Linear and non-linear relationships between job demands-resources and psychological and physical symptoms of service sector employees. When is the midpoint a good choice?

fpsyg-13-950908 September 23, 2022 Time: 14:30 # 1 TYPE Original Research PUBLISHED 29 September 2022 DOI 10.3389/fpsyg.2022.950908 OPEN ACCESS EDITED BY Gabriele Giorgi, European University of Rome, Italy REVIEWED BY Vicente Martinez-Tur, University of Valencia, Spain Beáta Sz. G. Pató, Eötvös Loránd University, Hungary *CORRESPONDENCE Alicia Arenas [email protected] SPECIALTY SECTION This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology RECEIVED 23 May 2022 ACCEPTED 24 August 2022 PUBLISHED 29 September 2022 CITATION Sanclemente FJ, Gamero N, Arenas A and Medina FJ (2022) Linear and non-linear relationships between job demands-resources and psychological and physical symptoms of service sector employees. When is the midpoint a good choice? Front. Psychol. 13:950908. doi: 10.3389/fpsyg.2022.950908 COPYRIGHT © 2022 Sanclemente, Gamero, Arenas and Medina. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Linear and non-linear relationships between job demands-resources and psychological and physical symptoms of service sector employees. When is the midpoint a good choice? Francisco J. Sanclemente, Nuria Gamero, Alicia Arenas*and Francisco J. Medina Department of Social Psychology, University of Seville, Seville, Spain Related to the research of working conditions, the link between organizational factors and health was traditionally analyzed using linear models. However, the literature analysis suggests inconsistencies in linear models predicting workers’ health levels. To clarify this issue, this exploratory research compares the linear and non-linear relationships between job demands-resources (task complexity, time pressure, contact with users, and job autonomy), and the psychological and physical symptoms of employees working in the main five service subsectors: commerce, horeca (hotels, restaurants, and cafés), public administration, education, and healthcare. With a final sample of 4,047 participants, our study data were extracted from the II Andalusian Working Conditions Survey. Following the theoretical framework of JD-R Model and considering the Vitamin Model theoretical approach for nonlinear relationships, our results showed that there were significant differences among the five subsectors analyzed regarding the linear and non-linear relationships between job demands-resources and psychological and physical symptoms of employees. Furthermore, task complexity generated non-linear relationships in higher proportion than time pressure and contact with users. Likewise, non-linear relationships found showed a U-shape. Moreover, the findings of non-linear relationships suggested that medium levels of task complexity should not be exceeded to avoid further negative impact on psychological and physical symptoms for service sector employees, preserving their health. Finally, some general practical implications of work environment interventions are suggested. KEYWORDS linear and non-linear relationships, job demands, job resources, psychological and physical symptoms, service sector Frontiers in Psychology 01 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 2 Sanclemente et al. 10.3389/fpsyg.2022.950908 Introduction Stressful working conditions and their negative impact on employees’ health are considered an important issue for both the scientific community and society, generating changes in policies and human resources management. In article 151 of the Treaty on the Functioning of the European Union, the EU recognizes how important improvements in working conditions are. Furthermore, European social partners recognized the relevance of psychosocial risks by signing the Framework Agreements on Work-related Stress in 2004 (EU-OSHA,2004). In Europe, 25% of employees experience work-related stress, reporting that work affects their health negatively (Eurofound, 2016). Additionally, differences in working conditions between groups of employees are sector-related (Eurofound and EUOSHA,2014). These statements mean that the health of millions of employees is negatively affected by work stress in the EU and reveal differences among work activities. Likewise, the sixth European Working Conditions Survey (6th EWCS) indicates that Spanish employees are exposed to high work intensity, which is linked with a negative impact on their health and wellbeing, revealing a high proportion of job activities related to high levels of time pressure, task complexity, and contact with users or clients (Eurofound,2016). Additionally, the Spanish labor market is characterized by a working population primarily occupied in the service sector (Instituto Nacional de Estadística [INE],2021). Guided by the magnitude of the negative impact of stressful working conditions on employees’ health, the purpose of this exploratory study was to investigate how employees in five main service subsectors are negatively affected in terms of psychosomatic symptoms due to specific job demands that act as work stressors. To contribute to this understanding, first, we separately analyzed psychological and physical symptoms. Second, job demands were considered as specific job demands (task complexity, time pressure, and contact with users). And third, the impact of these specific job demands and job autonomy as job resource on employees’ psychosomatic symptoms was analyzed in five main service subsectors with different job activities through linear and non-linear models. The service subsectors considered in our study were commerce, horeca (hotels, restaurants, and cafés), public administration, education, and healthcare. Our objectives were threefold. First, to examine what specific job demandsresources affect psychological and physical symptoms. Second, to analyze whether specific job demands-resources impacted the psychosomatic symptoms of employees in a linear or nonlinear way, and to define the shape of these relationships. And finally, to determine whether specific job demands-resources had different effects on the five different subsectors analyzed. Therefore, the main contribution of this exploratory study was to provide a comprehensive overview of the effects of the three main job demands and job autonomy as job resource on the psychosomatic symptoms of employees in the service sector, considering the physical and psychological symptoms separately, and showing the linear and non-linear nature of these relationships detailed by subsectors. Specifically, considering the five main service subsectors of the service sector, our study will show how task complexity, time pressure and contact with users, as job demands, will have a linear versus non-linear impact on health of employees depending on the subsector. Moreover, our results will shed light on how job autonomy, as job resource, will show a linear versus non-linear pattern in different subsectors. Likewise, our study will