Employers often invest a substantial amount of time and money in scrutinizing job candidates for value fit, for example, through multiple interviews, before extending a job offer (Barrick & Parks-Leduc, 2019; Bermiss & McDonald, 2018; Cable & Judge, 1997; Rynes & Gerhart, 1990). Generally, employees also care about whether their values are congruent with those of the supervisor (person-supervisor [P-S] fit); group or team (person-group [P-G] fit); and the organization in general (person-organization [P-O] fit; Chatman, 1989, 1991). Indeed, greater perceived fit is associated with positive outcomes such as higher job satisfaction, performance, well-being (Kristof-Brown et al., 2005; Oh et al., 2014); pursuit of career development programs (Hentschel et al., 2021) and less job stress (Edwards & Rothbard, 1999; Edwards et al., 1998). According to the Attraction-Selection-Attrition framework (ASA; Schneider et al., 1995), employees with poor fit are expected either to become increasingly assimilated to the organization’s values trough socialization processes or to eventually leave the organization. Empirical research generally supports this framework, suggesting that individuals within an organization do tend to become more similar over time (De Cooman et al., 2009; Oh et al., 2018), and that employees who do not embrace the organization’s values face the risk of becoming unfairly treated, such as being passed over for promotion or excluded (Stone-Romero et al., 2003; Turco, 2010).

A growing body of literature has demonstrated adverse relationships connected to poor P-O fit, such as higher levels of turnover intentions, lower levels of job satisfaction, poorer social behavior, and higher levels of burnout (Akkaya & Serin, 2020; Coşkun et al., 2022; Maloba & Pillay-Naidoo, 2022; Vandevelde et al., 2020). Although much is known about the importance of P-O fit for employee retention, health, and thriving at work, the explorations of fit perceptions have typically focused on individual and organizational values that are related to, but conceptually different from, political ideology (Bermiss & McDonald, 2018). Whereas values are defined as guiding principles about which behaviors and end-states are desirable (Chatman, 1989; Schwartz, 1992), political ideology is defined as a set of interrelated ideas about the proper order of society and how it can be achieved (Erikson & Tedin, 2003). Unlike work-related values, which can be shaped by socialization processes within organizations (Edwards, 2008), political ideology tends to persist as individuals move between organizations (Block & Block, 2006), and seems to be activated relatively frequently in workplace contexts, for example, in casual conversations among employees (Bermiss & McDonald, 2018). As political ideology has become important as an identity marker and an arena for self-expression in modern secular democracies (Caprara & Vecchione, 2017), it is important to study how it is expressed in organizational settings.

In this paper, we propose that fit pertaining specifically to political ideology matters for several work-related attitudes such as turnover intentions, job satisfaction, burnout, and social relations in the workplace, and that an important mechanism is the willingness to self-censor one’s political views in the presence of others. By testing a model of how perceived political fit and self-censorship relate to work-related outcomes, the present study makes a novel contribution to the P-O fit literature, by adding knowledge about how, and in what way, political fit and political self-censorship relates to employee health, retention, and social relations in the workplace.

Literature review

Negative outcomes associated with poor political fit

Although research in this area is surprisingly scarce, political polarization and its consequences in organizational settings have received increased interest in recent years. For example, some studies have focused on how political ideology as a social identity can influence social dynamics at work (e.g., willingness to cooperate with coworkers; Sinclair et al., 2023; for a review see Swigart et al., 2020). There are also a handful of studies focusing on correlations between political identity dissimilarity and work-related attitudes. He et al. (2019) found that political identity dissimilarity predicted burnout, job satisfaction and turnover intentions, and that this was mediated by self-reported experiences of workplace incivility, in an American Mechanical Turk sample during the 2012 U.S. presidential election. Similarly, employees who felt that their political ideology fit with that of their supervisors have been found to experience less job-related stress and higher job satisfaction (Foley et al., 2018). Furthermore, Bermiss and McDonald (2018), who followed investment professionals in the U.S. private equity industry over time, found that ideological misfit between employee and firm (as measured through the proxy of monetary donations to a political party), predicted subsequent departure from the firm. However, none of these studies has investigated the role of self-censorship of political opinions at work.

