Danish University Colleges
Stereotypes in Context: How and When Do Street-Level Bureaucrats Use Class Stereotypes?
Harrits, Gitte Sommer
Published in:
Public Administration Review
DOI:
https://doi.org/10.1111/puar.12952
Publication date:
2019
Document Version Peer reviewed version Link to publication
Citation for pulished version (APA):
Harrits, G. S. (2019). Stereotypes in Context: How and When Do Street-Level Bureaucrats Use Class Stereotypes? Public Administration Review, 79(1), 93-103. https://doi.org/10.1111/puar.12952
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal
Stereotypes in Context:
How and When Do Street-Level Bureaucrats Use Class Stereotypes?
By Gitte Sommer Harrits
Abstract
Stereotype use in street-level work and discretionary judgements is important due to the possible introduction of bias. This study contributes to the study of stereotypes in street-level work by suggesting that stereotypes are related to both cognitive and social uncertainties. A distinction between uncontrolled stereotype activation in the interpretation of information, and controlled stereotype application in decision making is presented to allow for a more nuanced study. Also, a unique design, combining semi-structured interviews and a vignette experiment is presented to accommodate this theoretical framework. Results show that stereotype activation is conditioned by class difference and involves both categorization and simplified assumptions. Stereotype application is conditioned by class as well as by homogeneity of the social context of street-level institutions. These results suggest that to decrease stereotypical bias in frontline encounters, one solution may be to increase social heterogeneity.
Practitioner points
• Street-level work can benefit from becoming more reflective about the use of stereotypes.
• To understand how stereotypes work it is helpful to distinguish between stereotype activation, as an uncontrolled reaction due to deep-seated cultural learnings, and
stereotype application, as the more controlled use of stereotypes when weighing different types of information in decision making.
• Both stereotype activation and application are conditioned by social class difference between street-level bureaucrats and clients. Stereotype application is also conditioned by the homogeneity of the social context of frontline institutions.
• Efforts to decrease the impact of stereotypes should focus on stereotype activation and could involve increasing social heterogeneity in frontline encounters. This could be done by organizing work to increase interaction between street-level bureaucrats and citizens from different classes. Also organizing work emphasizing the collection of different types of information and perspectives may be effective in reducing stereotype application.
Introduction
The literature on street-level bureaucracy has consistently demonstrated how discretionary judgements are shaped not only by political, administrative, or organizational structures, but also by frontline social relations, identities, and citizen encounters (Brodkin 2012; Dubois 2010;
2014; Epp, Maynard-Moody, and Haider-Markel 2014; Lipsky 2010; Maynard-Moody and Musheno 2003, Soss, Fording, and Schram 2011). As pointed out by e.g. Lipsky (2010) and Maynard-Moody and Musheno (2003), one important mechanism shaping street-level
judgements is the use of social stereotypes, but not many studies have tried to disentangle more precisely how and when street-level bureaucrats use stereotypes. This lack of specific studies of stereotype use at the street-level is unfortunate, as stereotypes involves the possible introduction of bias in encounters between state and citizens.
In light of this, the article makes three contributions. First, a theoretical framework for studying stereotypes, drawing on sociological and psychological theory, is suggested. Here, it is suggested to see stereotypes as reducing both cognitive and social uncertainty as they function as a frame for the interpretation of information. Also, the distinction between stereotype
activation and application is suggested as key to studying how and when street-level bureaucrats use stereotypes. Second, a methodological framework using vignette experiments that allows for interpretive analysis of reasoning and systematic assessment of final decisions is suggested.
Third, empirical results show how stereotype use is linked to class differences between SLBs and citizen-clients, and to the social context of street-level work.
Discretion and stereotypes: What do we know?
The existing literature typically includes stereotypes as one among several mechanisms linked to frontline discretion, and there is a consensus that stereotypes function as a tool to reduce the uncertainty of information and situations which characterizes street-level work. For Lipsky, discretionary powers are constitutive for street level bureaucracy due to task complexity, “the human dimension of the situation” (Lipsky 2010, 15), and SLB–citizen interaction. At the same time, SLBs work under conditions of (perceived) resource scarcity, severe workloads, and conflicting demands. This facilitates different coping mechanisms (Tummers et al. 2015), and one such mechanism, Lipsky suggests, is “people-processing”, i.e. the sorting, screening, and prioritizing of clients using social stereotypes. Lipsky thereby links the use of stereotypes to both cognitive and organizational aspects of street-level work (Lipsky 2010, 115).
The literature on bounded rationality further underlines how inadequate and ambiguous information, as well as limited human cognitive capacity, facilitates the use of short- cuts, such as stereotypes, in discretion and decision-making (Simon 1947; Jones 2003; Keiser 2010). Also, the literature on citizen-agency (Dubois 2010; Maynard-Moody and Musheno 2003) show how frontline work is embedded in social and cultural relations, and how social
uncertainties facilitate the use of social stereotypes as a way of navigating social interactions (Raaphorst 2017). However, most of the existing literature focus more on the possible effects of stereotypes than on details on how and when they are used or not used (e.g. Wenger and Wilkins 2009).
A few recent studies are notable exceptions, though. Soss, Fordring, and Schram (2011) find that even though SLBs do not directly express crude stereotypes, they nevertheless draw on racial stereotypes when exercising discretion regarding welfare benefits and sanctions.