show the differential effects of sociodemographic variables, such as gender, age, and job tenure, on the job demands-health relationship. Relationship between job demands and psychosomatic symptoms of employees The negative effect of job demands on physical and psychological employees’ health has been shown in numerous studies (Karasek,1979;Bakker et al.,2003;Väänänen et al., 2003;Bakker and Demerouti,2007,2013). Theoretical models that explain this impact, such as the Job Demands-Resources (JD-R) Model (Bakker and Demerouti,2007) and the Vitamin Model (Warr,1987) signal how effects of job characteristics are. The JD-R model is a comprehensive model characterized by differentiating organizational factors such as job demands that affect employees’ health and job resources as job autonomy that protects employees’ health (Bakker and Demerouti,2007). Job demands are defined as physical, psychological, organizational, or social aspects of work that involve sustained effort and that have a physical and psychic toll (Bakker and Demerouti, 2013) signaling that job demands have effects on physical and psychological employees’ health. Likewise, job resources are those physical, psychological, organizational, or social aspects of work that can reduce job demands and associated physiological and psychological costs, achieving job goals, or stimulating personal growth, learning, and development (Bakker and Demerouti,2013). Thus, job demands act as the main predictors of variables such as exhaustion or psychosomatic health problems, and job resources act as the main predictors of job satisfaction, motivation, and engagement. At the same time, there is an interaction between job demands and job resources so that the job resources decrease the impact of job demands on strain and, on the other hand, job demands increase the impact of job resources on motivation-engagement (Bakker and Demerouti,2013). Although the JD-R model signals relationships between job demands-resources and their outcomes, the authors do not indicate if these relationships are linear or non-linear. However, the traditional research, based in the framework of JD-R and other models, assumed linear effect of job characteristics (Ferris et al.,2006). One Frontiers in Psychology 02 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 3 Sanclemente et al. 10.3389/fpsyg.2022.950908 of the most commonly statistical method used in traditional organizational research is the linear regression model. The linear regression model assesses whether one or more predictor variables explain the dependent variable, and is based in four main assumptions as linearity, homoscedasticity, independence, and normality. In this sense, Ferris et al. (2006) suggest that theoretical development and data analytic methodology created together a linear organizational science based in linear relationships. However, the assumption of linearity implies ignoring other possible effects (Warr,2007), leading in possible inconsistency in results when data are analyzed exclusively from a linear perspective (Ferris et al.,2006). Likewise, traditional research assuming linear relationships only is too simplistic, thus future research should focus on the dynamic relations among the concepts in the framework of JD-R model (Schaufeli and Taris, 2014). In this line, the Vitamin Model considers the effects of job characteristics on health and wellbeing as an analogy to the role of vitamins on physiological health. In consequence, further vitamin intake produces constant effects without an improvement in health, this constant effect occurs with vitamins C and E (CE pattern). In the worst of the cases this further intake could cause a decrease in health. This additional decrease effect occurs with vitamins A and D (AD pattern with an inverted U-shape), this way, job demands are identified as a job characteristic with an AD pattern effect on affective wellbeing with an inverted U-shape. This model assumes that both CE and AD effects are non-linear (Warr,1987). Thus, the Vitamin Model (Warr,1987) gives theoretical support to non-linear relationships between work conditions, health, and wellbeing. However, previous research on non-linear relationships between job conditions and personal outcomes shown in Table 1 suggest some contradictory findings. For example, Lawson et al. (2009) found only linear relationships between psychosocial working conditions and psychological health in a sample of statebased police officers. Nevertheless, Noblet et al. (2009) found curvilinear links between workload and psychological health in a similar sample. Additionally, Warr (1994) carried out several recommendations for detection of non-linear relationships. First, using large samples, more than 500 participants. Second, using homogeneous work activities and a wide range of work characteristics. Third, using specific (better than general) job demands. And finally, distinguishing between psychological health (mental health) and physical health. However, as we can see in Table 1, few studies have large numbers of participants, some of them analyzed heterogeneous job activities and some of them only considered psychological health. According to Warr’s suggestions, to analyze non-linear relationships, a large sample and a wide range of job characteristics are required. In this sense, the use of a large work conditions survey is a chance to test this research topic. The impact of job demands on the service sector The service sector is composed of different subsectors with specific characteristics and job activities (Eurofound,2014b). Regarding the health of service sector employees in Europe, it is emphasized that service-oriented sectors and workers employed in manual occupations experienced significantly worse working conditions in terms of exposure to physical and psychosocial risks (Eurofound,2014a). Specifically, administrative services and food and beverage services showed lower job quality indicators such as low pay rates, irregular working time schedules, no control over working time, and high levels of exposure to physical and psychosocial risks (Eurofound,2014a). Additionally, the Spanish labor market is characterized by a population mainly occupied in the service sector. According to the National Institute of Statistics (INE) in 2019 (before the COVID-19 pandemic situation) the percentage of people employed in the service sector in Spain was 75.5%. Likewise, data from the 6th European Working Conditions Survey EWCS (Eurofound,2017) indicates the percentage of Spanish employees whose job involves working at very high speed (64%), working to tight deadlines (64%), being in situations that are emotionally disturbing (53%), handling angry customers (53%), dealing directly with people such as customers, pupils, patients (67%) and being required to hide their feelings (54%). These data reveal that Spanish employees are exposed to high levels of time pressure, task complexity, and contact with users (Eurofound,2017). Time pressure