Radicalization and attenuation in the construction of tourism employment images: a multi-level frame analysis

Purpose – Tourism employment is very diverse ranging from precarious, exploitative study to high-quality workplaces. However, poor employment images dominate the tourism industry, which makes attracting employees difficult. This study aims to examine the processes that lead to such image construction. Design/methodology/approach – Using a qualitative methodology, the study develops a multilevel framing cycle comprising a media analysis of newspapers and magazines (macro-level), a conversation analysis of peer communication/negotiations (meso-level) and a content analysis of single employee/ manager interviews (micro-level); and a comparative analysis of the macro-, mesoand micro-level findings. Findings – The multilevel frame cycle identifies image-construction processes that pass through working conditions, payment, seasonality and human resource problems. These processes are shaped by the two cross-level dynamics of radicalization and attenuation. The latter consists of rationalized and repressed framings of tourism employment images (TEI) and the former consists of ideological and emotional framings. © Andreas Mölk, Manfred Auer and Mike Peters. Published by Emerald Publishing Limited. This is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and noncommercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode The research presented in this article is part of the research project EBKOM, which is funded by the Tyrolian State Government’s Translational Research Programme. However, the funder did not play a role in the study design; collecting, analysing, and interpreting the data; writing the report; and the decision to submit the article for publication. They thank all members of the EBKOM research group for their helpful support in the development of this paper. The University of Innsbruck funded the open access publication of this article. Multilevel frame analysis Received 23 December 2020 Revised 27April 2021 15 July 2021 23 September 2021 Accepted 26 September 2021 International Journal of Contemporary Hospitality Management EmeraldPublishingLimited 0959-6119 DOI 10.1108/IJCHM-12-2020-1490 The current issue and full text archive of this journal is available on Emerald Insight at: https://www.emerald.com/insight/0959-6119.htm


Introduction
Research on employment images is important for the tourism and hospitality industries, which often struggle with labor shortages (Fes, 2018) and problems attracting a qualified workforce (Baum, 2019b;Daniel et al., 2017). Poor employment work images are a major cause of those difficulties, as tourism employment is often associated with unacceptable and demanding conditions (Baum, 2018) and precarious employment structures (Robinson et al., 2019). However, tourism employment is diverse in terms of both workplace quality and work experiences (Baum, 2019a(Baum, , 2018. Companies and entire regions have been striving to change poor employer and work images and highlight the positive aspects of tourism employment. These efforts range from employer branding (Gehrels, 2019) to sustainable human resource (HR) strategies (Baum, 2018). Despite these efforts, negative images of tourism employment persist.
The literature on tourism employment images (TEI) does not fully explain this situation because it typically measures (potential) employee perceptions of aspects of tourism employment (McGinley et al., 2018;Schneider and Treisch, 2019;Wan et al., 2014). Two interlinked considerations are underdeveloped in this research as follows: (1) The multilevel nature of TEI, including the role of nonemployee stakeholders (e.g. managers), social encounters (e.g. peer conversations) and institutional influences (e.g. media organizations).
Focusing on the interplay between the individual, interactional and institutional levels can help explain why the negative facets of tourism employment prevail. Research needs to go beyond analyzing the effects of issues, such as payment levels, career opportunities, work schedules and (potential) employee evaluations of job attributes by focusing on the processes and dynamics in the construction of TEI across multiple levels. Therefore, we ask as follows: Q1. How are TEI constructed within and across individual, interactional and institutional levels?
To answer this question, we adopt Cornelissen and Werner's (2014) framework, which sensitizes framing within and across the macro-(institutional), meso-(interactional) and micro-(individual) levels. We then investigate the framing processes and dynamics in the construction of TEI, using media products, peer conversations and interviews to empirically examine the joint construction of TEI in Tyrol (Austria). Tyrol presents an excellent case for our purposes given the economic and sociocultural importance of tourism there, the problematic nature of TEI, the many initiatives taken to improve workplace quality and the image of tourism work in the region. We first discuss institutional, interactive and individual image construction processes as they relate to four overarching themes. Then, we discuss the cross-level dynamics that shape these processes, which are pivotal for the construction of TEI within and across different levels. This study introduces a holistic and dynamic understanding of TEI as the first multilevel frame analysis performed within this context.
2. Literature review 2.1 Tourism employment: the ambiguity of working practices Employment in tourism and hospitality is diverse and it is difficult to generalize about working conditions, job attributes and quality (Baum, 2019a). Much of the literature emphasizes the negative facets of tourism employment, identifying poor and demanding working conditions (Armstrong, 2017), low pay (Baum et al., 2020), limited career opportunities (Goh and Okumus, 2020), stress and burnout related to emotional labor (Lee and Madera, 2019) and precarious employment structures (Baum, 2019a;Robinson et al., 2019). These studies highlight seasonality, contingent and nonstandard work forms and the utilization of cheaper worker cohorts, such as youth, women and migrant workers. Large parts of the workforce in tourism and hospitality experience job insecurity, exploitation and high levels of vulnerability (Baum et al., 2020;Robinson et al., 2019). However, the literature also describes the positive facets of tourism employment (Baum, 2019a;McPhail et al., 2015;Mooney et al., 2016), characterizing it as satisfying and fulfilling (Goh and Baum, 2021). It also explains how "employers of choice" (Baum et al., 2020(Baum et al., , 2018 attempt to overcome the problems of tourism work and provide a set of progressive HR management policies.