Political self-censorship

Recent findings suggest that political ideology does indeed contribute to fit perceptions in organizations (Foley et al., 2018; He et al., 2019), and that it predicts turnover (Bermiss & McDonald, 2018). However, many individuals whose ideology differs from that of the majority may not be able to leave their job. These employees may feel stigmatized or anxious about expressing their views, leading them to put on a “façade of conformity” and hide their opinions or their political identity in order to fit in, which may negatively impact their well-being (Hewlin et al., 2017).

Self-censorship refers to intentionally withholding facts or opinions based on one’s perception of the climate of opinion (Hayes et al., 2005; Nets-Zehngut & Fuxman, 2017). Thus, self-censorship reflects individuals’ inclination to impose restrictions on themselves. It requires active assessment of the communication climate, which may include the probability of being embarrassed, hurting someone’s feelings, or starting an interpersonal conflict (Hayes et al., 2005). Put differently, because of some kind of expected cost or sanction associated with revealing a certain opinion, the person chooses to conceal it. Even though there can be positive consequences of self-censorship, such as enhanced group cohesiveness and solidarity as a result of prevented disagreements and controversies, there are also negative consequences to consider. Whereas a free flow of information and views allows for dynamic change of opinions and can stimulate the development of tolerance (Bar-Tal, 2017), a “culture of silence” may develop if people feel the need to screen out certain topics, opinions, and questions from discussions, preventing communication about important matters (Sunstein, 2003). Because self-censorship hinders critical views from surfacing, the suppression of disagreement may create a false impression of consensus, which might impair the quality and creativity of group decisions and thus be detrimental for organizations (Bar-Tal, 2017; Janis, 1997). From the employee perspective, the perceived pressure to conceal one’s personal values and monitor one’s behavior might eventually lead to elevated stress levels and a decrease in job satisfaction (Chou et al., 2020; Hewlin, 2003). In this study, we therefore look closer at whether political fit and self-censorship at work is related to job satisfaction, social belonging, burnout, and turnover intentions.

Self-censorship occurs in every social space, in interpersonal, as well as in intra-organizational and intra-societal settings (Bar-Tal et al., 2017), and there is evidence to suggest that it has become increasingly common. In US polls, agreement with the statement that one cannot “openly express one’s opinions because others might find them offensive in the current political climate”, has increased from thirteen percent in the 1950’s to between forty and sixty percent (Ekins, 2020). In Sweden, where we conducted the current study, the situation is similar, with over 50% of the population expressing agreement with this statement according to a large-scale nationwide survey (Persson & Widmalm, 2022). Moreover, the results of the survey revealed that the political climate in Sweden is characterized by a widespread self-censorship among sympathizers of all political parties: At least a third of sympathizers in all Parliament parties state that they fully or partially consider themselves unable to openly express their opinions because others may find them offensive.

At work, people are forced to get along with others whom they would perhaps not have chosen to socialize with in their spare time, and political ideologies may thus be more diverse in professional compared to private networks. Being an ideological “misfit” may make an employee feel uncomfortable and less valued. As demands for censorship, or for conversations to be regulated, have been occurring more frequently in recent years (Persson & Widmalm, 2022), it is understandable if some people would think twice about revealing their political opinions or affiliation at work. Indeed, qualitative data from interviews conducted with investment professionals in the U.S. (Bermiss & McDonald, 2018) indicate that ideological misfits often avoid voicing their political opinions at work, and that they attribute this to career concerns and an implicit risk-reward calculus that favors silence.