is defined as the extent to which employees feel they do not have enough time to finish their work or need to work faster than usual (Binnewies and Wörnlein,2011), or the amount of work a person has, combined with the required speed to fulfill it (Debusscher et al., 2014). Task complexity can be defined from an objective or a subjective point of view. The objective perspective considers task characteristics independent of the task performer from a structuralist or resource requirements points of view. However, the subjective perspective considers task properties and task performer characteristics in an interaction viewpoint. Thus, task complexity is defined by Liu and Li (2012) from an objective perspective as the aggregation of intrinsic task characteristics influencing the performance of a task. In this sense, Campbell (1988) also defined task complexity as ‘related directly to the task attributes that increase information load, diversity, or rate of change’ (p. 43). However, from a subjective point of view, when the complexity of the task exceeds the capacity of the worker, the worker will perceive it as perceived psychological complexity (Liu and Li,2012). In this line, task complexity is defined as an individual’s perception of how complex a task is (Debusscher et al.,2014). Finally, contact with users is a common characteristic of the job among the service sector Frontiers in Psychology 03 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 4 Sanclemente et al. 10.3389/fpsyg.2022.950908 TABLE 1 Summary of revised works about non-linear research by kind of sample, measure of job demands, relationships found, and variables involved. Non-linear research papers Sample Job demands Relationships Variables Authors/Year/Journal NHomo JA HeteroJA As one variable Specific Linear Nonlinear Binnewies and Wörnlein (2011) Journal of Organizational Behavior 112 X X X X Time pressure-creativity-job control-affect Chung-Yan and Butler (2011) Canadian Journal of Behavioural Science 259 X X X Job complexity-demands-abilities fit-job satisfactionproactive personality Debusscher et al. (2014) PLoS ONE 45–52 130 X X X X Task complexity-work pressure-task performance-state neuroticism De Dreu (2006) Journal of Management 109 212 X X X X Workload-task conflict-relational conflict-innovation-resource scarcity-information exchange-collaborative problem solving-task interdependence-goal attainment Lawson et al. (2009) Health Promotion International 587 X X X Job demands-job control-social support-organizational justice-psychological health-job satisfaction Maltarich et al. (2010) Journal of Applied Psychology 11,892 X X X X Cognitive job demands-cognitive ability-voluntary turnover-job satisfaction Noblet et al. (2009) Stress and Health 2,085 X X X X Job demands-social support-psychological health-job satisfaction-organizational affective commitment Parkes (2016) Applied Ergonomics 901 X X X X Job demands-job controlphysical work environment-sleep-age Reis et al. (2017) European Journal of Work and Organizational Psychology 54 X X X X Time pressure-job control-vigor-absorption Rydstedt et al. (2006) Work and Stress 4,154 6,000 X X X X Job demandsjob control-social support-job satisfaction-psychological health Sawang (2012) International Journal of Manpower 307 X X X Job demands-work engagement-social support Schmitt et al. (2015) Journal of Personnel Psychology 191 X X X Time pressure-work engagement-illegitimate task Sheng et al. (2019) International Journal of Stress Management 67 X X X Time pressure-work engagement-psychological capital-sleep quality Tanskanen et al. (2016) Journal of Happiness Studies 7,867 X X X X Job demands-job control-social support-work engagement Van Ruysseveldt and Van Dijke (2011) Journal of Vocational Behavior 11,099 9,738 X X X Workload-workplace learning opportunitiesautonomy Total number of papers 8 7 7 8 10 14 Homo JA, homogeneous job activity sample; Hetero JA, heterogeneous job activity sample; sample size: 8 papers with N<500 and 7 papers with N>500. Frontiers in Psychology 04 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 5 Sanclemente et al. 10.3389/fpsyg.2022.950908 referred to interactions with users or customers (Szczygieł and Bazi´ nska,2013). The 6th EWCS also analyzed work intensity as an index related to the level of job demands, such as workload, job absorption in terms of mental and physical energy, jobs that require juggling with various demands, and difficulties performing tasks in the most effective way (Eurofound,2016). Our analysis, based on the work intensity index of the 6th EWCS (see Table 2), indicates that the five service subsectors considered had different levels of time pressure and task complexity. Although the time pressure work intensity index is quite similar to our study variable, the task complexity work intensity index is mainly related to emotional job demands, while in our study, task complexity is related to high levels of attention, complicated or difficult tasks, computer jobs and intellectually and emotionally demanding jobs. Likewise, Karasek (1979) pointed out the ‘tendency to describe all structurally determined work characteristics as job demands regardless of their drastically different effects on psychological functioning’ and also that a more specific distinction between job demands is needed due to the ‘inconsistent finding that time pressure demands are associated with strain symptoms, while intellectual demands are not’ (p. 286). This consideration suggests differences between effects on employees’ health of intellectual-cognitive job demands and other job demands such as time pressure based first on the aforementioned premise pointed out by Karasek (1979) about differences between the effects of job demands on employees’ health. Second, Bakker and Demerouti (2013) consideration that job demands involve a sustained effort with a physical and psychological toll. And third, effects of job characteristics on employees’ health and distinction between psychological and physical symptoms of Väänänen et al. (2003). Previous findings indicate that each job activity, and its specific job characteristics has specific psychosocial health determinants (Väänänen et al.,2003). Within the service sector, there are occupations that could differ in the type of demands they receive from users, in physical overload, and to what extent work interferes with employee personal life (Eurofound and EU-OSHA,2014). Similarly, employees in these occupations differ in their level of education and tenure and the perceived stability of their contract (Eurofound, 2016). All these aspects could influence the link between job demands and employees’ health. In this sense, the employees of the education and healthcare subsectors generally receive high emotional demands from users (Szczygieł and Bazi´ nska, 2013). These occupations usually spillover to personal life because there are less chances of recovering from work activity (Lockwood,2003). Similarly, job occupations could