A similar demonstration of the use of racial stereotypes is provided by Epp, Maynard-Moody, and Haider-Markel (2014) in their study of police stops, and Møller (2016) further demonstrates how use of social stereotypes linked to specific diagnoses influence the reasoning of Danish social workers exercising discretion on social security eligibility. These are valuable
contributions, but still focus is primarily on demonstrating the use of stereotypes and not on exploring further, how stereotypes work, and when they are used or not used. These questions, however, are important for understanding stereotypes as a mechanism in frontline discretions, and for developing strategies to avoid the possible bias and injustice that stereotypes can entail.
In the following, focus is given to the first question, whereas possible strategies will be addressed in the discussion.
Stereotypes at the Street Level: A theoretical framework
Although only few existing studies explore the use of stereotypes in detail, many existing theoretical insights can help build a theoretical framework and form a number of general expectations. In particular, existing psychological and sociological work can supplement the existing understanding of stereotypes within the street-level bureaucracy literature.
Following Lipsky, discretion can be understood as the space for decisions given to SLBs within the broader context of policy (Lipsky 2010, 15; Hupe 2013). However, following the notion of bounded rationality, and insights from the sociology of professions, discretion can also be seen as a continuous mode of practical reasoning in the context of uncertainty and ambiguity. Here, SLBs draw on different sources, such as policies, organizational procedures, professional knowledge, social norms, personal beliefs, and individual and societal values (Hupe 2013; Wagenaar 2004; Schön 1981; Molander and Grimen 2010). Stereotypes are thus one
possible tool applied in discretionary reasoning, and therefore it can be expected to be visible not only in concrete decisions, but also in the ways SLBs interpret information and how they make sense of and justify their exercise of discretion.
Within psychology, stereotypes are seen as cognitive categories that link group membership to certain attributes and behaviors (Hilton and Hippel 1996; Dovidio et al. 2010;
Lakoff 1987), and as “an over-simplified mental image of (usually) some category of person, institution or event” (Tajfel 1981b, 144; 1981a). This means, that while categories can be used in non-simplified ways, stereotypes designate the simplified use of categories including simplified assumptions about characteristics and behaviors. Sociologists typically adds to this that the function of stereotypes is not only to give cognitive shortcuts but also to navigate social relations and identities, and that oftentimes stereotypes result in the exaggeration and reproduction of differences between one’s own group (in-group) and others (out-groups) (Schütz 1960; Berger and Luckmann 1966). Also, the “shared-ness” of stereotypes is taken to indicate that stereotypes exist not only as individual, cognitive categories, but also as collective categories embedded in broader discourses and shared cultures (Harrits and Møller 2011; Yanow 2003; Lamont and Molnár 2002). Whereas psychology will tend to see stereotypes as universally shared understandings of different social groups (e.g. Fiske 2015), sociology will emphasize the
possibility that stereotypes are shared among e.g. national cultures or embedded in broader social groups defined by e.g. class or race (Lamont 1992; Bourdieu 1984). This, supports the
expectation that stereotypes are used not only to reduce cognitive uncertainty but also as a way of navigating and negotiating social identities and group boundaries.
Further, some psychology scholars have warned against understanding the use of stereotypes only as the direct expression of simplified assumptions on traits and behavior (e.g.
the assumption “criminal” when confronted with a black young boy, or the assumption
“laziness” when confronted with a young mother receiving welfare benefits). Such assumptions on specific behaviors are only one aspect of stereotype use, since stereotypes can also work as a frame, which people use to interpret and process a range of information not only on behavior but also on e.g. character, preferences and intentions (Jasso and Webster Jr 2008; Biernat 2003;
Biernat and Manis 1994). Within a sociological understanding, a similar point can be seen in Bourdieu’s conception of habitus, understood as an integrated system of cognitive, normative and aesthetic heuristics used to interpret information about the social world (Bourdieu 1984;
Bourdieu 1990).
Also, Devine (1989) distinguishes between stereotype activation and the possible control of stereotype activation in decision making, something which could also be termed stereotype application. Stereotype activation is primarily uncontrolled, since stereotypes are deep-seated, culturally and socially learned categories. However, the application of stereotypes, i.e. the application of the assumptions and frame embedded in stereotypes to one’s actions or decision can be controlled, meaning that some people can refrain from applying stereotypes.
Devine links this to psychological features such as beliefs and attitudes, however in a
sociological understanding the control of stereotype application could be linked to social context.
Bourdieu (1984) links the use of specific categories and stereotypes to specific social positions and thus to class differences; Coser (1975) argues that people in diverse, heterogeneous, and complex social settings will tend to develop reflexivity and mental flexibility which weakens the use of stereotypes, and Lamont argues that increased social differentiation will tend to produce different and competing sets of categories potentially weakening the impact of each individual set of categories and stereotypes (Lamont 2012).
Several general expectations for the study of class stereotype use among SLBs follow from this. First, the activation and application of class stereotypes may not only be seen as a direct activation of assumptions on behavior, but also as a frame that SLBs use to interpret different types of information, guide reasoning on problems and decisions, and navigate social relations. Second, the activation and application of class stereotypes as a specific type of simplified categories may be linked to the ways in which SLBs and clients are situated in different classes. Stereotypes may therefore be activated and applied more strongly in cases where clients are members of a social class different from the social class of SLBs (se also Harrits and Møller 2014). This expectation emphasizes the use of stereotypes as linked to SLBs own middle-class identity vis-à-vis the identity of citizens as either the middle class or not. The point is not, though, that SLBs will activate and apply similar stereotypes towards upper and lower classes (which they probably will not), but that they will apply stereotypes to a similar extent. It is thus class difference, and not class hierarchy that is seen as important for the use (but not necessarily content) of stereotypes. Third, the activation and application of stereotypes may vary depending on the social context of the street-level organization, so that SLBs working in heterogeneous communities will tend to apply stereotypes to a lesser extent than those working in homogeneous communities.