Tourism employment images: management of workforce perceptions
The TEI literature reflects the ambiguity of work and employment in tourism. It is largely based on the paradigm of effective HR management and thus attempts to identify the factors influencing (potential) employee perceptions of working in the tourism and hospitality industries (Kusluvan and Kusluvan, 2000;Kusluvan et al., 2010). Many studies have measured the impact of general management, HR or marketing practices, such as sustainable business practices (Baum, 2018;Murray and Ayoun, 2010), relationship marketing (Cameron et al., 2009) and career opportunities (Richardson, 2009) and satisfaction (Chang and Busser, 2020) on potential and/or existing employees' perceptions of tourism employers. One stream of research has dealt with worker perceptions of the nature and social status of their work, including low pay and seasonality in tourism and hospitality (Wan et al., 2014) and the role of organizational status on TEI and employee attraction (McGinley et al., 2018). Scholars have elaborated on the problematic aspects of tourism employment, such as work-life conflicts, workplace-induced isolation (McGinley et al., 2017) and bad management behavior (Hight et al., 2019). Another stream of research has examined the role of student and/or employee characteristics in TEI perceptions, including their generational behavioral peculiarities (Goh and Lee, 2018;Richardson, 2010;Solnet and Hood, 2008;Tsaur and Yen, 2018;Zopiatis et al., 2012), the importance they place on their careers (Brown et al., 2014), their attitudes at work (Kusluvan, 2003;Zopiatis, 2007), the evaluative repertoires used to assess job attributes (Schneider and Treisch, 2019) and their cultural backgrounds and social networks (Janta et al., 2012).

Considerations in tourism employment images research
Despite the sophisticated ways in which scholars have studied TEI, two considerations remain underdeveloped. First, these studies have mainly focused on the individual level, usually investigating (potential) employees' perceptions of specific features of tourism employment. They have only minimally dealt with the interpretations of other individuals (e.g. managers, customers) and institutional stakeholders (e.g. media organizations and trade unions). Moreover, with the exception of Schneider and Treisch (2019), studies have not addressed social, cultural or institutional factors, including interactions within groups of employees or managers. Therefore, we conducted a multilevel analysis to overcome this onesided perspective, focusing on the impact of individual, interactional and institutional levels. Second, aside from Goh and Lee (2018) and Hight et al. (2019), scholars have followed a functionalist and/or positivist epistemological orientation focused on measuring the impact of the factors influencing perceptions of employment images. They have largely neglected the processual and dynamic characteristics of TEI. Thus, we use a frame analysis approach common in general management (Ansari et al., 2013; and tourism (Leung et al., 2019) research, which allows us to learn more about the emergence and consolidation of TEI.

Levels and linkages in tourism employment images construction: a multilevel frame analysis
We follow Cornelissen and Werner's (2014) agenda for a multilevel frame analysis to examine TEI construction. This agenda shifts the focus from a unilateral understanding toward "an enlarged conceptualization of framing" (215), which goes beyond a concern with single actors' involvement and the impacts of specific negative or positive facets of tourism employment and work.
This agenda organizes frame constructs into three levels. First, the micro-level outlines the sensemaking processes of single actors. In our research, the micro-level emphasizes tourism stakeholders' work experience, knowledge and attitudes. Second, the meso-level maps out meaning construction in groups. For our purposes, the meso-level sheds light on how work is processed among tourism peers. The macro-level describes the activation of "broader cultural templates of understanding" (Cornelissen and Werner, 2014, p. 182). In our study, the macro-level draws attention to widely accepted and recognized assumptions about working cultures or industry logic in tourism. Cornelissen and Werner (2014) build three linkages between these levels to enable a multilevel frame analysis. First, the macro-micro linkage implies that the meaning constructions of single actors are "acts of framing in context" (215). Our study outlines how single stakeholders engage with working cultures and industry logic, such as the individual framing of job dissatisfaction in conjunction with precarious employment structures. Second, the micro-meso linkage indicates that frames are constructed via interaction and reveal participants' social identities. This linkage addresses, for example, the interactional framing of leadership behavior in tourism companies in relation to leader or follower identities. Third, the meso-macro linkage assumes that institutional frame constructs are built in a two-step manner as follows: first by establishing "personal common grounds" (217) during small-scale interactions between pairs of actors; and second through the emergence of "communal common grounds" (217) that spread across interactions, actors and time. This linkage outlines how institutionalized working cultures and industry logics, such as seasonality, are based on their framing within interactions.
2.5 Framing cycles in tourism employment images construction: extending a multilevel frame analysis This study extends the suggestion of Cornelissen and Werner (2014) to link levels to pairs. We connect the macro-, meso-and micro-levels toward a framing cycle ( Figure 1) to bring "ideas together in one location" (Entman, 1993, p. 51). We argue that frame analysis can also be theorized as a circulating system in which all three levels are linked. We aim to link the institutional, interactional and individual framing of TEI. A framing cycle enables the analysis to elicit the cross-level dynamics and thrusts of meaning construction in tourism employment. By applying a framing cycle, we examine how the meanings of negative or positive facets of tourism employment and work spread across institutional, interactional and individual framing and establish employment images of tourism.
3. Methods and research design 3.1 Contextual background of tourism employment images The Tyrolean tourism industry offers an exceptional opportunity to empirically investigate the construction of TEI. Tourism plays a major role in regional economies. According to the Tyrol Chamber of Labour (2020), the tourism industry employs 11% of the region's workforce. Moreover, the ambiguous character of tourism work and employment is especially pronounced in Tyrol. Critical employment reports highlight the negative facets of tourism employment and work (Biehl and Kaske, 2011). Likewise, many Tyrolean businesses enjoy employer-of-choice status among tourism students/interns. Moreover, due to a severe shortage of skilled and qualified labor (Austrian Hotelier Association, 2021), hotels and tourism associations have launched initiatives to improve the region's TEI (e.g. Job-Life-Achensee).

