bonfanti and Wagenknecht (2010)   Transit worker interviews revealed three dimensions of the service system related to customer misbehavior: ○ Service system resources are limited (e.g., in space/capacity or offered services) ○ Service regulations impose an ideal service that is not realistic, increasing the likelihood of conflict ○ Service practices mediate the relationship between regulations and reality and thus incur misbehavior Friis et al. (2020) Observation of a passenger performing an aggressive behavior toward a ticket inspector: ○ All passengers displaying aggression were 16 years or older Gilbert (2011) Taxi driver interviews regarding violence against taxi drivers indicated: ○ Passengers transported at night commonly under the influence of alcohol or substances.Drivers become party to illegal activity or the recipient of sexual advances or witness sexual behavior (especially distressing for foreign-born drivers) ○ Violence typically involves money issues including fare disputes, fare evasion, etc Krasnoshchechenko et al. (2019) Flight attendants with < five years of experience: 66% experienced passenger-related conflict at work often, 30% sometimes, and 4% never.Those with five or more years of experience: 36% experienced passengerrelated conflict at work often, 54% sometimes, and 0% never Lincoln and Gregory (2015) Focus groups on violence and anti-social behavior against bus drivers: ○ Drivers suggested vulnerability characterizes the role because of factors such as working alone, being unprotected etc Survey results explaining why passengers and others abuse bus drivers: ○ Late running ○ Fare costs ○ People's attitudes CCTV analysis of attacks against bus drivers indicated: ○ Incidents commonly experienced during afternoons and nights later in the week (Thursday through Saturday) ○ Passenger frustration propelled 23% of incidents ○ Disputes over fare propelled 24% of incidents ○ Profit motive propelled 21% of incidents ○ Boredom or showing off propelled 25% of incidents ○ Abusive incidents commonly perpetrated by groups of young males Bus company incident reports indicated: ○ 13% of incidents occurred between 3 and 4 pm, constituting the most common one-hour window ○ 16% of incidents occurred on Thursdays, Fridays, and Saturdays, constituting the key days ○ Common months for offending were October and November ○ The number of reported aggressive incidents decreased by 73% from August-October 2010 to August-October 2013 Stakeholder interviews on violence and anti-social behavior against bus drivers indicated: ○ Preventative measures are largely reactive and unimplemented ○ Incidents are declining in frequency but increasing in severity ○ Common precipitators are fare evasion and ticketing issues (including the automated ticketing system), and unrealistic timetabling ○ Young people perceived as main perpetrators Observational analysis of violence and aggression against bus drivers recorded on one 24/7 bus route indicated 17% of perpetrators were drunk or drugged Police data indicated 191 (0.19%; in 2011), 231 (0.23%; in 2012), and 149 (0.15%; in 2013) offences against property per 100,000 Commercial bus drivers and conductors, truck drivers, and taxi drivers experience of physical violence in the last 12-months: ○ 44% of physical violence occurred between 7am and 1 pm, 47% between 1 and 6 pm, and 9% between 6 pm and 12am ○ 10% of physical violence occurred at a motor park (i.e., transit hub), 49% at the passenger's destination, and 41% while driving ○ 29% reported (35% to park chairperson and 65% to police) Intensity Quantitative Findings Krasnoshchechenko et al. (2019) Among flight attendants with < five years of experience, relative to those with ≥ five years: ○ Motivational component of conflict resolution skills was higher.p = 0.001 ○ Cognitive component of conflict resolution skills was higher.p = 0.030 ○ Certain behavioral components of conflict resolution skills were higher: extraversion (non-significant), agreeableness (p = 0.030), conscientiousness (non-significant), openness to experience (non-significant), and emotional stability (p = 0.038)Among flight attendants: ○ Work experience was negatively associated (r = -0.37)with motivational component of conflict resolution skills ○ Conscientiousness was positively associated (r = 0.34) with motivational component of conflict resolution skills ○ Extraversion was positively associated (r = 0.36) with regulatory component of conflict resolution skills ○ Emotional stability was negatively associated (r = -0.59)with the cognitive component of conflict resolution skills ○ Openness to experience was positively associated with the regulatory (r = 0.29) and cognitive components (r = 0.30) of conflict resolution skills Qualitative or Descriptive Findings Gilbert (2011) Taxi drivers described interpersonal violence that commonly occurred (e.g., conflict emerging from efforts to deal with interpersonal violence or rule violations, like smoking) Gregory and Lincoln (2018) Analysis of observational (CCTV) data of physical violence between bus drivers and passengers