Data and Methods
The study presented here explores these three expectations, focusing on implementation of Danish preventive policies within healthcare, childcare, and education (see appendix for further description). Preventive policies present SLBs with the challenging task of identifying,
intervening in, and preventing possible future problems. Discretion is therefore broad and often
based on sketchy and incomplete information and observations (Møller and Harrits 2013). More precisely, focus in the following will be on the initial concern of SLBs with respect to a specific child or family, i.e. the exercise of discretion with regard to the initiation of small preventive initiatives, further investigation or screening, or the continuous “keeping an eye on” a specific child and family.
This case was chosen because of the broad nature of discretion, which allows for a high degree of uncertainty regarding both information and social relations between SLBs and citizens. Also, the nature and content of Danish preventive policies, which focus strongly on problems regarding social mobility and increasing opportunities for all children, increase the likelihood of class stereotypes being activated. This makes the case well suited for exploring the use of class stereotypes more in depth, thereby facilitating an empirical contribution to the street- level bureaucracy literature where studies of class stereotypes are rare. However, the nature of the case may limit possible inferences to other policy areas. This problem of inference will be further addressed in the discussion of results.
The study was conducted as a combination of semi-structured interviews and an experimental, qualitative vignette embedded within the interviews (Kvale and Brinkmann 2009;
Charmaz 2006). The semi-structured interviews facilitate in-depth and interpretive analyses of SLB reasoning and use of stereotypes, including how stereotypes are used when interpreting information. However, relying solely on this makes it difficult to compare the use of stereotypes across individuals and contexts, since each interviewee will be focused (as they should be) on their own specific context and experiences. Therefore, a vignette experiment was embedded halfway into the semi-structured interviews, asking the interviewees to comment on two fictive case-stories presented to them on a piece of paper.
The vignettes were designed to test the impact of class stereotypes, systematically varying information about social class background of the child, with names and information about employment of parents used as cues. Apart from these class cues, the problems described in the case stories were similar, and all problems were deliberately designed to be ambiguous.
Hence, consulting with field experts, each vignette described problems which on one hand could be interpreted as something many children will encounter as part of growing up, thereby not raising further concern, but which on the other hand also could be interpreted as signs of more serious problems and thus needing further screening or intervention. Moreover, the case stories were adapted to the specific group of SLBs in order to present a realistic story with a realistic set of problems. This also meant adapting the age of children and parents to the specific age-groups normally encountered by the SLBs. The vignettes were developed based on pilot observations and discussions with field experts belonging to each of the three professions. This helped make the vignettes as authentic as possible, and it helped describe problems similar in character and seriousness (cf. below).
Two stories were presented to each interviewee. The first story (vignette MC) provided a baseline with social class cues about middle class, i.e. a class background similar to that of the three professional groups. The second story, were given in two versions (vignette UC and LC), each containing cues on upper class and lower class respectively. The UC/LC stories were distributed randomly between the interviewees. Problems in the UC and LC stories were identical, whereas problems in the MC story were similar with respect to the nature and seriousness of the problems. Comparing the difference between how SLBs asses and reason when given these two stories (MC and UC/LC) will thus facilitate an analysis of stereotype use in situations with class difference (cf. above), whereas comparing reactions between the two
groups given either the UC or LC story, will facilitate an analysis of stereotype use and content in situations with class hierarchy.
[Table 1 about here]
The vignettes were presented to the interviewee mid-way through the interview, and they were kept in a closed envelope until then, preventing the interviewer from knowing which stories would be presented. The interviewees were asked to interpret the information in the stories, and probes were made regarding a final decision with respect to the SLBs degree of concern for the specific child and possible suggestions for interventions. Before the vignette experiment, the semi-structured interview touched on questions regarding the work practices and routines of the SLB, how they themselves would describe their workplace and the community they worked in, and how they approached the identification of children possibly at risk or in need of extra help.
After the vignettes were discussed, the interview turned to questions on the individual social background of the SLBs and their professional self-understanding (see table A.1 in appendix for interview protocol).
Asking interviewees to freely discuss and interpret the information presented in the vignette stories facilitates an interpretive analysis of stereotype activation in discretionary reasoning, whereas asking the SLBs to make a final decision on their level of worry in each of the two vignettes, as well as compare their level of worry between the two vignettes, facilitates an analysis of stereotype application in the final decisions, when different factors are weighed against each other. Due to the design, where differences between the two stories are mainly
related to social class background, an SLB decision with higher level of concern in the UC/LC vignette compared to the MC vignettes can be taken as an indication of stereotype application.
One caveat exists, though. Different measures were taken to ensure that the problems in vignette MC and UC/LC were as similar as possible, so that there was no
“objective” or professional reason to be more concerned for the one compared to the other. Even so, the different problems in the vignettes means that the experimental design itself cannot guarantee that differences in discretion were only related to class cues and not to difference in problems. However, since the interviews were asked to compare their decisions on the two vignettes, the interviews also captured (in almost all cases) the specific reasons given by SLBs for why they were more concerned about the UC/LC vignette compared to MC vignette. As can be seen in the appendix (table A.3), the analysis of these data strongly indicates that the class cues indeed were effective and therefore can be taken as an indication of stereotype application.