Research design
We used a qualitative research design for our multilevel frame analysis given the quantitative focus of most existing studies ( Figure 2). An employment image is understood as an abstraction with socially, culturally and experientially constructed meanings. The qualitative research provides insights into the construction of employment images beyond the restrictions of statistical data. Importantly, existing frame analyses call for more rigorous qualitative research designs (Cornelissen and Werner, 2014) to capture the processes and dynamics of the framing. This study had two main objectives. First, it established a holistic understanding of TEI by performing three levels of analysis as follows: a macro-level to examine the institutional frames/framing of media organizations, a meso-level to examine the interactional frames/ framing of peer groups and a micro-level to examine the individual frames/framing of single employees/managers. These levels were established based on the TEI literature. Second, the study sought a dynamic understanding of TEI and linked the three levels to analyze the relationships between the institutional, interactional and individual frames/framing, following the proposal of a multilevel frame analysis in Cornelissen and Werner (2014) and our own notion of "framing cycle." We investigated the macro-level through media organizations' reports on tourism work over time, the meso-level through peer communication/negotiations about work and the micro-level through single employees/managers work knowledge, experience and attitudes. Finally, the macro-meso-micro linkages were explored by examining the (overarching) themes in which media organizations, peer groups and employees/managers engaged while constructing the TEI.
3.3 Data collection 3.3.1 Macro-level. The first data set included media reporting over a period of approximately nine years (January 1, 2011, to December 31, 2019). The key inclusion criteria were as follows: The media coverage had to be important to the Tyrolean tourism industry; the target readership had to comprise potential or existing tourism employees/managers; and the coverage had to disseminate employment-relevant content. We included articles from the Tiroler Tageszeitung, the region's daily newspaper with the largest circulation in Tyrol (249 articles); Hotel and Touristik, a monthly hotel management magazine (56 articles); Tourist Austria International, published biweekly (135 articles); and Saison, a quarterly magazine for Tyrolean tourism managers (39 articles).
3.3.2 Meso-level. The second data set comprised 31 unstructured conversations conducted in 2015 during the so-called "Long Night of Research" with participants who were invited to discuss working in the Tyrolean tourism industry. The key inclusion criteria were that the participants required peer status and had to engage in conversations about tourism employment and peer-to-peer exchanges of employment knowledge and/or experiences. We observed 19 conversations among employees (ce1-19) and 12 conversations among managers or owners (cm1-12). The conversations varied in length (two to 16 min) and group sizes (two to five peers). The conversations were audio recorded and transcribed.
3.3.3 Micro-level. The third data set included 36 semi-structured interviews with 18 employees (ie1-18) and 18 managers (im1-18). The key inclusion criteria were that the interviewees had to work in one of the traditional sub-sectors of the Tyrolean tourism industry (e.g. hotel, restaurant) for at least three years or be in an active tourist employment relationship during the study period and had to describe themselves as either tourism employees or managers. Interviews were conducted between 2011 and 2019. The employees were between 21 and 37 years old and comprised both front-office (e.g. waitresses) and backoffice (e.g. chefs) staff. The managers were between 34 and 60 years old, including 13 top executives [e.g. chief executive officers (CEOs)] and five hotel owners. The choice of interviewees was random and depended on their willingness to participate. The interviews followed a storytelling approach (Czarniawska, 2004), with participants initially asked to "narrate" their work experience in the Tyrolean tourism industry. We also asked about career motives and opinions regarding past/current employers. The interviews took place in corporate offices or classrooms, averaged 44 min long and were audio recorded and transcribed.