also differ in the level of education of employees, with a higher level of employee education in the public administration, education, and healthcare subsectors (Eurofound,2016). Educational level could modify the personal appraisal of job demands, considering moderated levels of job demands as a challenge (LePine et al., 2005;Van den Broeck et al.,2010). In other job occupations with more unstable employment and lower educational level, job demands could be perceived by employees as a hindrance (LePine et al.,2005;Van den Broeck et al.,2010). Additionally, the JD-R Model (Bakker and Demerouti,2007) predicts that job resources, as job autonomy, protect employees’ health against the negative influence of job demands. Likewise, job autonomy was defined by Hackman and Oldham (1975) as ‘the degree to which the job provides substantial freedom, independence, and discretion to the employee in scheduling the work and in determining the procedures to be used in carrying it out’ (p. 162). Based on these premises, we could expect differences depending on the service subsectors analyzed. Thus in education and healthcare subsectors due to a more stable employment consequence of stable contracts as public servants in large public organizations (as public hospitals, high schools, and universities), higher requirements of educational level and higher level of job autonomy, the job demands related to the task, such as task complexity, could produce a lower negative influence on health when job demand levels are higher. Consequently, we could predict lower and linear effects of job demand on employees’ health in the education and healthcare TABLE 2 Work intensity index by sector EU28 (%) and their correspondence with our study variables time pressure and task complexity. Work intensity variables Commerce and horeca Public administration Education Healthcare Working at high speed (3/4 of time +) 41 21 19 35 Working to tight deadlines (3/4 of time +) 37 30 25 35 Three or more pace determinants 34 30 20 28 Time pressure (Mean) 37.33 27 21.33 32.67 Hide emotions (most of time/always) 36 37 35 44 Handling angry clients (3/4 of time +) 21 19 23 28 Emotionally disturbing situations (3/4 of time +) 9 14 13 24 Task complexity (Mean) 22 23.33 23.67 32.00 Source, Sixth European Working Conditions Survey (Eurofound,2016); Index from 0 to 100%. Frontiers in Psychology 05 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 6 Sanclemente et al. 10.3389/fpsyg.2022.950908 subsectors However, in the commerce and horeca subsectors where most of the Spanish companies in these subsectors are private small companies, offering unstable employment and low pay rates, with low requirements of educational level and low level of job autonomy, we could predict higher and nonlinear effects of job demand on employees’ health. Similarly, we could predict higher and non-linear effects of job demand on employees’ health in public administration due to low level of job autonomy as a specific subsector characteristic with a highly hierarchical and inflexible bureaucratic work system. Regarding the shape of non-linear relationships, as psychological and physical symptoms are negatively related to health and wellbeing, thus the higher the symptoms the lower the health, we could predict a U-shape. Because there is no previous evidence regarding the effect of job demands using a broad occupational analysis, we will elaborate our hypotheses with an exploratory character. Thus, regarding these precedents, we propose the following hypotheses: Hypothesis 1: The linear and non-linear relationships between specific job demands and psychological and physical symptoms of employees will differ depending on the service subsectors analyzed. Hypothesis 2: When the effect of job demands-resources on psychological and physical symptoms of employees is nonlinear a U-shaped relationship is expected. Materials and methods Sample and procedures To carry out this study, a systematic review of recent literature related to non-linear relationships between job demands-resources and their outcomes was carried out first. In order to systematize our literature exploration and analyses, we followed the recommendations of the Cochrane Handbook for Systematic Reviews (Higgins and Green,2008). To answer the question ‘which type of designs and methodologies are used in recent studies on non-linear relationships between job demands and their outcomes?’, we applied a specification of criteria for including and excluding studies in the literature review. The eligibility criteria were as follows: journals indexed in Web of Science (WoS) or SCOPUS, with high quality index Q1 or Q2 containing ‘non-linear relationship,’ ‘curvilinear relationship,’ and ‘job demands’ as keywords in the article title or in the abstract. Result records identified through database searching were 49 documents found in SCOPUS and 38 documents found in WoS. After removing duplicate articles and evaluating their eligibility, we selected 15 studies (shown in Table 1) to be included in our qualitative synthesis as a significant sample of research in this field. Data were extracted from the second Andalusian Survey of Working Conditions (II ASWC). The II ASWC was carried out in 2012 on a sample of 8,854 employees. This survey is carried out by the Andalusian Institute for Prevention of Occupational Risks (IAPRL) related to the Andalusia Regional Government, to know the perception that Andalusian employees have about their health in relation to the work activity they perform and the conditions in which they do it, as well as their knowledge about the organization and other preventive activities carried out by companies on their workplaces. The II ASWC survey was directed at the employed population, through a stratified random sampling proportional to the size of the population under study and representative of the employed population resident in municipalities larger than 5.000 inhabitants with a confidence level of 95.5%. The details of the II ASWC are available in a public document edited by the Andalusian Government (Instituto Andaluz de Prevención de Riesgos Laborales [IAPRL],2012). We have preferred to use a regional survey more than a national or European one, to increase homogeneity in participants and reduce cultural differences. Although a more recent ASWC is not available yet and even some variables could be sensitive to the economic and social context (Sanclemente et al.,2020), their nature is the same and the relationships between them (lineal or non-linear) might be maintained. Furthermore, the II ASWC sample shows a great interest for our study because the Andalusian economy depends mostly on the service sector (Instituto Nacional de Estadística [INE],2019). Thus, the final sample used for our study was composed of service sector employees, extracted from the II ASWC of 2012 (N= 4047), with a proportion of 52.5% women and 47.5% men, a mean age of 38.01 years (SD = 10.41), and a mean job tenure of 8.41 years (SD = 9.12). The