Interviewees were selected theoretically in several steps. First, the analysis includes three groups of SLBs: Public health nurses provide healthcare to families with pre-school
children (aged 0–6 years) and children in primary schools (ages 6–16); pre-school teachers provide childcare in general childcare institutions (children ages 3–6), covering almost all Danish children (European Commision 2014); and primary school teachers in this study work in the first three grades of primary schools (children ages 6–9). Although all are concerned with care, education or both, these are indeed three separate professions. Health nurses have an educational background in health and nursing, pre-school teachers have an education focusing primarily on child development and pedagogy (and they are literally referred to as ‘pedagogues’
in Danish), and primary school teachers have an education combining an academic curriculum in
different subjects (math, language, science etc.) with a didactic and pedagogical curriculum.
Also, these three professions are organized in three separate professional associations.
Further, with regard to the degree of professionalization, social class position and societal prestige, the three professions are very similar. They all have BA degrees aimed at professional work from non-research university colleges, and based on both their education, employment, salary level and societal prestige, they can all be considered as belonging to the Danish middle class (Goldthorpe 1996; Harrits 2014). Also, all three professions perform their work tasks in informal settings, interacting closely with children and families (see table A.4 in appendix for background information on individual interviewees).
Second, interviews were performed in two different communities in four
municipalities (eight communities in total). The municipalities were chosen with the intention of strengthening possible analytical inference, including large and small municipalities in rural and urban areas, and the local communities were chosen in order to vary the homogeneity of the frontline institutions’ social context. More specifically, within each municipality, a homogenous and primarily middle-class local community and a community with a high degree of social class heterogeneity were selected. Local communities were operationalized as the municipal
administrative unit of school districts, which is the basis for organizing both family healthcare, childcare, and schools in all of the municipalities included in the study. Information about population composition was obtained from the local administration in the four municipalities.
Third, within these eight local communities, managers at local institutions (healthcare agencies, childcare institutions, and primary schools) were contacted and asked for SLBs who would be willing to be interviewed. If not enough interviewees could be recruited in one institution, a second (and third) institution in the same local community was contacted. A
total of 58 interviews (22 teachers, 20 pedagogues, 16 public health nurses) lasting 1–3 hours were conducted by two interviewers. This method of recruiting interviewees through local managers involves the risk of some selection bias, either based on the managers selecting
employees who they think can represent their institutions well, or based on the most interested or active employees volunteering to do interviews. However, even if some selection bias is present along these lines, this will not necessarily jeopardize the results. Rather, such a selection bias could be seen as a hard case and most likely lead to underreporting of stereotype use as opposed to the opposite.
All of the interviews have been transcribed verbatim by student assistants.
Analyses have been performed by the author, based on coding frames (see table 2) applied systematically across all 58 interviews in Nvivo (Charmaz 2006; Saldana 2013; Miles,
Huberman, and Saldana 2014). Exploring SLB interpretations and the use of stereotypes, an open and grounded coding strategy was chosen. Open coding was performed on a subset of the
interviews, followed by focused coding across all interviews. Supplementing this, interpretive memos on each interviewee were made. Also, a focused coding on interviewee descriptions of local communities and their own social background was performed (see appendix for coding frame and summary analysis on social background of SLBs).
Finally, responses to the vignettes have been coded using a focused coding frame, followed by an individual categorization of each interviewee. This categorization was initially done for analyses in previous work (Harrits and Møller 2014) and repeated for the present analysis to test reliability, resulting in no change. For simplicity of comparison, and the ability to perform simple statistical tests, a recoding was performed, transforming the three qualitative responses into a binary scheme of concern or no concern (see table A.5 in appendix).
[table 2 about here]
In the following, analyses proceed focusing on the three expectations listed above. First, attention is given to the relationship between stereotype activation, application and social class difference, including a statistical analysis of the vignette experiment as well as an interpretive analysis of how two dimensions of stereotype activation (categorization and simplified
assumptions) are both related to class differences. Second, we turn to the analysis of how social context may impact stereotype activation and application, again including both a statistical analysis of the vignette experiment as well as a interpretive analysis of stereotype use. Both analyses address the expectation that stereotype activation and application are used to provide frame for interpreting different types of information and to navigate social relations.
Stereotypes and class difference
Table 3 displays the results of the vignette experiment and SLBs decisions with regard to their level of concern for the children in the vignette stories. As can be seen, there is a significant difference in level of concern for the middle-class boy compared to both the upper class and the lower-class boy, indicating that stereotype application is related to class difference. This is most evident among health nurses, but nevertheless significant in all three professions. On average, more than half of SLBs express concern about the upper and lower-class boy compared to only around 9% who express concern about the middle-class boy, and this is similar across
professions and when seen separately for the upper- and lower-class vignettes.
[Table 3 about here]
Since the nature of the problems in all vignette stories are similar, and since almost all
interviewees themselves suggest that the differences between the two vignettes are primarily the social class background of the families and not the problems described (cf. analysis in the appendix), this suggest both activation and application of class stereotypes by SLBs when confronted with children and families from social classes different from the middle class, i.e.
their own class.
Analyses of how SLBs interpret the vignettes support this conclusion, and further sheds light on stereotype activation and how stereotypes are used to interpret information and navigate social relations. As shown in table 4 and 5, two dimensions of stereotype activation were identified, both relating to class differences. The first dimension concerns the use of marked or un-marked class categories when interpreting information on the situation and the family, whereas the second dimensions has to do with the use of assumptions about
characteristics and behaviors of different classes.