Data analysis
A coding structure was established to process the media products, peer conversations and interviews at the beginning of our data analysis. We used MA codes for macro-level frames/ framing, ME codes for meso-level frames/framing and MI codes for micro-level frames/ framing. We built on these to create categories (e.g. ME1), sub-categories (e.g. ME1.1) and category systems for each of the three levels of analysis. To ensure validity, we used researcher triangulation (Mayring, 2014) in which all authors coded the three sets of data independently of one another. Different interpretations were discussed after the individual coding rounds and were incorporated into the analysis. To ensure effectiveness, we employed ATLAS.ti for interviews and conversations and Citavi for media products. Our data analysis [1] comprises two synchronized sequences ( Figure 3). 3.4.1 First sequence. This sequence aimed at framing the media (macro-level), peer groups (meso-level) and individual employees/managers (micro-level). Here, we analyzed the three sets of data in isolation, exploring the construction of TEI "within" the levels (summarized in Table 1).
3.4.1.1 Media analysis. We examined media products at the macro-level following the principles of media analysis Altheide and Schneider (2013). First, we defined seven thematic units based on the TEI literature as follows: payment/compensation, work quality/ conditions, career advancement, labor shortages, the characteristics of potential employees, sustainable business practices and organizational/social status. Building on these, we derived keywords for each thematic unit. For instance, "working conditions" included the keywords "stress" and "working time." The media products were subsequently analyzed using a computer-assisted keyword search (APA-Defacto-Campus). The results for each theme were initially organized according to the year and place of publication. This resulted in 232 articles (an average of 58 per year), most coming from Tiroler Tageszeitung (121). Second, we processed the search results of each thematic unit to form categories (e.g. "MA4.2"). As this task produced 56 categories, we increased the level of abstraction via summarization and created main categories ("MA1" -"MA6"). Third, we evaluated the temporal aspects of our categories by studying how media coverage developed over time. We then created a final category system for the macro-level.
3.4.1.2 Conversation analysis. We used conversation analysis Sacks et al. (1974) as a framework to examine the peer conversations. First, we divided the conversation transcripts into episodes, focusing on those that provided information on turn-taking practices; forms (e.g. "struggling with the response"); and the frequency with which speaker turns were taken, maintained and yielded. We discovered no interactional TEI construction but observed an individual (manager/employee) construction in 13 peer conversations. Second, we separately analyzed the employee and manager episodes, inductively creating categories for the recurrent turn-taking practices of these two peer groups. We found that the turn-taking practices in the manager conversations were often comprising "refusing" or "highlighting" but were comprising "embracing" in the employee conversations. The turn-taking practices of "refining," "stressing," and "marginalizing" were evident in both peer groups. Third, we introduced a thematic focus and organized the categories according to TEI topics. This task resulted in six main categories ("ME1" -"ME6"). However, as these main categories were often very extensive, we structured them through the extraction (e.g. "ME1.2," "ME1.3"). This process resulted in a final category system for the meso-level.
3.4.1.3 Qualitative content analysis. At the micro-level, we examined all interview transcripts following qualitative content analysis principles (Mayring, 2014). We initially examined 10 transcripts line-by-line, five employee interviews (E) and five manager interviews (M). We formed categories (e.g. "MI 1.5") when there was a fit between interview excerpts and  "within" levels the study's seven literature-based thematic units ("Media Analysis" section above). After stabilizing the initial category system (e.g. when no new categories were found), we conducted a revision loop in which the categories and coding were intercoder-evaluated. Points of revision were then discussed and if necessary, considered in the respective category systems. Second, we grouped categories (e.g. "excessive working hours" and "challenging schedules" were transferred to "MI1.4"), thereby creating a category system that embraced empirical insights from both the manager and employee interviews. Next, we analyzed the remaining 26 transcripts in accordance with this merged category system, verifying whether the revised categories were relevant and consistent with the rest of the interviews and adapting them where necessary. This resulted in 17 main categories and 45 subcategories. We then grouped similar categories again and built main categories ("MI2" -"MI8") to establish the micro-level category system.   IJCHM 3.4.2 Second sequence. In the second sequence, we determined how the framing of media, peer groups and individual employees/managers was reciprocally influenced, exploring the construction of TEI "across" different levels (summarized in Table 2).
3.4.2.1 Qualitative comparative analysis. First, we compared and positioned the individual, interactional and institutional framings vis-à-vis each other and searched for overarching themes. We examined all the category systems and identified the topics that played a role at the macro-, meso-and micro-levels. We based the TEI construction on the following four overarching themes (Figure 4): Table 2.  Second, we elaborated on the cross-level dynamics shaping the construction processes of these four themes. We began with the macro-level because we regarded media framing as an institutionalized, widely established discourse. We then explored how the meso-and microlevels within each theme were related to institutional framing. This procedure enabled us to determine that the "working conditions" and "payment" construction processes radicalized while "seasonality" and "HR problems" attenuated. Finally, we wrote a researcher's narrative consisting of topic-specific summaries. The data analysis was continuously checked for reliability issues. In addition to the mutual checks performed by the authors, each analysis sequence was subjected to an inter-coder check (Mayring, 2014) that evaluated category systems and coding processes. The first sequence was evaluated by six masters' students (two at each level) and the second sequence was evaluated by a research fellow with knowledge of multilevel frame analysis. Points of concern (e.g. the demarcation of "working conditions") were discussed and where appropriated, incorporated in the respective data analysis areas. The argumentation used was often emotionally charged due to the particularly critical discourse and had the potential to trigger an emotional escalation in readers. The latter can be seen in the following example:

Micro-level.
Negative press releases were highlighted in the employee interviews, with respondents mostly condemning poor working conditions (MI1). Employees reported a poor working atmosphere and environment (MI1.1), (considerable) rivalry among employees (MI1.2), poor communication practices (MI1.3) and unfavorable working hours and schedules (MI1.4). Poor tourism working conditions often upset the employees being interviewed, with interviewees reporting their own, often negative, work experiences (MI1.5). By contrast, managers distanced themselves from apparent and potential criticism about these issues (MI2). These stakeholders mostly criticized media reports on working conditions as being false or largely misleading (MI2.1). The managers also refuted potential criticism by asserting the unchangeable nature of tourism work (MI2.2; Table 3): 4.1.4 Meso-level. Interestingly, the managers largely ignored the apparent criticism in the peer discussions, instead highlighting the beneficial aspects of tourism employment (ME2). They mentioned work safety (ME2.1), the employability of professionals (ME2.2), the range of tasks (ME2.3) and the team element of tourism jobs (ME2.4). Discussing these beneficial aspects often triggered a positive emotional state (e.g. enthusiasm). The following excerpt offers an example: cm10: from (your perspective) speaking (.) what motivates employees the most. cm8: (0.5) difficult. cm9: I think it has something to do with the team (.) that there is a great team. cm8: I AGREE (.) but the host (.) The boss is also important. cm10: <so> entire crew (.) <ALL> cm9: that's (.) all about.
From a multilevel framework perspective, the media laid the foundation for an emotional discourse on working conditions in tourism. The employers and employees both intensified In another article, a representative of the largest association of hotel owners offered a generally positive view of developments concerning tourism salaries: The average gross wages mostly increased in restaurant services, about 47 to 51 euros per employee. "This is more than a whole week's wages", said Thomas Reisenzahn, Secretary of the ÖHV (Austrian Hotel Association). (Tourism Austria International, October 7, 2011, p. 10) The ideological distortions between employees and employers/managers about tourism wages became particularly obvious during negotiations on collective agreements. In this context, union representatives spoke of "modern slavery" (Tiroler Tageszeitung, August 17, 2012, p. 18), whereas employer representatives criticized union representatives for trying to "use the media politically" (Tiroler Tageszeitung, August 17, 2012, p. 18). 4.2.2 Meso-level. The peer discussions strongly confirmed the political controversies regarding tourism salaries. However, these discussions focused more on common grounds, as the managers and employees were both remarkably nuanced in their joint ideological payment positions (ME5), engaging in these conversations with radical advice (ME5.1), "friends and foes" schemata (ME5.2) and one-sided lines of argumentation (ME5.3). The following two conversations between employees and managers illustrate this: ce12: you just don't make good money {in a hotel}. ce13: mmm (.) okay (.) Do you think others {employees} say the same. ce12: DEFINITELY (0.3) >doubtless< cm1: the salary in our company is not that bad (0.2) NO serious complaints. cm2: treat people properly the salary is not the main aspect (.) ANYWAY.
(.) people forget the role of tips in tourism (.) that is the (main part) of. their income.

Micro-level.
The employees and employers maintained their ideological views on tourism wages (MI4). The individual framing was extremely rich in its detail and depth of focus, as the interviewees provided insight into how they arrived at their viewpoints. Both groups reported that these payment ideologies were based on personal experiences of financial exploitation (MI4.1). However, the managers emphasized the industry's heavy price pressure and competition (MI4.2) and the resulting impact on payment calculations (Table 4). From a multilevel framing perspective, the media portrayed two diametrically opposed payment ideologies, which were recurrently supported and deepened in both interactive and individual framing.

Radicalization
The results show that the "working conditions" and "payment" framing saw additional momentum across levels, intensifying their impact on TEI. Both themes are characterized by the cross-level dynamic of "radicalization." We argue that radicalization becomes evident through emotionalism and/or ideologism ( Figure 5).
Emotionalism implies an interpretation of the key indicators of effects, valence and arousal (Winkielman et al., 2007) and is most visible in the framing of working conditions.

