The managerial role and psychosocial factors of job satisfaction: a cross-sectional study among Wittyfits users

Background Job satisfaction is an emerging indicator for measuring workers occupational well-being, however this has been poorly studied in those with managerial roles. Objective We aimed to explore job satisfaction in workers and to determine and assess psychosocial factors that may influence this relationship. Methods Data from Wittyfits users were collected between January 2018 and February 2020. Volunteers anonymously provided socio-demographic data and responses to questionnaires about their levels of job satisfaction and psychosocial feelings (ambiance, meaning, organization, recognition, values, work-life balance). Results 10,484 employees (40.9% of women) and 836 managers (33.9% of women), i.e. 11,320 workers with median age of 45 years and seniority of median 10 years of service, were included in the study. Job satisfaction of workers was higher in managers than employees (mean {+/-} SD 68.1 {+/-} 20.4 vs 57.8 {+/-} 24.2, p < .001), as were their feelings about ambiance (71.2 {+/-} 20.9 vs 66.1 {+/-} 24.2), meaning (66.9 {+/-} 21.8 vs 56.1 {+/-} 23.1), organization (55.3 {+/-} 23.6 vs 46.6 {+/-} 24.6), recognition (62.8 {+/-} 23.9 vs 48.3 {+/-} 26.7) and values (66.3 {+/-} 21.7 vs 56.5 {+/-} 23.1) (p < .001). There was no difference in work-life balance (58.1 {+/-} 23.9 vs 59.2 {+/-} 23.4, p = 0.2). All psychosocial factors had an impact on job satisfaction for both managers and employees (p < .001). High job satisfaction was more prevalent in workers who were managers than in employees (84.6 vs 68.8%, p < .001). Even though the managerial position was the most influential factor of job satisfaction (OR = 2.65, 95% CI 2.18 to 3.23, p < .001), other socio-demographic variables such as age (0.87, 0.79 to 0.95, p = 0.002) and seniority (0.71, 0.65 to 0.79, p < .001) also had an influence, although three times less. Conclusions Managers seem to have higher job satisfaction and psychosocial feelings about their work than employees. Psychosocial factors, many which are modifiable, as well as socio-demographic factors such as age and seniority, may influence job satisfaction among workers.

52 Key words: job satisfaction; psychosocial factors; manager; Wittyfit, occupational health 53 54 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 31, 2022.

Introduction
56 Job satisfaction is a widely studied concept, with positive benefits both for workers and 57 companies (1-5). It is a major and increasingly used indicator to assess occupational health of 58 different aspects of work, making it a global importance at the workplace (6). One possible 59 definition of job satisfaction is "a pleasurable or positive emotional state resulting from the 60 appraisal of one's job or job experiences" (7). In addition to common models used to assess 61 occupational well-being of workers (8-12), job satisfaction is an important attribute in 62 measuring quality of life and stress at work (1,6). Therefore, understanding the predictors of 63 job satisfaction allows for the development of effective strategies to improve occupational well-64 being and business performance (13). To the best of our knowledge, job satisfaction has been 65 poorly studied in the general population, and nearly no studies assessed the differences in job 66 satisfaction between managers and employees, or they focused only on healthcare professions 67 (14). As a crucial role within a company, it therefore seems necessary to investigate the 68 importance of managerial status on perceived job satisfaction.  78 Thus, the main objective of our study was to explore the influence of the job position (manager 79 or employee) on job satisfaction. The secondary objectives were to assess the other . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. 102 questionnaire using the visual analogical scale (VAS; score range from 0 to 100). Wittyfit uses 103 VAS to assess health outcomes in the same way as it does to measure stress, for example (22-104 24). Workers could estimate their job satisfaction as many times as they wanted. For each . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint 6 105 indicator (job satisfaction and psychosocial feelings), we used the average score to evaluate the 106 feeling of a worker. We considered a worker to be satisfied when their score was higher than 107 50 and dissatisfied lower than 50. 119 Similar to job satisfaction, all psychosocial indicators were evaluated using the associated VAS 120 in the questionnaire. Workers could estimate their feeling in each category as many times as 121 they wanted and their average score was retained. Wittyfit's clients provided the age, seniority, 122 sex and position (manager or employee) of their workers who use Wittyfit. Workers were then 123 divided into two age categories (over or under 45 years old) and two seniority classes (over or 124 under 10 years of service).

126 Statistics
127 Quantitative data (job satisfaction and psychosocial feelings) were expressed in terms of mean 128 ± standard-deviation. In absence of data, assumed to be missing completely at random (25), a 129 mean imputation by client was applied to avoid complete-cases analysis bias (26). Qualitative . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint 130 data (worker's job satisfaction/dissatisfaction) were expressed using number of participants and 131 associated percentage. First, comparisons between groups (e.g. by sex, by job position) were 132 performed with ANOVA for continuous variables (job satisfaction and psychosocial feelings) 133 and χ 2 tests for categorical variables (satisfied/dissatisfied workers). Then, associations between 134 endogenous variables (job satisfaction) and exogenous variables (psychosocial and socio-135 demographic factors) were assessed with linear and logistic mixed regressions respectively.
136 The company to which the worker belongs was considered a random effect. Variance influence 137 factor values of exogenous variables were assessed and revealed no serious multicollinearity (a 138 value below 4 was considered acceptable). Coefficients of linear regression models have been 139 expressed as effect sizes (ES) and confidence intervals (95% CI) using a "refit" method (27) 140 and interpreted according to Funder (28). Coefficients of logistic regression models are 141 expressed as standardized odds ratio (OR) and 95% CI and interpreted according to Chen (29). Managers also tended to be more satisfied with their jobs, 160 with a higher prevalence of managers highly satisfied with their jobs than employees (84.7 vs 161 68.8%, p < .001) (Fig 2). Managers were twice as likely to be more satisfied than employees 162 (OR = 2.65, 95% CI 2.18 to 3.23, p < .001) (Fig 3).

169
170 Results of the "job satisfaction" questionnaire 171 Overall, the psychosocial feelings were higher in managers than in employees (Fig 4).
177  (Fig 5). For managers, job satisfaction was mainly driven by . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint 10 207 years of seniority by 19% (0.71, 0.65 to 0.79, p < .001). Finally, both job satisfaction level (-208 0.02, -0.06 to 0.02, p = 0.97) and prevalence of workers with a high job satisfaction (0.97, 0.88 209 to 1.06, p = 0.5) did not appear to differ by sex (Fig 3). 246 Meaning at work is defined as "the discovery of existential meaning from work experience, 247 work itself and work purpose/goals" (43). Thus, it is not surprising to find a direct link between 248 meaning and job satisfaction as it translates among other things the motivation and the 249 fulfillment of the worker. Organization commitment, known for being related to job satisfaction 250 among nurses (13,44), was also found to be positively related to job satisfaction. Lastly, work 251 values were positively related to job satisfaction, which is consistent with the findings of 252 previous studies on the subject (45,46). It has also been highlighted among doctors that those 253 with higher work values scores presented higher job satisfaction (39). Recommendations should 254 be made for organizations to strive for improvement of workers' psychosocial feelings, to 255 consequently improve their job satisfaction.

256
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 31, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint 13 282 inclusion of a random effect in our regression models, namely the "company" effect, allowed 283 us to measure the influence of the company on the outcome independently of the fixed effects, 284 namely the psychosocial and socio-demographic factors including the influence of the position 285 held by the employee. Even taking these considerations into account, we believe 286 generalizability of the findings from our study is possible. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 31, 2022. ; https://doi.org/10.1101/2022.07.30.22278228 doi: medRxiv preprint