First, when confronted with especially the lower-class vignette, SLBs tend to clearly mark the class identity of families (Yanow 2003), by referring to either the educational background, level of income, or employment status of the parents, by referring to the level of resources in the family, or even to a lifestyle characteristic typical of a specific class. This also happens when discussing some of the upper- and middle-class vignettes, however not as often.
Also, looking more closely at the use of categories (cf. table A.6 in the appendix), the marked class categories used when discussing middle-class vignettes refers mostly to employment, whereas the used of marked class categories in when discussing the upper-class vignettes refers also to resources and lifestyles, thus resembling more closely the use of marked class categories
when discussing the lower-class vignettes. Finally, there seems to be a wide tendency for SLBs to use un-marked class categories when describing the middle-class families, for example by characterizing the family as normal, by referring to “universal” and inclusive labels such as
“families” or “children”, or by referring to other characteristics, e.g. the age of the boy. This use of un-marked class categories is not very widespread when discussing either the upper- or the lower-class families.
The second dimension of stereotype application has to do with assumptions about characteristics and behaviors. Sometimes SLBs present simplified assumptions based on the class of the parents, and these assumptions are furthermore presented with a certain level of certainty, for example not presenting any possible alternative interpretations. Therefore, this is referred to as closed interpretations. Other times SLBs present more open interpretations, emphasizing the need to get further information, or suggesting alternative interpretations to supplement either their own previous interpretation or as an alternative to what they themselves portray as stereotypes. As seen in table 5, there is to be a tendency across professions and in all three vignettes for SLBs to present open interpretations, typically by referring to procedures founded in institutional routines and professionalism, for example when health nurses or pre- school teachers suggest screening children for developmental problems (see table A.6 in the appendix).
However, table 5 also shows how interpretations applying simplified and class- based assumptions are more often presented when discussing either the upper-class or the lower- class vignettes. The use of simplified assumptions thus seems to depend on class difference, but the content of these assumptions is somewhat different when discussing either the upper-class or the lower-class vignette (see table A-6 in appendix). SLBs typically make assumptions about the
resources and abilities of Mike’s lower-class family to take care of him and help him, whereas assumptions made when discussion Vitus’ upper-class parents are more focused on the
willingness of Vitus’ parents to take care of Vitus, due to their busy work schedule and high expectations to the child. Thus, whereas the lower-class family is met with the somewhat paternalistic assumptions about lack of competence, upper class families are met with the assumptions of the parents being rational, ‘cold’ and too hard, and therefore less willing to and capable of caring for the child. These stereotypical and simplified assumptions resemble class stereotypes found in other studies (Lamont 1992; Fiske 2015), and they also match the
stereotypes found in analyses of SLBs spontaneous narratives about their own practice (Harrits and Møller 2016).
Summing up these findings, there is a strong indication that SLBs rely on
stereotypes when interpreting information about families and children in the vignettes. The use of stereotypes is clearly related to class, with the use of un-marked class categories more often towards middle-class families, underlining the “universality” and normality of the, and the use of closes class-based assumptions more often towards both upper- and lower-class families. The interpretive analysis thus seems to confirm theoretical expectations about the activation of class stereotypes and about how stereotypes are used as a way to interpret and make sense of
information. Also, the use of marked and non-marked categories, and the ways in which the middle-class problems are clearly normalized, whereas both upper- and lower-class problems are marked as worth noting, suggest that stereotypes are also used as a way to navigate and negotiate class differences. Finally, stereotypes seem also to be applied to some degree in the final
decisions about level of concern.
Stereotypes and social context
Moving on to the expectation about use of stereotypes in different social contexts, table 6 displays an interesting pattern across the different social contexts of frontline institutions.
[Table 4 about here]
As can be seen, almost all health nurses and pre-school teachers in homogenous communities apply stereotypes in their final decision on level of concern, expressing more concern towards the upper and lower class families. Compared to this, very few health nurses and pre-school teachers in heterogeneous contexts apply stereotypes in their final decisions on concern. This seems to support the theoretical expectations described above with regard to the impact of homogeneity in the social context of street-level organizations. However, for teachers a similar pattern is not found. Here, the differences in concern between teachers in homogenous and heterogeneous communities are not significant.
The interpretive analysis of the use of stereotypes when interpreting the information from the vignettes, however, does not show a similar clear difference between heterogeneous and homogenous contexts. As can be seen from the detailed display of stereotype activation (table A.6 in the appendix), a similar pattern in the activation of both marked and un- marked class categories and of open and closed interpretations seem to be present both in homogenous and heterogeneous contexts. There may be seen a small tendency for SLBs in homogenous context to activate more stereotypes and for the degree of simplification and prejudice to be a little more pronounced in homogenous context. However, SLBs in
heterogeneous context overall does seem to activate stereotypes in ways much similar to SLBs in homogenous contexts.
To understand the seemingly contradictory results from the statistical and
interpretive analysis of differences across social contexts of frontline institutions, the distinction between stereotype activation and stereotype application is helpful. One possible interpretation of the results is thus that SLBs in both homogenous and heterogeneous contexts activate class stereotypes when confronted with information on families from different class backgrounds. As described above, they tend to activate the un-marked class categories when confronted with middle-class families, whereas they tend activate marked class categories and class-based simplified assumptions when confronted with both upper-class and middle-class families.
However, when weighing the different types of information and reaching a final decision on their level of concern, it seems to be SLBs in homogenous contexts that apply stereotypes, whereas SLBs in heterogeneous context do not. This finding, however, is valid only for health nurses and pre-school teachers working in child care, and not for teachers.