Figure 5. Radicalization
The media laid the foundation for a critical engagement by "addressing bad working conditions in tourism." Employees then "condemned bad tourism working conditions," consequently defending the sentiment the media maintained in its reporting. Institutional frame constructs, therefore, influenced individual thinking (Benner and Tripsas, 2012). In our case, the media provide an emotional compass for the framing of tourism working conditions and employees sympathetically accept this institutional guidance. This compass constructs an essentially negative picture of tourism employment and ignores the meaningful and satisfying work experiences of tourism employees (Baum, 2019a;Mooney et al., 2016), the positive attitudes of hospitality students and their potential to become tourism employees (Goh and Lee, 2018) and the decent working conditions offered by "employers of choice" (Solnet et al., 2015). This one-sided framing of working conditions in tourism is problematic for the industry, as good working conditions and positive working experiences are critical for achieving attractive employment images, to a greater extent than extrinsic rewards (Wan et al., 2014).
Managers were "distancing from apparent and potential criticism of bad tourism working conditions," thus turning away from this emotional compass. They specifically pointed to incorrect interpretations/one-sided perspectives and focused on defending institutional work (Lefsrud and Meyer, 2012) by using three main strategies in their individual framing. First, managers "naturalized" and "romanticized" working conditions by relating them to their own "tough" work experiences, constructing specific features of work as either "normal" or/and "exciting." For them, being able to deal with these characteristics proved whether an employee had the right personality and attitude toward tourism work. Second, they assessed or included little of the critical media depictions of the tourism work environment, despite the empirical data proving its complexity and difficulty (Robinson et al., 2019). Third, managers connected their positive constructions of tourism employment to a framing of themselves as insiders while "defensively attacking" (Lefsrud andMeyer, 2012, p. 1479) critics as outsiders, who did not understand or fit the business, thus devaluing their criticism and constructing an emotionally charged "us vs. them" situation.
However, the data show a revaluation of emotionalism in the interactional framing of working conditions. This was clearly evidenced when manager groups/pairs "highlighted the beneficial aspects of tourism employment" while employees tended to "embrace problematic manager-employee relationships." With a focus on the latter, joint activities may also serve as a breeding ground in which the negative impressions or experiences of individuals can turn into negative emotional states for all participants. In fact, bad managerial behavior experienced by individual employees paved the way for collective aversion. Although negative management behavior is not uncommon in tourism (Baum, 2018), peer conversations may generate extreme leverage. Here, it appears easy to go from "problematic relationships" to "managers from hell" (Hight et al., 2019).
Ideologism implies an interpretation based on radical ideas/beliefs that have "considerable staying power" (Benford and Snow, 2000, p. 613) and are, thus resistant to criticism. Our findings illustrate ideologism most clearly through the framing of payments. As the media highlighted "political controversies about tourism salaries," union and employer representatives both disseminated featured stories "sustaining their own ideological views on tourism wages," which were either pro-or anti-employer. These stakeholders attempted to use the media in their "information warfare" to convince the wider public of the legitimacy and credibility of their stance on payment in tourism. In this context, the media are also "a vehicle for carrying and delivering the message to others" (Snow and Byrd, 2007, p. 132). Disseminated messages radicalize the discourse, signal a problematic situation, and thus aggravate the potentially negative impact of commonly used industrial payment policies, which are sometimes used to enhance the attractiveness of tourism employment (McGinley et al., 2017). However, the media are not just an institutional megaphone by which tourism stakeholders disseminate their ideologies. They may also capitalize on the potential of these ideologies, thus supplying a "strategic imperative" (Westby, 2004, p. 287) for the framing of tourism salaries.
Aside from media controversies, the inherent logic of ideologies aims to condemn the meaning systems of opponents (Oliver and Johnston, 2000). This was evidenced in our peer conversations, where managers and employees often "refined joint ideological payment positions" This ideologically charged atmosphere of tourism employment relations adds to the already negative reputation of pay in the industry (Ladkin, 2011;Wan et al., 2014). It hardly allows a more rational discourse on payment, such as discussions on the extent to which payment structures derive from the characteristics of the tourism industry or the (formal) level of employee qualifications and the kinds of jobs available (Casado-Díaz and Hip olito, 2016). The ideological discourse prevents the stakeholders from taking a constructive approach to criticism and hinders innovative solutions.
4.4 Seasonality 4.4.1 Macro-level. Seasonality was represented in the media through 44 articles (9.2%), with broad temporal (covering all years) and spatial (across all newspapers and magazines) dissemination. This reporting highlighted the multifaceted nature of seasonal employment (MA2), with the media discussing so-called "boreout" during low seasons (MA2.1) and burnout following high seasons (MA2.2.), year-round employment approaches (MA2.3) and seasonal job rotation models (MA2.4). The following example illustrates the institutional framing in this context: In fact, there are thousands of individual work crises that occur every winter, when the cable cars stand still and employees can no longer find jobs. (Tiroler Tageszeitung, May 2, 2012, p. 25) 4.4.2 Meso-level. Employees and managers both marginalized the seasonal aspect of tourism work in peer conversations (ME4). In fact, seasonality discussions were mostly avoided in their social encounters (ME4.1). In the best-case scenario, employees provided insights into future seasonal jobs and workplaces (ME4.2) or their off-season vacation plans (ME4.3). The latter is exemplified in the following dialog: ce12: and where are you flying to in summer ce13: uhhhh (.) good question (0.2) but I don't decide that ce12: [laughing] (0.2), so this time I am flying to AUSTRALIA (.) for THREE months The peer conversations among managers were similar. However, they did not focus on future seasonal jobs or off-season plans but rather on recurring seasonal (key) events (ME4.4), as the following discussion illustrates: cm7: BOOOOM (0.2) suddenly IT was here. cm4: IIIIIIT. cm8: right after Christmas and New Year's Eve. cm7: EXACTLY. cm4: OOOOOK but that's nothing out of the ordinary. cm4: every year again.