indicated: ○ Most incidents lacked volatility and involved a sustained and escalating interaction (e.g., verbal aggression, accumulated issues) prior ○ Involvement of other passengers may discourage or encourage the severity of the outcome Krasnoshchechenko et al. (2019) Flight attendants with < five years of experience: 34% reported resolving passenger-related conflict to full extent, 44% were not sure if it was resolved, and 22% did not resolve the situation.Those with five or more years of experience: 64% reported resolving to full extent, 32% were not sure if it was resolved, and 4% did not resolve the situation Lincoln and Gregory (2015) Focus group participants described violent incidents against bus drivers as escalating quickly from 'nothing' Analysis of CCTV data of attacks against bus drivers indicated they are characterized by escalation and volatility Stakeholder interviews indicated violent and anti-social incidents against bus drivers are declining in frequency but increasing in severity Salomonson and Fellesson (2014) Interviews with train conductors and bus drivers on customer misbehavior indicated: ○ Helpful interaction strategies include verbal skills, being considerate, avoiding acting authoritatively, etc ○ Strategies based on appearance can be helpful including use of body posture, aesthetic appearance, etc. Can also be disadvantageous ○ Helpful strategies making use of the physical environment can include waiting until stopped to eject customers.Physical environment can also be a limitation (Couto et al., 2011;Lincoln & Gregory, 2015), and with increased whole body vibrations, temperature, poor lighting, and poor technical resources in another study (Assunção & Medeiros, 2015).Evidence implicated aspects of the transit service itself as facilitating (e.g., difficulty avoiding perpetrators) or obviating (e.g., timing passenger ejections with Interviews with transit workers who had recently experienced a traumatic event at work indicated: ○ 59% identified peers as a fundamental source of support.A peer support system for those with similar experiences was recommended ○ 66% emphasized the importance of managerial response after the event.33% described negative interactions with management as a barrier to seeking support and its own stressor ○ Supportive management is associated with favorable views of treatment seeking ○ 61% mentioned paperwork, and 50% the rigidity and scope of work-related procedures following an event, as daunting ○ Participants reported worrying about the impact of the event on their family.This prompted them to seek treatment ○ 24% noted an inability to function normally and cope with trauma symptoms.They reported treatment was their only option ○ 31% reported seeing a change in themselves and their relationships motivated them to seek treatment.In addition to post-traumatic stress disorder, some were being treated for depression ○ Several took time off work after the event, with this financial burden encouraging treatment seeking Geoffrion et al. (2017) Frequency of endorsement amongst bus drivers: ○ 26% perceived they would be judged by colleagues for complaining about severe workplace violence ○ 50% perceived they would be judged by employer for complaining about severe workplace violence ○ 28% perceived talking about workplace aggression without a visible wound as useless ○ 12% perceived severe violence in their workplace or role as normal ○ 61% were unaware of their workplace's aggression policies Lincoln and Gregory (2015) Focus groups discussing violence and anti-social behavior against bus drivers indicated: ○ Incidents are viewed as inevitable and implicit to the role and thus should be accepted ○ The potential for aggression is acknowledged as ever-present, but personal experiences are denied or minimized ○ Many participants recounted positive experiences within their role, with a sense of positivity and pride Survey of bus drivers indicated that while over half reported some incidents, they did not report all because they perceived nothing would be done (48%), dealt with it themselves (44%), was not serious enough (41%), and there would be no consequences (37%) Stakeholder interviews on violence and anti-social behavior against bus drivers indicated verbal violence is more common, but no less damaging, than physical violence Uzosike and Douglas (2017) Commercial bus drivers and conductors, truck drivers, and taxi drivers' experience of physical violence in the last 12-months: 29% reported (35% to park chairperson and 65% to police) For brevity non-significant findings are not reported here but are available in Table S8.PR Prevalence ratio; OR Odds ratio; BF Bayes Factor stops) hostility, according to train conductors and bus drivers (Salomonson & Fellesson, 2014).External environmental factors including late-running transit and traffic congestion (Assunção & Medeiros, 2015;Lincoln & Gregory, 2015), time of day (Bonfanti & Wagenknecht, 2010;Lincoln & Gregory, 2015), and location on bus route (Uzosike & Douglas, 2017) were also linked with the occurrence of hostility.