The analysis presented here has not, however, said anything about how the individual family class background and individual lifestyles of SLBs may also be shaping stereotype activation. In the appendix, a supplementary analysis of this questions is presented, showing no significant differences. Individual family class background and individual lifestyles of SLBs can therefore not in this analysis be seen to impact stereotype use in a manner similar to the social context of street-level organizations.
In sum, the analysis indicate that stereotypes activation is similar across social contexts, and related to social class differences. However, stereotype application in the final decisions seems to be conditioned by social context, with a heterogeneous social context clearly
lowering the tendency for using stereotypes, at least among health nurses and pre-school teachers working in child care. The results thus partly support the expectation that stereotypes are used less in heterogeneous social context. This may be due to the increased mental flexibility in heterogeneous contexts as suggest by Coser (1975), but this mental flexibility is, then, only effective in the controlled application of stereotypes in decisions, and not in the uncontrolled activation of stereotypes in information interpretation.
Discussion
The results presented above suggest that whereas stereotypes are activated among almost all interviewees, there are differences in the ways in which stereotypes are activated towards different classes, and differences in the extent to which they are applied across different social contexts. Where stereotype activation thus seems to be related to social class, stereotype application is related to social class but further conditioned social context. Two dimensions of stereotype activation, both related to class, were found, namely the use of un-marked vs. marked class categories, and the presentation of closed and simplified class-based assumptions open interpretations and rejection of assumptions.
These findings support the expectations presented above, as well as theoretical understanding of stereotypes in general. Especially, the finding that activation of stereotypes is related to class differences and contains both the use of marked categories and simplified assumptions, support the idea that stereotypes are used to reduce both cognitive and social uncertainty (Raaphorst, 2017). Stereotypes are not only used to interpret information on concrete situations, but also to interpret and categorize people. The clear class differences in stereotype activation and application further suggest that SLBs use stereotypes to interpret and navigate the
social relations of frontline work, and to negotiate class boundaries and differences between what is ‘normal’ and what is noteworthy and possibly concerning. Further, the finding that differences in stereotype application is conditioned by social context suggest that heterogeneous social relations can support the inhibition or control of stereotype application.
These results are based on a limited number of interviews, and inferences can therefore only be analytical. Supporting analytical inference is the fact that saturation is high across the data. This means that inclusion of data from different types of municipalities and from different frontline institutions did not introduce any significant new dimensions. Also, taken as a whole, the ways in which interviewees talk about the topic and vignettes is very comparable, and the content of activated class stereotypes is remarkably similar across the three professions. This supports a cautious analytical inference to a larger population of at least Danish street-level bureaucrats. Furthermore, the experimental design of the vignette experiment lends supports to the conclusion that SLBs activate social stereotypes differently depending on the social
background of the children and families they meet. However, when it comes to the variation across social context of the communities, in which SLBs work, a more hesitant interpretation should be made. With the design applied here, differences across social contexts could be the result of selection bias, i.e. reflecting the fact that SLBs choosing to work in heterogeneous communities simply apply stereotypes in a different way, perhaps due to a specific set of personal beliefs (Devine 1989). Still, the analysis of the individual social background of SLBs (see appendix) does not support the existence of such a selection bias.
Also, it should be remembered that teachers display a different pattern here, compared to the two other groups. The analysis here has not explored this difference further, but one can speculate about possible reasons. Most importantly, teachers encounter children older
than the children encountered by health nurses and pre-school teachers, and they encounter children and families in a more institutionalized setting. This means that they have less interactions with parents, which could result in the heterogeneous social context being less salient for them. Also, differences could be related to the specific content of teacher’s professional background compared to health nurses and pre-school teachers, with teacher education focusing somewhat more on a didactic curriculum. Finally, it could be related to individual factors such as beliefs or values not explored in this analysis.
Finally, as mentioned when discussing the case of Danish preventive policies and the selection of the three professions, inference beyond preventive policies and professions focused on care and education could be problematic, since the content and function of these policies could make class differences especially salient for SLBs. Still, many welfare and health policies have a similar preventive purpose, focusing on the identification of social risk as well as redistribution of resources. Other professionals like social workers, physiotherapist and medical doctors, thus have similar discretion with regard to interpreting situations of families and
determining risks and concerns, and the mechanisms found here may be relevant for these groups also.
Taken together, the results contribute to understanding how stereotypes work at the street level, and also point towards a discussion on how to inhibit or control the impact of
stereotypes, prejudice and bias. First, it is worth underlining the fact that stereotype activation is mostly uncontrolled, and that stereotypes work as a way of navigating cognitive and social uncertainties. We should therefore expect stereotype activation to be widespread at the street- level, and instead focus on ways to control stereotype application (see also Bodenhausen and Macrae, 1998; Devine, 1998).
One suggestion for controlling bias comes from the public administration literature on representative bureaucracy, where results suggest that equal group representation in
bureaucracies will weaken bias (Bradbury and Kellough 2011). Striving towards equal
representation with regard to such demographic characteristics will therefore expectedly work as a measure against bias. However, this may be an obvious solution when considering bias of gender, race and e.g. physical ability, equal representation of social classes in frontline
institutions will be very difficult to obtain, since there is an obvious link between employment and class positions of SLBs. Hence, representative bureaucracy as a solution to the use of class stereotypes is not viable. Also, the representative bureaucracy literature seeks to counteract bias by securing equal representation and not by controlling stereotype application.