Micro-level.
A similar framing of seasonality occurred in the interviews. However, the stakeholders spent more time discussing this issue, marginalizing seasonality within a content-related dimension as a result (MI6). The employees stated that seasonality is an unavoidable requirement in tourism work in Tyrol (MI6.1) and the managers likewise took seasonality for granted (MI6.2; Table 5). A multilevel framing perspective shows how the strong media coverage of seasonality was marginalized in the peer conversations and interviews, either by questioning its impact or positing it as a self-evident component of tourism work. Both stakeholder groups argued in a highly rational manner and streamlined the issue into its basic elements. In doing so, they limited the role of seasonality in the construction of employment images in Tyrol.
4.5 Human resource problems 4.5.1 Macro-level. Reports on HR problems comprised 154 articles (32.2%) and had broad temporal (covering all years) and spatial (covering all newspapers and magazines) dissemination. The media emphasized the decreased attraction and loyalty among (potential) employees (MA1), with articles discussing labor shortages (MA1.1), job dissatisfaction (MA1.2) and high labor turnover (MA1.3). The following excerpt provides an example: The total number of seasonal workers has drastically decreased and the labor shortage in the industry has worsened. (Tiroler Tageszeitung, November 15, 2017, p. 21) 4.5.2 Meso-level. Unlike the media, the vast majority of managers refused to discuss employment-related difficulties with their peers (ME1). Furthermore, they attempted to avoid discussions of illness statistics (ME1.1), employee dissatisfaction (ME1.2) and professional shortages (ME1.3). The latter is illustrated by the following interaction  . Interestingly, these issues were framed during the interviews in far less detail than were working conditions or payments, although an interrelation between these topics is apparent. When managers were directly asked about HR problems, most responded as briefly and vaguely as possible, as exemplified in Table 6. From a multilevel framing perspective, the strong media coverage of HR problems was not reflected in the peer conversations and interviews. The employees completely ignored them and the managers largely refused to discuss them, leading to a partial repression of HR problems in the construction of TEI.

Attenuation
The framing of "seasonality" and "HR problems" can be characterized by the dynamic of attenuation, as it lost substance across the levels of analysis. We argue that attenuation is expressed by rationalization and/or repression ( Figure 6).
Rationalization implies an interpretation along with abstract logics and/or formal standards. This is seen in the framing of seasonality, wherein employees and managers were marginalizing "seasonal aspects of tourism work" and "seasonality on a content-related dimension". Both Table 6. HR problems Categories Representative interview experts for HR problems (Interviewee; sub-categories)

MI3
Managers: "I don't actually look at illness reports, only when they threaten me, our business . . .. But so far it has always turned out well." (im15; MI3.1) "We get a not-inconsiderable share (of our workforce) from the east (Eastern Europe). . . . Hungary, Czech Republic, or the former East Germany." (im11; MI3.3) Figure 6. Attenuation stakeholder groups collectively and individually distanced themselves from the media stance, "embracing the multifaceted nature of seasonal employment." Instead, they provided inductive rationales (Lounsbury and Glynn, 2001) based on their work experiences, interests and knowledge. Given the continuous engagement with seasonal work, stakeholders may internalize this component of employment to the extent that it transforms into a taken-for-granted reality. Yıldırım (2021) argues that employees can internalize these problems as a normal, natural part of their job, creating a frail, superficial consent. However, the employment experiences of seasonal workers may undermine this consent. For example, Ainsworth and Purss (2009) found a strong contrast between employers' expectations during recruitment and the reality of their work, where seasonal workers were treated "as a cost and commodity" (230). Similarly, Guillaume et al. (2019) showed that seasonal workers experienced less job quality than standard workers. Nevertheless, sociocultural factors such as the relevance of a "sense of community" (McCole, 2015) for seasonal workers can contribute to more satisfying, sustainable employment and a more positive employment image.
The repression of pessimism also supports attenuation and becomes apparent in the framing of HR problems. On the one hand, the media were "emphasizing the decrease in attraction and loyalty from (potential) employees." This shows that the HR work of an organization influences the construction of TEI. The managers in our study were "refusing (to discuss) employment-related difficulties" in peer conversations and were "downplaying personnel facts and figures" in the interviews, confirming that they are often unaware of recruitment and retention problems (Baum, 2018, p. 727). Paradoxically, the managers still created intersubjective meanings even though they deliberately withheld first-hand knowledge. Consequently, common ground among managers may also emerge through alienation and passiveness. Importantly, the managers did not build common ground regarding HR problems. Rather, they established taboo zones, which prevented the emergence of substantial common ground, stabilizing the common image of poor employment practices. To attract the next generation of tourism employees (Goh and Okumus, 2020), employers must tackle structural HR problems. Attenuation in the context of HR problems gives rise to serious concerns in the construction of tourism employment because it prevents employers and managers from seeing, understanding and changing the management of people in tourism. It precludes the evaluation of many constructive suggestions for fostering more sustainable employment, careers and working practices (Baum, 2019a;Solnet et al., 2015;Chang and Busser, 2020), which could also improve TEI.

