Stereotype content of occupational groups in Germany. Maria-Therese Friehs, Felicia Aparicio Lukassowitz, Ulrich Wagner. Journal of Applied Social Psychology, March 31 2022. https://doi.org/10.1111/jasp.12872
Abstract: The stereotype content model (SCM) is a prominent model of social perception proposing two dimensions of evaluation: Warmth and competence. Occupational stereotypes have rarely been assessed in this model, especially in the German context, albeit their important impact on how individuals experience gainful occupation and navigate everyday social interactions. Responding to recent methodological critiques regarding the SCM's scale performance, we developed a context-adapted, well-performing German-language SCM scale and assessed warmth and competence ratings of 13 occupational groups in a heterogeneous sample (N = 425). Using the alignment optimization procedure to allow for more reliable latent mean value comparisons, we found occupational stereotypes to differ substantially on both dimensions, with Firefighters presenting the most favorable and Politicians and Unemployed people showing the least favorable evaluations. We discuss our findings in terms of their content-wise and methodological meaning as well as their implications for research and in occupational contexts.
5 DISCUSSION
In this study, we pursued the double goals of developing a scale to measure perceived warmth and competence, the two fundamental dimensions of social perception as defined by the SCM (Fiske et al., 2002), and employing it to describe current occupational stereotypes in Germany. Using an online survey in a heterogeneous adult sample and applying the state-of-the-art alignment optimization procedure to compare latent warmth and competence means, we found substantial differences between the perception of the 13 occupational groups included in the survey. The results as well as their implications will be discussed in the following.
5.1 Development of a stereotype content scale to assess occupational stereotypes
One goal of this article was to develop and apply a stereotype content scale that assesses the social perception of occupational groups with adequate reliability, dimensionality, and comparability. This was necessary as the functionality of established German and English scales has been challenged due to its unclear dimensionality and because preconditions for (latent) mean value comparison were often not given (Friehs et al., 2022; Kotzur et al., 2019, 2020). Though initial and exemplary steps have been taken to remedy this issue (Halkias & Diamantopoulos, 2020), existing well-performing SCM scales have not proven their applicability for assessing human targets and social groups, but rather focused on products or countries. Thus, we carefully selected indicators suitable for the context of assessing occupational stereotypes and applied a comprehensive factor-analytical scale development procedure. As a result, we can present a scale with a well-defined dimensionality, good model fit for all occupational groups we assessed, and acceptable internal consistency. We hope that this scale will assist in producing more structurally valid SCM findings and will provide options for cumulative research on stereotype content by using identical scales in future research.
Naturally, this is only the first application of the newly developed scale, and as scale construction and validation are ongoing processes (Flake et al., 2017), future applications and continuous careful examinations of the SCM scale are needed. In a minor internal replication study again assessing occupational stereotypes, we found independent support for our proposed SCM scale (see Supporting Information Material L). Nonetheless, future research should apply the scale to occupation-unrelated groups (e.g., social groups defined by gender, origin, or other features, or experimental conditions) to critically evaluate its applicability. For all further applications, we call for a careful examination of the measurement models using confirmatory factor analysis. Moreover, the used alignment optimization procedure might have led to differing results compared to other more traditional or conservative procedures of comparing latent mean values, like MGCFA (Friehs et al., 2022; Kotzur et al., 2020). Indeed, the data of Kotzur et al. (2019) showed somewhat diverging results depending on the method of analysis (alignment optimization in Kotzur et al., 2019; MGCFA in Friehs et al., 2022), and this finding has also been reported elsewhere (Magraw-Mickelson et al., 2020; Seddig et al., 2020).
Allowing for a residual covariation between the two competence indicators, as we did in this study, is new to (factor-analysis-based) SCM research. Although the procedure is in line with theoretical considerations about subdimensions of warmth and competence (Abele et al., 2021; Stanciu, 2015), it somewhat hinders applications of the scale to analyze observed means, because only advanced modeling approaches can account for this residual covariation. The residual covariation could also have contributed to the relatively low internal consistencies of the scale. Additional contributing factors could be the relatively low number of indicators per scale and the fact that we reported McDonald's ω values, which—unlike Cronbach's α—do not assume τ-equivalence of indicators (Hayes & Coutts, 2020). ω values have very rarely been reported in SCM research (but see Friehs et al., 2022; Kotzur et al., 2020), and indeed, if we had computed αs, the average internal consistencies would have been substantially higher. We do not believe that the internal consistency disqualifies the further usage of our scales, as we have presented plentiful additional evidence of the scales’ dimensionality and structural validity (Flake et al., 2017). Interestingly, we found that the internal consistency of the warmth scale was lower than that of the competence scale, which is in line with recent findings that warmth is assessed more idiosyncratic and less consistent across raters (Koch et al., 2020).