The current study

Human capital is a vital resource for organizations, and management of employees is a challenging task that many firms struggle to handle effectively (Coff, 1997). In the current study, we build on previous work that have studied political fit, by focusing on the role of self-censorship. Relying on a social identity framework (Tajfel, 1981), we reason that perceiving a poor fit between one’s own and the organizations’ political ideology, and feeling the need to self-censure one’s opinions, will correspond to a weaker sense of community at the workplace, and a higher fear of isolation. We further examine whether political fit and self-censorship are related to three outcomes that are key to stimulating the wellbeing and retention of employees: job satisfaction, burnout, and turnover intentions. Specifically, we hypothesized that willingness to self-censor political opinions at work, and poor perceived fit between one’s own political views and those of other employees and management, are negatively associated with job satisfaction (H1a; H2a), and with perceived social community at the workplace (H1b; H2b), and positively associated with burnout (H1c; H2c), turnover intentions (H1d; H2d), and fear of social isolation (H1e; H2e). A conceptual model of the hypothesized relations is presented in Fig. 1.

Fig. 1
figure 1

Conceptual model of the hypothesized relations between constructs. Expected directionality of the relationship (positive or negative) is indicated within parentheses

Full size image

Method

Participants and procedure

The sample size (N = 710) was determined by the availability of resources. A post hoc power analysis conducted in G*power 3.1.9 (two-tailed) with a conservative alpha of 0.005 suggested that this sample size yielded 80% power to detect a correlation of |ρ|= 0.136. Being currently employed and over 18 years old was a requisite to participate. No exclusion criteria were used in the sampling procedure. The participants were recruited at public places such as train stations and libraries, as well as on a university campus and various workplaces, with the aim of obtaining a diverse sample with regards to age, gender, and socioeconomic background. The gender distribution was 44.5% men, 54.6% women, 0.9% with another gender identity, and 12 participants who did not respond to the question (Mage = 40.4, SD = 13.8). Political self-placement was on average 4.91 (SD = 2.31) on a 10-point scale from left to right. Although this is slightly to the left (t[681] = -6.68, p < 0.001) of the theoretical midpoint of the scale (5.5), it indicates that we succeeded in recruiting a heterogeneous sample in terms of political affiliation. Whereas all participants had some form of employment, 85.2% reported working as their main occupation, 13.3% were (working) students, and 1.3% responded “other”. Further, type of occupation varied widely, with 47.4% of the sample reporting high-skilled and 45.4% low-skilled occupations (2.6% of responses were difficult to categorize, and 4.7% did not respond to the question). In terms of education, 3.7% had only completed elementary school, while 29.4% had completed upper secondary education, 16.1% were currently enrolled in university studies but had not received a degree, 34.1% had a bachelor’s degree, 14.8% had a master’s degree, and 1.3% a Ph.D. The sample thus contained more highly educated individuals compared to the general population. After being introduced to the study, the participants provided their informed consent and responded to the measures in the order that they are presented below, followed by demographic questions.

Measures

Willingness to Self-censor at Work

To capture willingness to self-censor political opinions at work, we adapted seven items from the Willingness to Self-Censor scale developed by Hayes et al. (2005). We modified the items so that they referred specifically to expressing one’s political views in a workplace context. We also included six self-constructed items concerning willingness to share political opinions or preferences at work (see Table 1). The goal was to construct simple, clear, and general items without double-barreled formulations, unnecessary negations, or ideological or culture-specific content. We further aimed for varied content and formulations to prevent methods factors from obscuring the factor structure. Following our pre-registration [https://osf.io/g9ry6/?view_only=b2b023ad0e9d4e03a12289391b54c5e7], we ran an exploratory factor analysis with maximum likelihood extraction and direct oblimin rotation (delta = 0). The results indicated that there were two distinct factors accounting for 38.4% and 15.9% of the variance in the responses respectively (the eigenvalues were 5.00 and 2.06). No other factor was above the inflexion point of the scree plot or had an eigenvalue larger than 1. As shown in Table 1, the seven items from the original Willingness to Self-Censor scale (Hayes et al., 2005) loaded strongly on the first factor, while four of the new items loaded strongly on the second factor. The last two items, which cross-loaded substantially were dropped, following the pre-registered plan. Based on these results, we calculated separate scores for general willingness to self-censor political opinions at work (α = 0.84, ωh = 0.80) and willingness to share political positions with others at the workplace (α = 0.84, ωh = 0.82). These two scale scores were used as separate factors in all subsequent analyses.