This makes it all the more important to look for other possible ways to weaken bias and stereotype application. The psychological literature on stereotype inhibition focuses mainly on individual characteristics such as self-regulation and anti-prejudicial belief systems
(Bodenhausen and Macrae, 1998; Devine, 1989, 1998; Gordijn et al., 2004). Surely,
strengthening, for example, anti-prejudicial beliefs among street-level bureaucrat, for example by strengthening professional and organizational norms about equal treatment, could be a way forward.
However, as has been underlined throughout this article, stereotypes are also embedded in social relations, and in the present study both to relations of class and to the social heterogeneity of the communities of street-level institutions. Following sociological insights also mentioned above (Coser, 1975; Lamont, 2012; see also Devine, 1998), a possible strategy for inhibiting stereotype application could, therefore, be related to the organization of street-level work and the social encounters and relations of this work. One suggestion could be to design
public institutions so that they maximize heterogeneous social relations and experiences, for example by maximizing interactions between SLBs and citizen-clients from different class backgrounds. Also, the installment of procedures insisting on the collection of different,
heterogeneous, sources of information and perspectives (i.e. open interpretation), may be helpful.
Conclusion
Research on street-level bureaucracies often mention stereotypes as one possible mechanism shaping discretion, but not many studies have explored this mechanism in detail. As
demonstrated in this article, SLBs use stereotypes when interpreting information and making decisions. Marked categories and simplified assumptions are activated more when SLBs are faced with citizens from a social background different from their own, and stereotypes are applied more often by SLBs working in a homogenous social setting.
The results give important contributions to both practical discussions and further research. Working to control stereotype activation, solutions may be found in both the
psychological literature focusing of anti-prejudicial beliefs, and in the sociological literature emphasizing the heterogeneity of social relations and experiences. However, as the present study has focused on SLBs use of stereotypes, no definite recommendations on strategies for
controlling stereotype application can be made. Future research may thus benefit from focusing more specifically on different strategies for controlling stereotype application.
Notes:
This article was supported by funds from the Danish National Research Foundation for Independent Research. The author thanks Marie Østergaard Møller for collaboration on data
collection. Also, the author thanks Marie Østergaard Møller, Nadie Raaphorst, Vibeke Lehmann Nielsen, and Steven Maynard-Moody for valuable comments and suggestions to earlier drafts of the article.
References
Berger, Peter L., and Thomas Luckmann. 1966. The Social Construction of Reality. Garden City:
Doubleday.
Biernat, Monica. 2003. “Toward a Broader View of Social Stereotyping.” American Psychologist 58 (12): 1019–27.
Biernat, Monica, and Melvis Manis. 1994. “Shifting Standards and Stereotype-Based Judgments.” Journal of Personality and Social Psychology 66 (1): 5–20.
Bourdieu, Pierre. 1984. Distinction. A Social Critique of the Judgement of Taste. London:
Routledge.
———. 1990. The Logic of Practice. Stanford: Standford University Press.
Bradbury, Mark, and J. Edward Kellough. 2011. “Representative Bureaucracy: Assessing the Evidence on Active Representation.” The American Review of Public Administration 41 (2):
157–67.
Brodkin, Evelyn Z. 2012. “Reflections on Street-Level Bureaucracy: Past, Present, and Future.”
Public Administration Review 72 (6): 940–49.
Bodenhausen G V. and Macrae CN (1998) Stereotype Activation and Inhibition. In: Wyer RS (ed.), Stereotype Activation and Inhibition, Mahwah: Lawrence Erlbaum Associates, pp. 9–
51.
Practice and the Production of Difference in Danish Day-Care Institutions.” Journal of Ethnic and Migration Studies 32 (1): 145–55.
Campbell, Ian 2007. Chi-squared and Fisher–Irwin tests of two-by-two tables with small sample recommendations. Statistics in Medicine 26(19): 3661–3675.
Charmaz, Cathy. 2006. Constructing Grounded Theory. A Practical Guide through Qualitative Analysis. London: Sage.
Coser, Rose Laub. 1975. “The Complexity of Roles as a Seedbed for Individual Autonomy.” In The Idea of a Social Structure. Papers in Honor of Robert K. Merton, 237–64. New York:
Harcourt Brace Jonanovich.
Devine, Patricia G. 1989. “Stereotypes and Prejudice: Their Automatic and Controlled Components.” Journal of Personality and Social Psychology 56 (1): 5–18.
———. (1998) Beyond the Isolated Social Perceiver: Why Inhibit Stereotypes? In: Wyer RS (ed.), Stereotype Activation and Inhibition., Mahwah: Lawrence Erlbaum Associates, pp.
64–74.
Dovidio, John F, Miles Hewstone, Peter Glick, and Victoria M Esses. 2010. “Prejudice, Stereotyping and Discrimination: Theoretical and Empirical Overview.” In The SAGE Handbook of Prejudice, Stereotyping and Discrimination, edited by John F Dovidio, Miles Hewstone, Peter Glick, and Victoria M Esses, 3–28. London: Sage.
Dubois, Vincent. 2010. The Bureaucrat and the Poor. Encounters in French Welfare Offices.
Farnham: Ashgate.
———. 2014. “The State, Legal Rigor, and the Poor: The Daily Practice of Welfare Control.”
Social Analysis 58 (3): 38–55.
Epp, Charles, Steven Maynard-Moody, and Donald P. Haider-Markel. 2014. Pulled Over. How
Police Stops Define Race and Citizenship. Chicago: Chicago University Press.