5.2 Occupational stereotypes in Germany
Another goal was to describe occupational stereotypes in Germany. Previous research applying the SCM or related constructs provided ample empirical evidence to predict occupational stereotypes for some groups. As such, we were able to confirm our assumptions concerning both the warmth and competence assessments of Firefighters and Police officers, the warmth prediction of Bankers, Child care workers, and Politicians, as well as the competence expectations concerning Unemployed people, Physicians, and Teachers. Nonetheless, some of our hypotheses were contradicted outright, such as the high warmth rating of Teachers or the high competence perceptions of Politicians, or deviated slightly from our expectations, for instance for Judges, Farmers, and Craftspeople. We will not offer a detailed discussion of all findings, but in the following, we will outline some select and unexpected findings.
Based on the results reported in the literature, we expected both Physicians (Asbrock, 2010; Gesellschaft für Konsum-, Markt-, und Absatzforschung e.V., 2018; He et al., 2019; Imhoff et al., 2013) and Teachers (Eckes, 2002; Gesellschaft für Konsum-, Markt-, und Absatzforschung e.V., 2018; He et al., 2019; Imhoff et al., 2013) to be perceived as highly warm. What is more, we would have assumed contextual circumstances (i.e., the global COVID-19 pandemic during the time of data collection; see below) to reinforce this positive warmth assessment due to increased public salience and appreciation of the contributions these occupational groups make to societal functioning. However, we found both occupational groups to score medium on warmth, with at least two occupational groups showing significantly higher warmth ratings. Similar findings were recently reported in the Swedish context (Strinić et al., 2021). Our data cannot provide explanatory information for this deviation from theory, so these issues might be worth investigating in future research.
We are also surprised by the prominently negative occupational stereotypes of Politicians, which were rated lowest on warmth and second-lowest on competence. Though previous international literature would have led us to expect higher competence ratings (Fiske & Durante, 2014; He et al., 2019), these findings are not singular in the German context (e.g., forsa Politik-und Sozialforschung GmbH, 2019; Gesellschaft für Konsum-, Markt-, und Absatzforschung e.V., 2018; Wagner et al., 2020) and consistent with results focusing on other information sources, such as the screening of occupational groups mentioned frequently and negatively on the Internet (Gesellschaft für Konsum-, Markt-, und Absatzforschung e.V., 2018). Nonetheless, they give rise to substantial societal concerns: Cuddy et al. (2007) proposed that warmth and competence stereotypes are predictive of emotional and behavioral responses. Consequently, the negative occupational stereotypes of politicians might in part be responsible for current political issues, such as the rise of right-wing populist parties, which proclaim their distinctness from established politicians and vote for fundamental changes in the political system, or the recent reports of hate mail threatening the lives of various politicians. In the long run, these negative perceptions of politicians might impair the functioning of the democratic system through a loss of interest in and disengagement from political parties and initiatives, reduced voter participations, and support for nondemocratic movements and ambitions.
Finally, the SCM usually predicts frequent observations of ambivalent stereotypes (i.e., high ratings on one dimension paired with low ratings on the other; Abele et al., 2021; Cuddy et al., 2009; Durante et al., 2013, 2017; Fiske, 2015; Fiske et al., 2002). In contrast, our findings showed overall strong and significant positive correlations between warmth and competence both within and between occupational groups. This led to a distinct absence of ambivalently rated occupational groups. Importantly, this pattern is not indicative of a unidimensional stereotype scale, as the EFA results indicated (at least) bidimensional solutions in all occupational groups except one. Naturally, from a statistical perspective, the across-group relation between the warmth and competence dimensions is highly dependent on the selection of occupational groups. Thus, our finding might just be explained by a tendency to select nonambivalently stereotyped occupational groups for assessment (Fiske et al., 2002). However, comparable findings have been reported elsewhere (Durante et al., 2013, 2017; Kervyn et al., 2015). One explanation might be the fact that for both warmth and competence, it is assumed desirable to be rated highly, and therefore these dimensions correlate positively with general evaluations (Kervyn et al., 2013; Osgood et al., 1957). In fact, Sayans-Jiménez et al. (2017) found the support of a bifactor model of stereotype content featuring both the SCM dimensions and an independent global evaluation factor. On the other hand, high correlations between warmth and competence factors within and across occupational groups could be indicative of acquiescence or halo-effects (Judd et al., 2005). The lack of ambivalently evaluated groups is also in line with the findings of Durante et al. (2013, 2017), which predict little ambivalent stereotypes in societies with relatively low inequality and low conflict, such as Germany.