Table 1 Factor loadings based on the structure matrix
Full size table

Perceived political fit

Similar to He et al. (2019), we captured perceived fit between own and others’ political opinions with two self-constructed items: “Generally speaking, if you compare your own political opinions with those of your co-workers, to what extent do you think they are similar?” (1 = “extremely dissimilar”, 7 = “extremely similar”) and “Generally speaking, how well do you think that your political opinions align with those of your superiors?” (1 = “do not align at all”, 7 = “completely align”; α = 0.65). These were recoded so that a higher value reflects less perceived fit.

Social community

The measure of perceived social community at the workplace consisted of three items, taken from the Copenhagen Psychosocial Questionnaire (COPSOQ III; Burr et al., 2019), e.g., “Do you feel included in a community at your workplace”? (1 = “never/almost never”, 7 = “always”; α = 0.85).

Job satisfaction

We captured job satisfaction with the four-item measure from COPSOQ III. Specifically, the participants were asked how satisfied they were with their job in general, regarding their future prospects in the job, the physical work conditions, the way their knowledge/skills are utilized, and their job in total (1 = “very unsatisfied”, 7 = “very satisfied”; α = 0.84, ωh = 0.83).

Turnover intentions

Turnover intentions were captured with a one-item measure from COPSOQ III: “How often do you consider/contemplate applying for a new job?” (1 = "never/almost never”, 7 = “always”). Although it is not possible to estimate internal reliability for a one-item measure, a recent review found that the majority of single-item measures in the organizational sciences have showed strong psychometric properties, including moderate to high test–retest reliabilities and extensive criterion-related validity (Matthews et al., 2022). Single-item measures can therefore have acceptable properties, particularly when concerning unidimensional and homogenous constructs (Wanous et al., 1997).

Burnout

We measured indications of burnout with a three-item measure from COPSOQ III, asking participants to report how often they have felt a lack of endurance and energy, physically exhaustion, and emotional exhaustion during the previous four weeks (1 = “not at all”, 7 = “all the time”; α = 0.83).

Fear of social isolation

Fear of social isolation was captured with the five-item Fear of Social Isolation scale developed by Hayes et al. (2013), e.g., “One of the worst things that could happen to me is to be excluded by people I know.”; “It is important to me to fit into the group I am with.” (1 = “do not agree at all”, 7 = “completely agree”; α = 0.84, ωh = 0.76).

Statistical procedure

As the outcome measures are conceptually related to each other, we initially performed confirmatory factor analyses in AMOS 29.0 to examine their distinctness. Calculations based on the maximum likelihood method and a model with four correlated latent factors (social community, job satisfaction, fear of isolation, and burnout) and turnover intentions as a separate manifest variable that was allowed to correlate with the factors yielded adequate fit: χ2(95) = 331.6 (p < 0.001), CFI = 0.954, RMSEA = 0.060 (90% CI: [0.053, 0.067]), SRMR = 0.0421 (λ ≥ 0.60). Although the job satisfaction and social community factors correlated very strongly (r = 0.69), collapsing these two into a common factor would have reduced fit substantially, Δχ2(4) = 413.8 (p < 0.001), χ2(99) = 745.4 (p < 0.001), CFI = 0.876, RMSEA = 0.096 (90% CI: [0.090, 0.103]), SRMR = 0.0592. These results confirm the distinctness of all outcome variables.

For scales with at least four items, we report McDonald’s hierarchical omega coefficients in addition to Cronbach’s alpha as estimates of internal reliability, based on analyses that were run in the “psych” package in R (Revelle, 2024). The hierarchical omega coefficient represents the extent to which a general factor saturates the scale. It does not require tau equivalence (i.e., equal factor loadings and true score variances) and is therefore widely considered a superior measure of internal homogeneity in the modern psychometric literature (Dunn et al., 2014; Zinbarg et al., 2005).