age ranged from 18 to 65 years, and the job tenure from 6 months to 48 years. Additionally, from this final sample, we extracted five subsamples by service subsectors related to commerce (n= 1,498) with a proportion of 50.2% women and 49.8% men, a mean age of 36.18 (SD = 10.60) and a mean job tenure of 7.29 (SD = 8.47). Horeca (n= 655) with a proportion of 46% women and 54% men, a mean age of 34.71 (SD = 10.11) and a mean job tenure of 5.34 (SD = 7.06). Public administration (n= 778) with a proportion of 42.4% women and 57.6% men, a mean age of 41.83 (SD = 8.87) and a mean job tenure of 12.1 (SD = 10.27). Education (n= 519) with a proportion of 58.8% women and 41.2% men, a mean age of 40.08 (SD = 9.81) and a mean job tenure of 9.52 (SD = 9.28). And healthcare (n= 597) with a proportion of 72.9% women and 27.1% men, a mean age of 39.45 (SD = 10.5) and a mean job tenure of 8.85 (SD = 9.19). It should be noted that in the healthcare and education subsectors, the proportion of women is higher than that of men, whereas the opposite occurs in public administration. Regarding job tenure, employees in the horeca subsector showed the lowest mean job tenure, suggesting that this sector is not as stable as the rest of subsectors analyzed. Frontiers in Psychology 06 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 7 Sanclemente et al. 10.3389/fpsyg.2022.950908 Measures In terms of job demands, two general job demands were selected, task complexity and time pressure, and a specific job demand found in the service sector, which is contact with users. Likewise, job autonomy was considered as a job resource. Scales and items used to measure task complexity, time pressure, and contact with users were extracted from the Spanish version of the Copenhagen Psychosocial Questionnaire (COPSOQ) (Moncada et al.,2005). In addition, the employees also reported the number of symptoms they suffered recently and most frequently. In order to measure the psychosomatic symptoms of the employees, we distinguished between psychological symptoms and physical symptoms. Based on Väänänen et al. (2003), our criteria for classifying the symptoms into both categories were made based on whether the symptoms reported by employees were manifested on a physical level (soma), i.e., perception of discomfort, pain, or fatigue in a particular part of the employee’s physical body, or whether the symptom was manifested as the perception of some kind of discomfort on a psychological or mental level (psyche), i.e., low mood or feeling tense and irritable. Although some physical symptoms considered in our study could be of mental or emotional origin, they are experienced by employees as pain or physical discomfort (Ford et al.,2011). However, determining the origin of these symptoms was not part of the objective of this study. As it is well established in crosssectional studies psychological and physical symptoms predict employee absenteeism due to sickness (Väänänen et al.,2003). Nevertheless, longitudinal study conducted by Väänänen et al. (2003) suggests that symptoms first cause short absences, and in long-term (9 months to 1 year) sick leaves prescribed by a physician. Thus, the relationship between psychosomatic symptoms and employees’ health is not direct and requires a relatively long process influenced by many factors (Väänänen et al.,2003). Finally, the Cronbach alpha values for the scales reached the threshold of (α= 0.70) suggested by Nunally (1978), showing good internal consistency. Task complexity Following Liu and Li (2012) objective definition of tax complexity as the aggregation of intrinsic task characteristics influencing the performance of a task, task complexity was measured with five items like the following: ‘To what extent does your job involve perform complex, complicated, or difficult tasks?’ The response ranged from 1 ‘Almost never or never’ to 5 ‘Always or almost always.’ The five items of task complexity regarding intrinsic task characteristics were related to sensory, cognitive, and emotional psychological demands as requirements of high level of attention, performing complicated or difficult tasks, working with computers, intellectually demanding jobs, and emotionally demanding jobs. The factor structure was tested twice through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). EFA showed a one-factor solution and a Kaiser–Meyer–Olkin (KMO) index of 0.72 above values considered by Kaiser (1974) as satisfactory for factor analysis (between 0.70 and 0.80). The Bartlett sphericity test was significant (χ2= 4011.39, df = 10, p<0.001) showing data suitability for factor analysis. The one-factor CFA showed an adequate fit to the data (χ2= 201.11, df = 5, p<0.001, RMSEA = 0.088, CFI = 0.96, GFI = 0.98, NFI = 0.96). Additionally, factor loadings ranged from 0.49 to 0.83 above the minimum factor loading cut-off of 0.30 commonly accepted in factor analysis (Bandalos and Finney,2010). Cronbach alpha coefficient was 0.71. Time pressure It was measured with four items like the following: ‘To what extent does your job involve working very fast?’ The response ranged from 1 ‘Almost never or never’ to 5 ‘Always or almost always.’ The four items of time pressure were related to working fast, working with tight deadlines, doing repetitive tasks in a very short time, and performing various tasks at the same time. EFA showed a one factor solution, a KMO index of 0.75, and a significant Bartlett sphericity test (χ2= 3356.01, df = 6, p<0.001) showing data suitability for factor analysis. The onefactor CFA showed an adequate fit to the data (χ2= 24.08, df = 2, p<0.001, RMSEA = 0.052, CFI = 0.99, GFI = 0.99, NFI = 0.99). Additionally, factor loadings were adequate, ranging from 0.68 to 0.79. Cronbach’s alpha coefficient was 0.74. Contact with users It was measured with a single item: ‘To what extent does your job involve dealing directly with people who are not employees?’ The response ranged from 1 ‘Almost never or never’ to 5 ‘Always or almost always.’ Contact with users is related to contact with clients, users, or pupils. Considering that the concept is simple and clear, the use of a single item to measure it was well justified (Fisher et al.,2016). Job autonomy It was measured with four items like the following: ‘At your job, can you choose or change the order of tasks?’ The response ranged from 1 ‘Almost never or never’ to 5 ‘Always or almost always.’ The four items of job autonomy were related to the possibility of change or modify the order of tasks, the rhythm of work, the distribution, and/or duration of work breaks, and put your own ideas into practice at work. EFA showed a one-factor solution, a KMO index of 0.77, and a significant Bartlett sphericity test (χ2= 6810.03, df = 6, p<0.001) showing data suitability for factor analysis. The one-factor CFA showed an adequate fit to the data (χ2= 72.51, df = 2, p<0.001, RMSEA = 0.093, CFI = 0.99, GFI = 0.99, NFI = 0.99). Furthermore, the factor loadings were adequate ranging from 0.59 to 0.88. Cronbach’s alpha coefficient was 0.82. Frontiers in Psychology 07 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 8 Sanclemente et al. 10.3389/fpsyg.2022.950908 Psychosomatic symptoms It was measured by showing to the participants a list of eleven symptoms and then asking them to report the number of symptoms on the list that they had suffered recently and frequently. Subsequently, symptoms were classified as five types of psychological symptoms and six types of physical symptoms, generating two separated variables as follows. Psychological symptoms: the five types of psychological symptoms considered were: sleep problems, difficulty concentrating, forgetting things, feeling tense/irritable, low mood. Each case could present a minimum of zero to a maximum of five psychological symptoms. Therefore, this quantitative variable represents the number of psychological symptoms reported by each employee. Physical symptoms: the six types of physical symptoms considered were: feeling tired, having headaches, dizziness, digestive disorders, eye problems, or other muscle aches. Each case could present a minimum of zero to a maximum of six physical symptoms. Therefore, this quantitative variable represents the number of physical symptoms reported by each employee. Considering each symptom on the initial list as a dichotomic item (presence or absence of symptom), reliability was assessed applying the Kuder–Richardson 20 (KR20) reliability test for dichotomic items (Kuder and Richardson,1937). Reference values for this index are the same as the Cronbach’s alpha coefficient. The Kuder–Richardson index (KR20) was 0.89 showing adequate reliability. ANOVA applying Cochran’s Qtest for dichotomic items was significant [Q(9,4046)= 11520.60, p<0.001] showing statistically significant between-subjects discrimination. Control variables Occupational gender roles and age show a significant relation to workplace stress and mental health in workers (Kerr et al.,2021). Furthermore, job tenure is associated with symptoms of musculoskeletal disorder (Warren et al., 2015). Thus, gender, age, and job tenure were selected as control variables due to their relationship with the psychological and physical symptoms of our study output variables. Before including the gender variable in our study, it was converted into a dummy variable that assigns a cero value to men and one value to women. Data analysis In order to test our hypotheses about non-linear relationships, we conducted hierarchical multiple regression analysis (Cohen et al.,2003). Our study variables were introduced in three successive steps. In the first step, control variables such as gender, age and job tenure; in a second step, linear variables (task complexity, time pressure, and contact with users); and finally, in a third step, quadratic variables (task complexity squared, time pressure squared, and contact with users squared). Variables were mean-centered before we calculated multiple regression models (Cohen et al.,2003). Models with a significant incremental of R2(1R2), when squared variables are entered in the model, signal a significant incremental variance adding explanatory power to the model (Cohen et al.,2003). Effect size was reported as Cohen’s d statistic (Cohen,1988). Results Table 3 presents means, standard deviations, and correlations for the study variables. Regarding the correlations for the job demand variables first, task complexity and time pressure were significant and positively related to physical symptoms and psychological symptoms. Second, contact with users was significant and positively related to psychological symptoms. Regarding the correlations for job resources, job autonomy was significant and positively related to task complexity and contact with users, and negatively related to time pressure. Furthermore, both physical and psychological symptoms were significant and positively related. Finally, regarding the correlations for control variables, gender, age, and job tenure were significant and positively related to physical and psychological symptoms. Additionally, a mean comparison analysis showed some gender differences. Therefore, women in the service sector (N= 4047) showed a significant higher number of physical and psychological symptoms than men [physical symptoms: women: M= 0.74, SD = 1.03; men: M= 0.53, SD = 0.84; t= 7.12, p<0.001; effect size: d= 0.22); psychological symptoms: women: M= 0.45, SD = 0.92; men: M= 0.34, SD = 0.76; t= 4.09, p<0.001; effect size: d= 0.13)]. Relationships between job demands-resources and psychological and physical symptoms by sectors For commerce (Table 4), regarding psychological symptoms, the results of Model 5 showed positive linear relationships among task complexity, time pressure, contact with users, and psychological symptoms. Likewise, Model 5 showed a positive non-linear relationship between task complexity and psychological symptoms. Regarding physical symptoms, the results of Model 5 showed a positive linear relationship between time pressure and physical symptoms. For horeca (Table 5), regarding psychological symptoms, Model 5 results showed positive linear relationships among task complexity, contact with users and psychological symptoms. Additionally, Model 5 showed positive non-linear relationships among task complexity, contact with users and psychological Frontiers in Psychology 08 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 9 Sanclemente et al. 10.3389/fpsyg.2022.950908 TABLE 3 Descriptive statistics and correlations for the study variables. Variable Mean SD 1 2 3 4 5 6 7 8 9 (1) Gender − − − (2) Age 38.01 10.41 −0.11** − (3) Job tenure 8.41 9.12 −0.12** 0.64** − (4) Task complexity 3.18 0.94 0.02 0.04** 0.13** − (5) Time pressure 3.05 0.98 0.00 −0.08** −0.01 0.42** − (6) Contact with users 4.36 1.14 0.07** −0.08** −0.07** 0.14** 0.03 − (7) Job autonomy 3.38 1.10 −0.02 0.15** 0.09** 0.10** −0.05** 0.10** − (8) Physical symptoms 0.64 0.94 0.11** 0.06** 0.06** 0.14** 0.16** 0.02 0.01 − (9) Psychological symptoms 0.39 0.85 0.06** 0.08** 0.06** 0.14** 0.10** 0.04** 0.02 0.50** – **p<0.01; two tailed; (N= 4047). TABLE 4 Hierarchical multiple regression analysis models predicting psychological and physical symptoms in commerce. Psychological symptoms Physical symptoms Variables M1 M2 M3 M4 M5 dM1 M2 M3 M4 M5 d Gender 0.13** 0.12** 0.13** 0.12** 0.12** 0.24 0.11** 0.11** 0.11** 0.11** 0.11** 0.22 Age 0.10** 0.11** 0.09** 0.09** 0.09** 0.14 0.11** 0.12** 0.11** 0.11** 0.11** 0.17 Job tenure −0.02 −0.02 −0.03 −0.02 −0.02 −0.02 −0.02 −0.02 −0.02 −0.02 Task complexity 0.07*0.06*0.08*0.08** 0.14 0.02 0.02 0.02 0.02 Time pressure 0.08** 0.09** 0.09** 0.09** 0.16 0.14** 0.15** 0.15** 0.15** 0.26 Contact with users 0.05*0.04 0.13*0.12* 0.11 0.03 0.02 0.10 0.10 Job autonomy 0.09** 0.07** 0.08** 0.14 0.05*0.05 0.05 Task complexity Sq 0.06** 0.06* 0.12 0.01 0.01 Time pressure Sq 0.02 0.02 0.01 0.01 Contact with users Sq 0.08 0.08 0.08 0.08 Job autonomy Sq 0.02 0.01 R20.022 0.037 0.048 0.055 0.055 