Event factors
Aspects of the hostility event were linked with its intensity and consequences.For example, three studies (Gregory & Lincoln, 2018;Lincoln & Gregory, 2015;Salomonson & Fellesson, 2014) linked passenger-worker conflict with escalation, volatility, and accumulation of unresolved conflict.Whilst verbal aggression was observed to commonly precede bus passenger physical violence (Gregory & Lincoln, 2018), bus driver focus groups suggested violent incidents are spontaneous (Lincoln & Gregory, 2015).
Other aspects of hostile events, such as their frequency, recency, and target, were also associated with adverse outcomes.For example, repeated exposure to violence was related to post-traumatic stress symptoms at one-month, six-month, and twelve-month follow-up after controlling for demographics, social support, and prior exposure (Zhou et al., 2018).However, these symptoms generally reduced over time up to 12-months post-incident.Another study indicated that attacks on co-workers vicariously led to a feeling of vulnerability amongst taxi drivers (Gilbert, 2011).

Discussion
Given the potential implications of an additional stressor on transit workers in an already stressful occupation, identifying the occurrence, consequences, and correlates of passenger hostility is a vital first step in addressing this phenomenon.The current review was the first to systematically examine the phenomenon of passenger-to-transit worker hostility with the goal of developing a comprehensive and generalizable understanding of its multiple facets.Strong evidence emerged regarding both the occurrence and consequences of passenger hostility, whilst mixed evidence surrounded many factors related to its intensity.

Occurrence of passenger hostility direct toward transit workers
Strong evidence was found for the occurrence of three types of passenger hostility (i.e., abuse, aggression, and physical violence).Overall, passenger hostility was shown to be commonly experienced by transit workers, across a diverse range of sociocultural contexts, with some minor variations between employee roles.This variability may reflect workers' differing exposure to passengers: for instance, train drivers, who are traditionally enclosed within a cabin, report fewer experiences than bus drivers, who are more visible and accessible to passengers, including those who are intoxicated, showing off, and/or in groups.It may also reflect other factors such as the reliability of the service, the stressfulness of the environment, the number of passengers, and their characteristics.The available evidence suggests that the frequency of passenger hostility decreases as severity increases.This widespread and tiered occurrence of hostility reflects established trends in other industries (Yagil, 2008), including administration, finance, and health (Ben-Zur & Yagil, 2005).
The findings related to hostility occurrence contrast with those pertaining to rates of incident reporting.That is, abuse makes up approximately one quarter of reported incidents while aggression and physical violence around half, suggesting lower-level hostility is underreported.This could be due to a lack of an accessible and convenient reporting mechanism for low level hostility or a perception that these levels are not worth reporting (Arnetz et al., 2015), despite industry sentiment suggesting all hostility is harmful (Lincoln & Gregory, 2015).Indeed, disregarding one excluded study that operationalized passenger incivility as an act of fare evasion that was not directed at the driver (Lincoln & Gregory, 2015), the occurrence of incivility was not reported by any identified study, despite its prevalence and acknowledged consequences within other service industries (Yagil, 2008(Yagil, , 2021)).
The evidence suggests hostility is more common in certain interactions, such as those involving authoritative (e.g., placing blame) or dominant (e.g., physically confining space) employee actions.Specific service characteristics appear to be a source of frustration for passengers and potentially trigger hostility, as predicted by established aggression theories such as the frustration-aggression hypothesis (Berkowitz, 1969).