In line with our pre-registration, we tested our hypotheses in terms of a series of correlational analyses followed by hierarchical regression analyses to probe the robustness of the hypothesized associations. In the first step of the regression analyses, self-censorship, political sharing, and perceived political fit were included as predictors; in the second step, political self-placement was added; and in the final step, demographic variables (education, age, and gender) were added. Dependent variables were fear of social isolation, burnout, job satisfaction, social community, and turnover intentions respectively. Separate regression models were run for each dependent variable. To manage the risk of Type 1 error, we set a conservative alpha threshold of 0.005. We also investigated the robustness of the hypothesized associations through structural equation modeling (SEM), which separates the systematic variation in constructs from measurement error, as the reliabilities of some scales were limited. These analyses were conducted in Jamovi 2.3 with the DWLS estimator, robust standard errors, and items as indicators. Self-censorship, political sharing, and perceived political fit were specified as predictors of fear of social isolation, burnout, job satisfaction, and social community. Model fit was adequate: Comparative Fit Index (CFI) = 0.98, the Root Mean Square Error of Approximation (RMSEA) = 0.066 (95% CI [0.062, 0.070]), Standardized Root Mean Square Residual (SRMR) = 0.058.

Thereafter, we ran a follow-up mediation analysis using the PROCESS macro version 4.1 in SPSS 29.0 (Hayes, 2022) to test whether the associations between perceived political fit and job satisfaction, social community, turnover intentions, and burnout were statistically mediated by willingness to self-censor. We report unstandardized bootstrapped confidence intervals for the indirect effect, but all variables were standardized prior to the analyses. These mediation analyses were, unlike the correlational and hierarchical regression analyses, not pre-registered.

Although there was a negative skew (-1.05) and one low extreme value on social community, results calculated without this extreme value and with non-parametric tests showed no indication that these deviations had any effect on the results. We therefore report results of parametric tests without exclusions. To provide more robust estimates, we report correlations bearing on our hypotheses with 95% bias-corrected bootstrap confidence intervals (10,000 resamples).

Results

Correlations between all variables are displayed in Table 2 and results from hierarchical regression analyses are displayed in Table 3. Consistent with the hypotheses, willingness to self-censor was strongly associated with higher fear of social isolation (r = 0.31[0.23, 0.37]), burnout (r = 0.32[0.25, 0.39], and turnover intentions (r = 0.27[0.20, 0.34]), and with lower job satisfaction (r = -0.34[-0.41, -0.27]) and social community (r = -0.39[-0.46, -0.32], p < 0.001), and these associations were highly robust when adjusting for other predictors (see Table 3). Political sharing was associated weakly to moderately strongly with higher job satisfaction (r = 0.15[0.08, 0.23]) and social community (r = 0.23[0.15, 0.31]) and with lower fear of social isolation (r = -0.13[-0.20, -0.05], p < 0.001) and turnover intentions (r = -0.12[-0.19, -0.04], p = 0.002), but not with burnout (r = -0.09[-0.16, -0.01], p = 0.022). Although these correlations are consistent with the hypotheses, none of the associations were significant when the effects of willingness to self-censor and perceived political fit between self and others at the workplace were adjusted for (see Table 3). Low perceived political fit between self and others was very robustly associated with lower job satisfaction and social community and higher turnover intentions with moderate to large effect sizes, consistent with the hypotheses. It was weakly associated with burnout as well, but this effect was marginal when adjusting for other variables, and it was not associated with fear of social isolation (although a suppression effect opposite to the hypothesized effect arose in the regression models); contrary to the hypotheses (see Table 2, 3).

Table 2 Correlations between all scales
Full size table
Table 3 Standardized regression coefficients (with 95% confidence intervals) based on linear hierarchical regression analyses
Full size table

In sum, hierarchical regression analyses suggested that the dependent variables burnout and fear of social isolation were uniquely and strongly predicted only by willingness to self-censor in support of H2c and H2e, while lack of job satisfaction, social community, and turnover intentions were uniquely and strongly predicted both by poor perceived political fit at work and willingness to self-censor, supporting H1a, H2a, H1b, H2b, H1d, and H2d. H1c and H1e were consequently not supported. In SEM-based robustness tests, which are reported in supplemental documents (Fig. A1), all dependent variables were, once again, strongly and uniquely predicted by willingness to self-censor (|β|≥ 0.40, p < 0.001). In these more sensitive analyses, perceived political fit at work did uniquely predict social isolation (β ≥ 0.18), as well as social community and job satisfaction (|β|≥ 0.28, p < 0.001), but it still failed to predict burnout (β = -0.09, p = 0.11).