0.018 0.042 0.045 0.046 0.046 1R20.022** 0.019** 0.007** 0.007*0.000 0.018** 0.024** 0.002*0.001 0.000 *p<0.05, **p<0.01; βstandardized regression coefficients are displayed; M, model; Sq, squared; d, effect size for M5 coefficients; (n= 1498). TABLE 5 Hierarchical multiple regression analysis models predicting psychological and physical symptoms in horeca. Psychological symptoms Physical symptoms Variables M1 M2 M3 M4 M5 dM1 M2 M3 M4 M5 d Gender 0.04 0.05 0.05 0.06 0.06 0.08*0.09*0.09*0.09*0.09* 0.19 Age 0.07 0.08 0.08 0.09*0.09* 0.17 0.07 0.09*0.10*0.10*0.10* 0.18 Job tenure −0.01 −0.03 −0.03 −0.04 −0.04 0.07 0.07 0.07 0.07 0.07 Task complexity 0.21** 0.21** 0.22** 0.23** 0.43 0.13** 0.14** 0.14** 0.15** 0.27 Time pressure −0.01 −0.01 −0.03 −0.04 0.15** 0.14** 0.13** 0.13** 0.23 Contact with users −0.01 −0.01 0.18*0.18* 0.16 0.01 0.01 −0.01 −0.01 Job autonomy −0.01 −0.03 −0.04 −0.06 −0.07 −0.07 Task complexity Sq 0.18** 0.18** 0.36 0.08*0.08* 0.16 Time pressure Sq −0.04 −0.04 −0.04 −0.05 Contact with users Sq 0.22*0.21* 0.19 −0.02 −0.02 Job autonomy Sq 0.05 0.03 R20.006 0.050 0.050 0.093 0.095 0.019 0.074 0.077 0.085 0.085 1R20.006 0.044** 0.000 0.043** 0.002 0.019** 0.055** 0.003 0.007 0.001 *p<0.05, **p<0.01; βstandardized regression coefficients are displayed; M, model; Sq, squared; d, effect size for M5 coefficients; (n= 655). Frontiers in Psychology 09 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 16 Sanclemente et al. 10.3389/fpsyg.2022.950908 References Aittomäki, A., Lahelma, E., Roos, E., Leino-Arjas, P., and Martikainen, P. (2005). Gender differences in the association of age with physical workload and functioning. Occup. Environ. Med. 62, 95–100. doi: 10.1136/oem.2004.014035 Bakker, A. B., and Demerouti, E. (2007). The job demands-resources model: state of the art. J. Manag. Psychol. 22, 309–328. doi: 10.1108/02683940710733115 Bakker, A. B., and Demerouti, E. (2013). Job demands-resources model. J. Work Organ. Psychol. 29, 107–115. doi: 10.5093/tr2013a16 Bakker, A. B., Demerouti, E., and Schaufeli, W. (2003). Dual processes at work in a call centre: an application of the job demands-resources model. Eur. J. Work Organ. Psychol. 12, 393–417. doi: 10.1080/13594320344000165 Bandalos, D. L., and Finney, S. J. (2010). “Factor analysis: exploratory and confirmatory,” in Reviewer’s Guide to Quantitative Methods, eds G. R. Hancock and R. O. Mueller (Milton Park: Routledge), 98–122. doi: 10.4324/9781315755649-8 Binnewies, C., and Wörnlein, S. C. (2011). What makes a creative day? A diary study on the interplay between affect, job stressors, and job control. J. Organ. Behav. 32, 589–607. doi: 10.1002/job.731 Burke, R. J., and McKeen, C. A. (1994). “Work, career, and life experiences associated with different career patterns among managerial and professional women,” in Job Stress in a Changing Workforce: Investigating Gender, Diversity, and Family Issues, eds G. P. Keita and J. J. Hurrell (Washington, DC: American Psychological Association), 301–310. doi: 10.1037/10165-019 Campbell, D. J. (1988). Task complexity: a review and analysis. Acad. Manag. Rev. 13, 40–52. doi: 10.5465/amr.1988.4306775 Chung-Yan, G. A., and Butler, A. M. (2011). Proactive personality in the context of job complexity. Can. J. Behav. Sci. 43, 279–286. doi: 10.1037/a002 4501 Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd Edn. New York, NY: Erlbaum Associates. doi: 10.4324/9780203771587 Cohen, J., Cohen, P., West, S. G., and Aiken, L. S. (2003). Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences, 3rd Edn. New York, NY: Erlbaum Associates. doi: 10.4324/978020377 4441 De Dreu, C. K. (2006). When too little or too much hurts: evidence for a curvilinear relationship between task conflict and innovation in teams. J. Manag. 32, 83–107. doi: 10.1177/0149206305277795 Debusscher, J., Hofmans, J., and De Fruyt, F. (2014). The curvilinear relationship between state neuroticism and momentary task performance. PLoS One 9:e106989. doi: 10.1371/journal.pone.0106989 EU-OSHA. (2004). Framework Agreement on Work-Related Stress. Bilbao: Legislation. Eurofound (2014a). Working conditions and job quality: Comparing sectors in Europe. Dublin: Eurofound. Eurofound (2014b). Occupational Profiles in Working Conditions: Identification of Groups with Multiple Disadvantages. Dublin: Eurofound. Eurofound (2016). Sixth European Working Conditions Survey – Overview report. Belgium: Publications Office of the European Union. Eurofound (2017). European Working Conditions Survey, 2015. [data collection], 4th Edn. Colchester: UK Data Service, doi: 10.5255/UKDA-SN-8 098-4 Eurofound and EU-OSHA (2014). Psychosocial Risks in Europe: Prevalence and Strategies for Prevention. Belgium: Publications Office of the European Union. Ferris, G. R., Bowen, M. G., Treadway, D. C., Hochwarter, W. A., Hall, A. T., and Perrewe, P. L. (2006). “The assumed linearity of organizational phenomena: implications for occupational stress and well-being,” in Research in Occupational Stress and Well-Being, Vol. 5, eds P. L. Perrewe and D. C. Ganster (Amsterdam: Elsevier Science). Fisher, G. G., Matthews, R. A., and Gibbons, A. M. (2016). Developing and investigating the use of single-item measures in organizational research. J. Occup. Health Psychol. 21, 3–23. doi: 10.1037/a0039139 Ford, M. T., Cerasoli, C. P., Higgins, J. A., and Decesare, A. L. (2011). Relationships between psychological, physical, and behavioural health and work performance: a review and meta-analysis. Work Stress 25, 185–204. doi: 10.1002/ 14651858.CD006066.pub3 Hackman, J. R., and Oldham, G. R. (1975). Development of the job diagnostic survey. J. Appl. Psychol. 60, 159–170. doi: 10.1037/h0076546 Higgins, J. P., and Green, S. (2008). Cochrane Handbook for Systematic Reviews of Interventions: Cochrane Book Series. Hoboken, NJ: Wiley. doi: 10.1002/ 9780470712184 Instituto Andaluz de Prevención de Riesgos Laborales [IAPRL] (2012). II Encuesta Andaluza de Condiciones de Trabajo. Andalusia: Junta de Andalucía. Instituto Nacional de Estadística [INE] (2019). Mercado Laboral, Actividad, Ocupación Y Paro. Available online at: https://www.ine.es/dyngs/INEbase/ es/categoria.htm?c=Estadistica_P&cid=1254735976595 (accessed February 15, 2022). Instituto Nacional de Estadística [INE] (2021). España en Cifras 2021. Available online at: http://www.ine.es/prodyser/espa_cifras (accessed February 15, 2022). Kaiser, M. O. (1974). Kaiser-Meyer-Olkin measure for identity correlation matrix. J R. Stat. Soc. 52, 296–298. doi: 10.1097/MD.0000000000029209 Karanika-Murray, M., Antoniou, A. S., Michaelides, G., and Cox, T. (2009). Expanding the risk assessment methodology for work-related health: a technique for incorporating multivariate curvilinear effects. Work Stress 23, 99–119. doi: 10.1080/02678370903068520 Karasek, R. (1979). Job demands, job decision latitude, and mental strain: implications for job redesign. Adm. Sci. Q. 24, 285–308. doi: 10.2307/239 2498 Kelliher, C., Richardson, J., and Boiarintseva, G. (2019). All of work? All of