Conclusions
This study found that radicalization in the construction of TEI leads to polarized positions, intense rivalry between employers and employees and the reduction of tourism employment questions to problematic issues or the refusal to acknowledge structural employment problems. By contrast, attenuation produces problematic, taken-for-granted assumptions about important facets of tourism employment, thus creating indifference and undermining a constructive discourse on the central aspects of tourism work.

Theoretical implications
This study supports the general use of frame analysis in tourism and hospitality research (Leung et al., 2019), broadening the scope of its application to TEI. Frame analysis is highly useful for exploring the processes and dynamics that create TEI. This approach goes beyond studying the role of individual perceptions about specific features of tourism work, such as bad management behavior (Hight et al., 2019) and the widely accepted, nearly objective employment structures such as limited career opportunities (Goh and Okumus, 2020). Instead, it focuses on the interplay of the individual, interactive and institutional frames/framing of different aspects of TEI, concentrating on the dynamics that characterize these processes and the problematic sociocultural and political constellations these dynamics create for tourism employment.
The research has not examined the dynamics that pervade the construction of TEI. We have conceptualized the main dynamics across all analysis levels as radicalization and attenuation. Radicalization sharpens language, cognition and culture, leading to a "contraction of meaning" that stabilizes and reinforces restricted frames to the "exclusion of alternative interpretations" (Cornelissen et al., 2014, p. 699). Radicalization intensifies a destructive, dysfunctional thrust of ideological and emotional antagonism that negatively affects tourism stakeholders' capacity to find productive compromises. By contrast, attenuation supports the development, preservation and extension of superficial forms of language, cognition and culture. This results in vague and frail meaning constructs that manifest as false issues or nonissues (Ansari et al., 2013), which hinder fundamental alignments of tourism employment.

Practical implications
The poor, one-sided image of employment in tourism industries is not driven exclusively by major employment issues but is also based on the dynamics that lead to these image constructions. Thus, gaining a deeper knowledge of these processes is important for tourism employment practices. The two cross-level dynamics of "radicalization" and "attenuation" make these processes both conflict-driven and destructive, as well as indifferent and superficial. Therefore, breaking through and changing these dynamics is an important step toward improving and consolidating TEI.
First, more pragmatism is required in the framing of tourism employment and work. Pragmatism from all stakeholders, including a more balanced media information policy, should aim to de-emotionalize and de-escalate image construction processes, allowing "the obvious contrast of glamour" and "low menial status" (Riley et al., 2002, p. 23) to be set aside. Pragmatism stresses the variety of work in tourism while exploring the potential for attractive jobs and decent working conditions within the industry. A more problem-oriented, rational framing of tourism employment could create more sustainable employment relations, a more productive and humane working environment, and thus an attractive employment image.
Moreover, the framing of tourism employment must be transformed based on open discussions and a critical review of the industry workplace (Baum, 2019a(Baum, , 2019b, including all aspects that are negative or positive, appealing or repellent, obvious or hidden. To strengthen innovative approaches, younger generations of stakeholders who stand for progressive ideas should be involved in these discussions. This could lead to a more differentiated understanding, overcome rationalization and repression and help tourism stakeholders to jointly transform tourism employment. Tackling radicalization and attenuation in the construction of TEI calls for a "new deal" between tourism stakeholders. This should aim to end "information warfare" while inaugurating a new era of tourism employment by fulfilling both the legitimate interests of employees (e.g. decent social standards of work) and employers (e.g. flexible working hours). Government institutions, particularly at the regional level, should actively support this new deal by, for example, offering sustainable childcare infrastructure or public transport, including in remote locations. The powerful divergent interests and economic and social pressures in this industry make this a complex task. However, given the massive economic (e.g. labor scarcity) and social (e.g. precarious work) problems of tourism employment, which were exacerbated by the COVID-19 pandemic (Baum et al., 2020), changing how employment images are constructed is crucial for the success of tourism industries.

Limitations and future research
This study has several limitations. First, we concentrated on employees and managers/ employers as the most relevant stakeholders. However, other stakeholders may also have considerable influence on the construction of TEI. For example, local residents could be of special interest for future studies in this field. Second, at the meso-level, we analyzed data collected in the specific context of arranged get-togethers. More spontaneous encounters could generate more authentic data. Third, at the macro-level, we were limited to input, such as media reports and comments over a specific period, which confined us to cultural conventions that garnered media attention. Finally, since the empirical part of this study was conducted prior to 2020, it is important to study how the COVID-19 pandemic has influenced the cross-level dynamics of radicalization and attenuation in the construction of TEI.

The data analysis was carried out in German and the empirical examples were translated into
English for the material's disclosure (interviews and media products were translated nonliterally and conversations were largely translated literally).