A series of follow-up mediation analyses indicated that the associations between low perceived political fit and job satisfaction (-0.056[-0.086, -0.029]), social community (-0.063[-0.097, -0.034]), turnover intentions (0.042[0.024, 0.067]), and burnout (0.057[0.030, 0.087]) were all statistically mediated by willingness to self-censor. We also note that the willingness to self-censor was slightly higher on average among those who were politically on the left, younger, less educated, and female, and perceived political fit was lower among the less educated (see Table 2). Because the sample was overall more highly educated than the general population and slightly left of center, we ran subgroup analyses with low educated and right-wing participants as a robustness check. Among the low educated, all of the hypothesized associations remained significant (p ≤ 0.025, |r|> 0.12). Among right-wingers, the association between political sharing and turnover intentions was not significant (r = 0.08, p = 0.20) but the other hypothesized associations were (p ≤ 0.038, |r|> 0.13).

Discussion

In some workplace situations, employees may choose to withhold personal information to protect their privacy (Allen et al., 2007). Furthermore, for some individuals, a low level of self-disclosure may simply be part of their personality (Spencer-Oatey, 2013). At the same time, humans seem to have a fundamental need to share and disclose their thoughts, feelings, information, and personal opinions (Bar-Tal, 2017). Sharing troubling events with others assists individuals in understanding them and creating a sense of meaningfulness, serves as a coping strategy, and adds to a positive sense of self-worth (Harber & Cohen, 2005). This might explain why self-censorship at work is associated with negative consequences (Bar-Tal, 2017). Indeed, our results showed that self-censorship at work is strongly and robustly associated with higher fear of social isolation, burnout, and turnover intentions, as well as lower job satisfaction and sense of social community at work. At the same time, willingness to share general political preferences with colleagues was not robustly and independently associated with these dependent variables over and above other predictors. A possible explanation for this is that being comfortable sharing political preferences is more of a personality orientation, which becomes problematic only insofar as there is a felt pressure to share opinions and a concomitant self-censorship.

The results further revealed that poorer perceived political fit between self and others at the workplace was robustly associated with lower job satisfaction and social community, and higher turnover intentions, with moderate to large effect sizes (although it was weakly associated with burnout and, unexpectedly, unrelated to fear of social isolation). These results thus add to previous research that suggests that perceived dissimilarity in political orientations can have negative implications for stress levels and for job satisfaction (Foley et al., 2018; He et al., 2019), as well as for interpersonal workplace interactions (Roth et al., 2017; Sinclair et al., 2023). In line with previous research (Bermiss & McDonald, 2018), our findings further suggest that political misfits might progress beyond mere unease and are likely to become increasingly inclined to leave their organization.

Finally, the associations between perceived political fit and the dependent variables were all statistically mediated by willingness to self-censor, suggesting that the perceived pressure to hide one’s opinions explains the negative effects of poor political fit. Presumably, this might be explained by feelings of alienation or a lack of authenticity (Hewlin et al., 2017).

The attitudes of individual employees are influenced by the norms, beliefs, values, and attitudes of their group under certain conditions and in the particular context in which people live and work. This includes the political, economic, and cultural characteristics of society (e.g., general tolerance levels; Bar-Tal, 2017), as well as the organization’s culture and climate. The tolerance level and the norms pertaining to pluralistic diversity in the organization are therefore likely to be important factors that determine whether employees feel the need to self-censor or not. Indeed, ideological misfits are more likely to encounter a hostile climate in organizations with strong values (Duarte et al., 2015; Tilcsik, 2011).