life? Reconceptualising work-life balance for the 21st century. Hum. Resour. Manag. J. 29, 97–112. doi: 10.1111/1748-8583.12215 Kerr, P., Da Torre, M. B., Giguère, C. É, Lupien, S. J., and Juster, R. P. (2021). Occupational gender roles in relation to workplace stress, allostatic load, and mental health of psychiatric hospital workers. J. Psychosom. Res. 142:110352. doi: 10.1016/j.jpsychores.2020.110352 Kuder, G. F., and Richardson, M. W. (1937). The theory of the estimation of test reliability. Psychometrika 2, 151–160. doi: 10.1007/bf02288391 Lawson, K. J., Noblet, A. J., and Rodwell, J. J. (2009). Promoting employee wellbeing: the relevance of work characteristics and organizational justice. Health Promot. Int. 24, 223–233. doi: 10.1093/heapro/dap025 LePine, J. A., Podsakoff, N. P., and LePine, M. A. (2005). A meta-analytic test of the challenge stressor-hindrance stressor framework: an explanation for inconsistent relationships among stressors and performance. Acad. Manag. J. 48, 764–775. doi: 10.5465/amj.2005.18803921 Liu, P., and Li, Z. (2012). Task complexity: a review and conceptualization framework. Int. J. Ind. Ergon. 42, 553–568. doi: 10.1016/j.ergon.2012.09.001 Lockwood, N. R. (2003). Work/life balance: challenges and solutions. Soc. Hum. Resour. Manag. 2, 1–11. Maltarich, M. A., Nyberg, A. J., and Reilly, G. (2010). A conceptual and empirical analysis of the cognitive ability-voluntary turnover relationship. J. Appl. Psychol. 95, 1058–1070. doi: 10.1037/a0020331 Moncada, S., Llorens, C., Navarro, A., and Kristensen, T. S. (2005). ISTAS21: versión en lengua castellana del cuestionario psicosocial de Copenhague (COPSOQ). Arch. Prev. Riesgos Labor. 8, 18–29. Noblet, A., Rodwell, J., and Allisey, A. (2009). Job stress in the law enforcement sector: comparing the linear, non-linear and interaction effects of working conditions. Stress Health 25, 111–120. doi: 10.1002/smi.1227 Nunally, J. C. (1978). Psychometric Theory. New York, NY: McGraw Hill. Parkes, K. R. (2016). Age and work environment characteristics in relation to sleep: additive, interactive and curvilinear effects. Appl. Ergon. 54, 41–50. doi: 10.1016/j.apergo.2015.11.009 Pató, B. S. G. (2015). The 3D job description. J. Manag. Dev. 34, 406–420. doi: 10.1108/JMD-11-2013-0151 Pató, B. S. G. (2017). Formal options for job descriptions: theory meets practice. J. Manag. Dev. 36, 1008–1028. doi: 10.1108/JMD-01-2016-0019 Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., and Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. J. Appl. Psychol. 88, 879–903. doi: 10.1037/0021- 9010.88.5.879 Reis, D., Hoppe, A., Arndt, C., and Lischetzke, T. (2017). Time pressure with state vigour and state absorption: are they non-linearly related? Eur. J. Work Organ. Psychol. 26, 94–106. doi: 10.1080/1359432X.2016.122 4232 Frontiers in Psychology 16 frontiersin.org fpsyg-13-950908 September 23, 2022 Time: 14:30 # 17 Sanclemente et al. 10.3389/fpsyg.2022.950908 Rydstedt, L., Ferrie, J., and Head, J. (2006). Is there support for curvilinear relationships between psychosocial work characteristics and mental well-being? Cross-sectional and long-term data from the Whitehall II study. Work Stress 20, 6–20. doi: 10.1080/02678370600668119 Sainani, K. L. (2012). Dealing with non-normal Data. Phys. Med. Rehabil. 4, 1001–1005. doi: 10.1016/j.pmrj.2012.10.013 Sanclemente, F. J., Gamero, N., Arenas, A., and Medina, F. J. (2020). Analysis of job satisfaction, job insecurity and job health in the employed population during the recent economic global crisis. UB J. Psychol. 50, 149–159. doi: 10.1344/ anpsic2020.50.14 Sawang, S. (2012). Is there an inverted U-shaped relationship between job demands and work engagement: the moderating role of social support? Int. J. Manpow. 33, 178–186. doi: 10.1108/01437721211225426 Schaufeli, W. B., and Taris, T. W. (2014). “A critical review of the job demandsresources model: Implications for improving work and health,” in Bridging Occupational, Organizational and Public Health, eds G. F. Bauer and O. Hämmig (Berlin: Springer). 43–68. doi: 10.1007/978-94-007-5640-3_4 Schmitt, A., Ohly, S., and Kleespies, N. (2015). Time pressure promotes work engagement: test of illegitimate tasks as boundary condition. J. Pers. Psychol. 14, 28–36. doi: 10.1027/1866-5888/a000119 Sheng, X., Wang, Y., Hong, W., Zhu, Z., and Zhang, X. (2019). The curvilinear relationship between daily time pressure and work engagement: the role of psychological capital and sleep. Int. J. Stress Manag. 26, 25–35. doi: 10.1037/ str0000085 Szczygieł, D., and Bazi´ nska, R. (2013). Emotional intelligence as a moderator in the relationship between negative emotions and emotional exhaustion among employees in service sector occupations. Pol. Psychol. Bull. 44, 201–212. doi: 10. 2478/ppb-2013-0023 Tadi´ c, M., Bakker, A. B., and Oerlemans, W. M. (2015). Challenge versus hindrance job demands and well-being: a diary study on the moderating role of job resources. J. Occup. Organ. Psychol. 88, 702–725. doi: 10.1111/joop.1 2094 Tanskanen, J., Taipale, S., and Anttila, T. (2016). Revealing hidden curvilinear relations between work engagement and its predictors: demonstrating the added value of generalized additive model (GAM). J. Happiness Stud. 17, 367–387. doi: 10.1007/s10902-014-9599-z Väänänen, A., Toppinen-Tanner, S., Kalimo, R., Mutanen, P., Vahtera, J., and Peiró, J. M. (2003). Job characteristics, physical and psychological symptoms, and social support as antecedents of sickness absence among men and women in the private industrial sector. Soc. Sci. Med. 57, 807–824. doi: 10.1016/S0277-9536(02) 00450-1 Van den Broeck, A., De Cuyper, N., De Witte, H., and Vansteenkiste, M. (2010). Not all job demands are equal: differentiating job hindrances and job challenges in the Job Demands-Resources model. Eur. J. Work Organ. Psychol. 19, 735–759. doi: 10.1080/13594320903223839 Van Ruysseveldt, J., and Van Dijke, M. (2011). When are workload and workplace learning opportunities related in a curvilinear manner? the moderating role of autonomy. J. Vocat. Behav. 79, 470–483. doi: 10.1016/j.jvb.2011.0 3.003 Warr, P. (2007). Work, Happiness and Unhappiness. Mahwah, NJ: Lawrence Erlbaum Associates. doi: 10.4324/9780203936856 Warr, P. B. (1987). Work, Unemployment and Mental Health. Oxford: Oxford University Press. Warr, P. B. (1994). A conceptual framework for the study of work and mental health. Work Stress 8, 84–97. doi: 10.1080/02678379408259982 Warren, N., Dussetschleger, J., Punnett, L., and Cherniack, M. G. (2015). Musculoskeletal disorder symptoms in correction officers: why do they increase rapidly with job tenure? Hum. Fact. 57, 262–275. doi: 10.1177/001872081454 7569 Žutautien˙ e, R., Radišauskas, R., Kaliniene, G., and Ustinaviciene, R. (2020). The prevalence of burnout and its associations with psychosocial work environment among Kaunas Region (Lithuania) Hospitals’ Physicians. Int. J. Environ. Res. Public Health 17:3739. doi: 10.3390/ijerph17103739 Frontiers in Psychology 17 frontiersin.org