Consequences of passenger hostility
Whilst preliminary evidence was found as to the detrimental correlates of passenger hostility, a scarcity of wellcontrolled, longitudinal explorations hinders the ability to draw valid conclusions about their status as consequences.Nonetheless, the research suggests that passenger hostility is associated with some severe psychological symptoms for transit workers including suicidal ideation, burnout, and traumatic symptoms.It seems likely, although the evidence is weaker, that these emotional and physiological symptoms affect workers' abilities to fulfill their role, reduce their confidence in coping with future aggressive passengers, and may lead to prolonged absenteeism and other long-term outcomes.For example, acute stress and repeated hostility exposure were associated with posttraumatic stress symptoms up to 12-months post-incident.Such detrimental associations are not unexpected, given evidence of similar cognitive, emotional, physiological, and job-related consequences in retail workers (Harris & Daunt, 2013).Taken together, the findings suggests hostility is likely a significant source of stress, ill-health, and under-performance for transit workers alongside the plethora of already established stressors relating to their job environment, role, and organizational context (Tse et al., 2006).

Correlates of passenger hostility
Findings suggest aspects of the transit environment (e.g., overcrowding, excessive heat, or noise) may prime hostile behaviors.Supporting this, research in other service industries shows that customer hostility is often the result of accumulated dissatisfaction (Yagil, 2008).Observational data obtained in transit settings similarly indicates that hostile incidents commonly escalate as unresolved issues between the parties accumulate over time.Seemingly in contrast, other evidence suggests that drivers often view incidents as spontaneous.
At an organizational level, given job demands were associated with the experience of hostility, and bureaucracy with worse consequences, it appears the incidence and consequences of passenger hostility are exacerbated by other such stressors (e.g., working overtime and paperwork).This complements reports suggesting hostility contributes to a lack of job control, high job demands, and poorer stress and health outcomes (Bowles et al., 2017), consistent with the job demands-resources model (Demerouti et al., 2001).Clearly, more research is needed to confirm the unique and interactive impacts of hostility, by examining it concurrently with other stressors within this model.
Social support has been shown in past research to buffer the effects of high job demands (Gonzalez-Mulé et al., 2021).The current review provided similar evidence regarding supportive behaviors in the transit industry.Strong evidence linked social support to better outcomes following hostility, whilst counter-supportive behaviors from management seemingly hindered support-seeking.

Strengths and limitations
This review represents the first known systematic effort to examine the occurrence, intensity, and consequences of passenger hostility toward transit workers.Utilizing PRISMA guidelines, established critical appraisal tools, and wide inclusion criteria ensured broad coverage of the empirical literature.The examination of this issue across diverse sociocultural boundaries and transit services helps establish this phenomenon as universal.Such efforts constitute a guide in this underserved area, providing suggestions for future studies and interventions to explore this issue in unique sociocultural contexts and within underexplored transit services (e.g., airlines).
Important limitations in the literature should be noted.It is evident that passenger hostility has been conceptualized and measured in varied ways.Such variability weakens the ability to make valid comparisons between studies.Adding to this problem is the mixed quality of analysis and reporting: statistical controls vary greatly between studies, and critical definitional and methodological details are lacking from several reports.Other notable methodological limitations include the prevalence of self-report measures taken at various arbitrary times following an incident of hostility, a reliance on descriptive statistics, the use of poorly operationalized (e.g., Krasnoshchechenko et al., 2019) or unvalidated (e.g., Schwer et al., 2010) self-report measures, and the small number of multi-wave studies exploring predictors and consequences (e.g., Zhou et al., 2018).In addition, most of the research available focusses on bus and taxi services, highlighting a need to expand this body of research to incorporate other types of transport (e.g., trains, ferries, and planes).
Our review also has limitations.For example, it is restricted to studies published in English language outlets over a 10.5-year period, and, while it examines likely antecedents of passenger hostility (e.g., poor service quality), it does not explore possible causes of these antecedent factors (e.g., under-staffing, prior passenger misbehavior).Although the study of such antecedents of factors triggering hostility was beyond the scope of the current study, this may represent an area for future research and intervention.Identification of studies is also limited to those retrieved using the current search engines and search terms, which may have inadvertently omitted additional articles.Further, some findings may have been overlooked during the extraction process.