As prejudice against political opponents often comes from both the political right and left (e.g., Sinclair et al., 2022, 2023), it is likely that persons leaning to the left would feel the need to self-censor in occupations where right-voting individuals are in majority, while the reverse would happen in occupations dominated by left-leaning people. In addition, the role and the status of the employee should be of importance, as this is likely to influence personal expectations about a damaged reputation and future sanctions (Bar-Tal, 2017). Indeed, this might explain why younger people with lower education levels tended to self-censor to a higher extent, as they may have more to lose from voicing deviant opinions at the workplace. As for the gender difference in reported levels of self-censorship, this might be explained by women being more likely to take on a relational approach to work and to engage in more interdependence-oriented behaviors in the workplace (Matthew et al., 2013), making them more likely to self-censor opinions that deviate from the norm.

Theoretical contributions

To the best of our knowledge, this is the first study to develop measures that assess political self-censorship as well as political sharing in the workplace, and investigate their relationship with fear of social isolation, job satisfaction, burnout, and perceived social community. By doing so, we build on and extend previous research, which has demonstrated relationships between perceived ideological dissimilarity and outcomes such as job satisfaction, burnout, and turnover intentions (He et al., 2019). In this study, self-censorship emerged as a stronger predictor of all outcomes (with the exception of turnover intentions). By identifying a possible mechanism, self-censorship, in the relationship between perceived fit and work-related outcomes, we extend previous findings and contribute to the P-O fit literature by providing novel information about when and why political ideology may influence vocational behavior.

Taken together, the findings are consistent with a social identity and social categorization perspective (Turner et al., 1987) on political affiliation. These perspectives would suggest that political affiliation could become a basis for the development of in‐ and out‐groups at work, which could result in feelings of alienation among political minority employees (He et al., 2019) and threaten their sense of belonging (Baumeister & Leary, 1995). Indeed, it is common for employees to spend a substantial amount of their time cooperating with co-workers to reach organizational goals, and the workplace often serves to fulfill essential social needs (Scott & Thau, 2013). Threatened belonging can thus lead to reduced social support (Beehr et al., 2000) and relational attachment (Ehrhardt & Ragins, 2019), which may explain the observed associations between political fit and self-censorship and outcomes such as job satisfaction, social community, turnover intentions, and burnout. Indeed, this would be consistent with the well-known effects of social support as an essential resource for preventing stress-related problems such as burnout (Bakker et al., 2023). Not only does such support protect employees from harmful consequences of stressful experiences (Cohen & Wills, 1985); in addition, it can be functional in achieving work goals, as instrumental support from coworkers can help to get the work done and thus alleviate the straining impact of work overload (Van der Doef & Maes, 1999).

However, political fit was not related to fear of social isolation, suggesting that fear of isolation may not necessarily arise because of poor political fit, as long as the employee can self-regulate by self-censoring their opinions. Hence, it is possible that self-censoring can serve as an adaptive mechanism to protect the sense of belonging, yet deplete individual resources leading to strain, and subsequently result in higher ratings of burnout symptoms and intentions to quit.

Practical implications

The results suggested that perceived political dissimilarity and political self-censoring in the workplace may be of importance for individuals’ well-being and work-related attitudes, and these findings have implications for practice. On one hand, organizations may strive to be apolitical or hold a neutral stance to political ideologies, when such values are not clearly work-related. This requires employees to self-regulate and adapt by refraining from political expression in the workplace, which could be positive for organizational processes, but may come with psychological costs. These costs might be particularly pronounced in the presence of strong informal political norms that a minority of employees do not conform to. On the other hand, a strategy of promoting openness, political plurality and free political expression in the workplace, may not be unproblematic either. Although it has been suggested that political viewpoints should be included in diversity initiatives within organizations to increase a sense of inclusivity (He et al., 2019), it is important to note that some ideologies or political views could be perceived as controversial, as standing in contrast with the organization’s values, or even as harmful to certain groups. The perceived inclusivity of one group may then come at the expense of another. The organization could consequently find itself caught between a rock and a hard place, where either advocating political neutrality or actively encouraging political inclusivity both would have consequences for their employees, albeit in different ways.