5.3 Relevance of the research results
Our findings might be applied in the investigation of social interactions and processes in specific work contexts: Oftentimes, workplaces are characterized by the intimate collaboration of differently stereotyped occupational groups (e.g., nurses and physicians in hospitals, teachers and child care workers in schools). Employees holding lower-status positions may be stereotyped as less competent and, therefore, passed over, so that information exchange and collaboration are disturbed (Koch et al., 2021). Reversely, employees in high-status positions stereotyped as highly competent might not be informed about smaller issues and problems due to strong perceptions of hierarchy, which might lead to “blind spots” and impaired decision-making processes based on incomplete information (Tourish, 2005). Acknowledging the ways in which different staff members could potentially be biased by social perception processes is crucial for well-functioning teamwork, which is a necessity in most contemporary working environments.
As mentioned before, occupational stereotypes might also strengthen occupational segregation (i.e., the distribution of individuals from different demographic backgrounds across occupations; He et al., 2019). Groups such as women, physically or mentally impaired people, or those with low socioeconomic status might be underrepresented in occupations scoring high on competence (He et al., 2019), a circumstance by which occupational stereotyping is reinforced. Occupational segregation can be reduced by predicting the social groups that might be underrepresented in a particular job and specifically encouraging and promoting their access to that occupation (e.g., with nondiscriminatory job advertisements). By knowing about occupational stereotypes and intervening accordingly, future labor shortages might be prevented (He et al., 2019). Thus, our research might also be applied to define and examine strategies to change occupational stereotypes.
5.4 Future research directions
As discussed above, we acknowledge that our findings concerning the warmth and competence assessments of different occupational groups are relational and dependent on the specific other occupational groups we assessed. Future research comparing different occupational groups might thus come to somewhat different conclusions. What is more, our study contained only a small number of occupational groups (compared to other SCM research, e.g., Asbrock, 2010; Eckes, 2002; He et al., 2019), thus limiting the descriptive and comparative informational value. Most certainly, the number and choice of groups in our study does not reflect the full range of occupational groups relevant in any society. Nonetheless, we collected data from a heterogeneous sample, most of whom had no substantial prior experience with filling in online surveys. Thus, we needed to keep survey length and participant strain to a minimum (Halkias & Diamantopoulos, 2020). Further research may investigate the stereotypes associated with more or other occupational groups.
We would also like to draw attention to the potential influence of the context this study was conducted in, as this may influence the occupational stereotypes of some groups. During the data collection period, Germany just experienced a relaxation of severe restrictions of everyday life and personal freedom due to the first wave of the global COVID-19 pandemic as well as a decreasing number of severe medical treatments. This context might impact the evaluations of some occupational groups, such as Physicians and Hospital and elderly care nurses. Likewise, schools and nurseries were closed for the most part, and parents were forced to care for their children at home, which might affect the perception of Teachers and Child care workers. This period was also marked by a large number of short-term and extensive political decisions, mainly to stabilize Germany's economy and to provide more extended health care, potentially impacting the stereotype content of Politicians. Our data collection period also overlapped somewhat with the lamentable incident of George Floyd's death in the United States on May 25, 2020, which initiated a wave of protests and a fierce public debate about racism in the police force both in the United States and in Germany. Consequently, the social perception of Police officers might be influenced by these circumstances. Previous SCM research has not, to the best of our knowledge, focused on the impact of relevant external circumstances, nor on the change of occupational stereotypes over time. Thus, further research applying repeated cross-sectional or longitudinal surveys might help answer these questions.
Finally, future research could apply the assumption that warmth and competence perceptions are predictive of emotional and behavioral reactions toward the assessed occupational groups (Cuddy et al., 2007). Thus, on the base of the presented findings, future research could predict and investigate the affective and conative responses certain occupational group memberships elicit in professional interactions or societal discourses. This approach might be applied to a variety of contemporary problems, for instance the striking contrast between the highly positive social perceptions of professions in the child, hospital, or elderly care sector on the one hand, and their precarious working conditions and insufficient remuneration on the other hand (DGB Niedersachsen, 2020). Another application might lie in the investigation of reported phenomena of actively harming or hindering representatives of different occupational groups fulfilling their occupational role (e.g., attacking firefighters and paramedics in action). The SCM and related theories may be put to the test as a theoretical framework to describe and explain these phenomena.