Research, managerial, and policy implications
To begin addressing this prevalent and complex issue, efforts should be made to improve the quality of research designs and reporting.Multi-wave studies are needed, as are objective and third-party measures of critical variables.There is also a clear need for standardized conceptualizations and operationalizations of passenger hostility variables, including standard durations over which occurrence is measured (e.g., past-month experience of low-level hostility, and past-year experience of more severe levels of hostility).Between-study comparisons will be more valid if such conventions are adopted.We propose that going forward, the umbrella term passenger hostility be adopted to refer to the spectrum of hostile behaviors.Under this rubric, specific acts of hostility should be clearly and consistently operationalized per severity as either incivility, abuse, aggression, or physical violence.Whilst these types of hostility often occur either alongside each other or in an escalating sequence, each may have distinctive causes, consequences, and correlates, and would therefore benefit from being treated as such in future explorations.
This review has identified several gaps in the extant research.In the future, efforts should be made to capture hostility across all levels, with special attention paid to low level acts of incivility.Also unexplored, is the full range of possible consequences of passenger hostility, with most past research focusing on employee mental health outcomes to the exclusion of broader consequences affecting organizations, other passengers, and the victim's family.Related to this, is the vicarious impact of passenger hostility on co-workers, with one study (i.e., Gilbert, 2011) suggesting such trauma may be common and severe, even for non-fatal altercations.
Findings from this review suggest several management strategies that can be implemented to address passenger hostility.Primary prevention by organizations may start with a well-publicized and consistently enforced zerotolerance policy toward hostility.As highlighted in this review, specific aspects of service delivery, such as inflexible and inconsistent procedures, may trigger passenger frustration and lead to acts of hostility.As such, there is likely to be opportunities to minimize the incidence of passenger hostility or at least carefully de-escalate it before its severity peaks.By re-designing workplaces and schedules, many of the triggers can be avoided, and by training staff in problem resolution skills, the duration and intensity of hostile acts may be minimized.Some strategies may be best implemented in a targeted manner.For example, given that employee gender, and possibly immigration status, are risk factors for passenger hostility, intervention efforts can be implemented with such groups in mind.
Rates of passenger hostility are high, but due to likely under-reporting, their true prevalence is unknown.Transit managers need to consider deterrents in the existing systems to the reporting of hostile acts, and devise ways to reduce the burden and stigma often associated with reporting.The findings pertaining to the high prevalence of passenger hostility also suggest that the problem may be difficult to eradicate entirely.As such, management should give consideration to ways in which to protect and support their staff.Managers should ensure that effective security mechanisms are in place and should encourage workers to take advantage of opportunities for vertical and horizontal workplace social support.
Interventions at the organizational levels are likely to have partial, rather than full, success in mitigating the problem of passenger hostility.The efficacy of these organizational strategies can be strengthened by broader policies implemented at industry and governmental levels.These may take the form of mandated safety hardware and security systems installed in all vehicles, and/or mandated staff training in de-escalation skills and stress management.Closer surveillance, stricter rule enforcement, and tougher penalties may deter many would-be hostile passengers.

Conclusion
This review has taken initial steps to identify potential correlates of the occurrence, intensity, and consequences of passenger hostility.Several of these (e.g., environmental triggers, interpersonal action, social support) may be malleable for use when developing or implementing interventions, and several of the findings may be generalizable to other service sectors.Given the methodological limitations of the research included in this review, further exploration of these correlates and consequences would be a valuable next step.Doing so will make meaningful strides in developing effective interventions to reduce passenger hostility and thereby protect the health and well-being of transit workers.

Fig. 1
Fig. 1 PRISMA Flowchart for the Identification, Screening, and Inclusion of Eligible Studies

Table 1 (
Percentages have been rounded to whole numbers.-=detailnot reported in study a Missing data within this sample resulted in the number of observations differing from variable to variableTable 2Quality appraisal information using CASP Tools (CASP, 2020) for all included studies ✓ = criterion was met; ✗ = criterion was unmet, or the information was unavailable; -= not applicable as this was a cross-sectional study

Table 3
Summary review question one findings: occurrence of passenger hostility

Table 3 (
a = Construct was not clearly defined.b = Measure was not clearly defined

Table 5
Uzosike and Douglas (2017)2014)monson and Fellesson (2014)Interviews with train conductors and bus drivers about customer misbehavior indicated: ○ Female and junior staff more frequently affected.Misbehavior directed towards females commonly involved comments relating to appearance, age, and gender ○ Verbal abuse often linked with appearance, age, origin, and gender ○ Misbehavior often arises when ejecting passengers from train ○ Abuse or threats often arises from ticket checks ○ Alcohol and other drugs are often involvedUzosike and Douglas (2017)