Highlighting this complexity, Swigart et al. (2020) suggested that value-based diversity may result in higher quality of work output, but that there also is a risk of it evoking conflicts that could impede performance. It may therefore be wise to keep in mind that a certain amount of self-censorship can serve as a positive competency that employees and leaders develop to make sure they adhere to the expectations of their role and norms for appropriate workplace behavior. Nevertheless, although these adaptive competencies may benefit the organization and work processes, they can come with psychological costs for individual employees. We do not aim to resolve how organizations should navigate these complex issues; rather, we call for a dialogue within organizations about the implications of political expression and censorship in the workplace.

Limitations

The results confirmed that the included instruments displayed the expected factor structure as well as high internal consistency, although the new scales should be tested further. Although we relied on self-reports, employees’ own reports should be considered valid indicators of their perceived work environment (Chan, 2009). It is possible that people’s perceptions of the political identities of their coworkers and superiors are less accurate; nevertheless, it is the perceived rather than actual fit that is of primary interest here, as it is the former that is expected to have effects on employee attitudes.

An important limitation to this study is that the design does not permit causal inferences. Although it is plausible based on previous research that self-censorship has negative consequences for the well-being and retention of employees, there are possible alternative explanations. For example, it may be that individuals who are experiencing increased levels of stress and discomfort at work tend to become less engaged, to withdraw from discussions, and to feel more alienated. It may also be that a neurotic and rejection-sensitive personality makes people likely both to censor themselves and to feel alienated, unsatisfied, and burnt out at work. Besides adding more control variables when examining the influence of political self-censorship on work-related outcomes, future research could test interventions designed to promote an inclusive and tolerant atmosphere, high diversity of perspectives, and low self-censorship. Longitudinal studies could also be helpful in order to test whether increasingly feeling like a misfit and beginning to self-censor as a result of naturally occurring changes prospectively predicts feelings of isolation, dissatisfaction, and burnout, as well as turnover intentions rather than the other way around. Currently, studies with such designs that permit causal inferences are rare in the literature.

There may also be some limitations associated with the sample. Although we succeeded in obtaining diversity with regards to occupation, age, gender, and political orientation, our sample of convenience may not be nationally representative in terms of political ideology and identity. This means that although our study shows that there are effects, a probability sample would be a better pointer to the precise size of the effects in the general population.

Finally, his study was conducted in Sweden, which is a Western-European post-industrial democracy with high levels of secularism, liberalism, egalitarianism, and support for democratic principles (Inglehart & Welzel, 2010). Further, Sweden has a multiparty system, with several distinct groups of voters with different ideological worldviews on the left and the right (Nilsson et al., 2020). It is therefore important to keep in mind that the results obtained here may not generalize equally well to all cultures. Political ideology may be particularly important as an identity marker in a modern secular democracy such as Sweden, where politics has become an arena for self-expression while religion has receding importance (Caprara & Vecchione, 2017). Furthermore, P-O fit tends to have stronger relationships with attitudes and turnover intentions in countries that are relatively more individualistic than collectivistic (Oh et al., 2014; Treviño et al., 2020).

To strengthen external validity, future research should examine the influence of political fit and self-censorship in other populations and sub-groups within these populations, as well as under what circumstances political self-censorship is more likely to occur. For example, it could be particularly interesting to explore whether political self-censorship vary depending on the type of organization, comparing the public and private sector. As the public sector is more susceptible to political governance, taking directives and enacting decisions made by political bodies, self-censorship may have a particular impact in this sector.

Conclusions

Non-political organizations may find themselves in tricky positions when they try to navigate in an increasingly politically polarized landscape. Indeed, employees bring their political ideologies to work, and associated values and identities are likely to impact their attitudes, judgments, decisions, and social interactions (Swigart et al., 2020). The current research contributes to the literature on person-organization fit by focusing on the role of political ideology – an often neglected, yet potentially influential feature of modern organizations. Even though there are of course many other core determinants of job satisfaction, turnover, burnout, and belonging, the current research point to political fit and self-censorship as factors within organizations that seem to have significant effects on employee well-being and work-related attitudes. For ideological misfits, silence may precede departure, suggesting the need for organizations to work proactively in order to create a tolerant climate.