European Commision. 2014. Key Data on Early Childhood Education and Care in Europe.
Bruxelles: Education, Audiovisual and Culture Executive Agency.
Fiske, Susan T. 2015. “Intergroup Biases: A Focus on Stereotype Content.” Current Opinion in Behavioral Sciences 3. Elsevier Ltd: 45–50.
Goldthorpe, John H. 1996. “Class Analysis and the Reorientation of Class Theory : The Case of Persisting Differentials in Educational Attainment.” British Journal of Sociology 47 (3):
481–505.
Gordijn EH, Hindriks I, Koomen W, et al. (2004) Consequences of Stereotype Suppression and Internal Suppression Motivation: A Self-Regulation Approach. Personality and Social Psychology Bulletin 30(2): 212–224.
Harrits, Gitte Sommer. 2014. “Professional Closure Beyond State Authorization.” Professions and Professionalism 3 (2).
Harrits, Gitte Sommer, and Marie Østergaaard Møller. 2016. Forebyggelse og bekymring i professionel praksis. København: Hans Reitzels Forlag.
Harrits, Gitte Sommer, and Marie Østergaard Møller. 2011. “Categories and Categorization:
Towards a Comprehensive Sociological Framework.” Distinktion: Scandinavian Journal of Social Theory 12 (May 2012): 229–47.
———. 2014. “Prevention at the Front Line: How Home Nurses, Pedagogues, and Teachers Transform Public Worry into Decisions on Special Efforts.” Public Management Review 16 (4). Routledge: 447–80.
Hilton, James L, and William Von Hippel. 1996. “Stereotypes.” Annual Review of Psychology 47: 237–71.
Hupe, Peter. 2013. “Dimensions of Discretion: Specifying the Object of Street-Level Bureaucracy Research.” DMS-Der Moderne Staat, 425–40.
Jasso, Guillermina, and Murray Webster Jr. 2008. “Double Standards in Just Earnings for Male and Female Workers.” Social Psychology 60 (1): 66–78.
Jones, Bryan D. 2003. “Bounded Rationality and Political Science: Lessons from Public
Administration and Public Policy.” Journal of Public Administration Research and Theory 13 (4): 395–412.
Keiser, Lael R. 2010. “Understanding Street-Level Bureaucrats’ Decision Making: Determining Eligibility in the Social Security Disability Program.” Public Administration Review 70 (2):
247–57.
Kvale, Steinar, and Svend Brinkmann. 2009. Interview. Introduktion Til et Håndværk. Second.
København: Hans Reitzels Forlag.
Lakoff, G. 1987. Women, Fire, and Dangerous Things. Chicago: Chicago University Press.
Lamont, Michèle. 1992. Money, Morals and Manners. The Culture of the French and American Upper-Middle Class. Chicago: Chicago University Press.
———. 2012. “Toward a Comparative Sociology of Valuation and Evaluation.” Annual Review of Sociology 38 (1): 201–21. doi:10.1146/annurev-soc-070308-120022.
Lamont, Michèle, and Virág Molnár. 2002. “The Study of Boundaries in the Social Sciences.”
Annual Review of Sociology 28 (1): 167–95.
Lipsky, Michael. 2010. Street-Level Bureaucracy. Dilemmas of the Individual in Public Service.
New York: Russel Sage Foundation.
Maynard-Moody, Steven, and Michael Musheno. 2003. Cops. Teachers, Counselors. Stories from the Front Lines of Public Service. Ann Arbor: University of Michigan Press.
Miles, Matthew B., A. Michael Huberman, and Johnny Saldana. 2014. Qualitative Data Analysis. A Methods Sourcebook. Third. London: Sage.
Molander, Anders, and Harald Grimen. 2010. “Understanding Professional Discretion.” In Sociology of Professions. Continental and Anglo-Saxon Traditions, edited by Lennart G.
Svensson and Julia Evetts, 167–87. Gøteborg: Daidalos.
Møller, Marie Østergaard. 2016. “‘She Isn’t Someone I Associate with Pension’— a Vignette Study of Professional Reasoning.” Professions & Professionalism 6 (1): 1–20.
Møller, Marie Østergaard, and Gitte Sommer Harrits. 2013. “Constructing at-Risk Target Groups.” Critical Policy Studies 7 (2): 155–76.
Raaphorst, Nadine. 2017. “How to prove, how to interpret and what to do? Uncertainty experiences of street-level tax officials”, Public Management Review (E-pub ahead of print).
Saldana, Johnny. 2013. The Coding Manual for Qualitative Researchers. Edited by 2. Second.
London: Sage.
Schütz, Alfred. 1960. On Phenomenology and Social Relations. Chicago: Chicago University Press.
Schön, Donald. 1981. The Reflective Practitioner. How Professionals Think In Action. Farnham:
Ashgate.
Simon, Herbert A. 1947. Administrative Behavior: A Study of Decision-Making Processes in Administrative Organizations. New York: Free Press.
Soss, Joe, Richard C. Fording, and Sandford F. Schram. 2011. Disciplining the Poor. Neoliberal Paternalism and the Persistent Power of Race. Chicago: Chicago University Press.
Tajfel, Henri. 1981a. Human Groups and Social Categories. Cambridge: Cambridge University
Press.
———. 1981b. “Social Stereotypes and Social Groups.” In Intergroup Behaviour, edited by John C. Turner and Howard Giles, 144–67. Oxford: Basil Blackwell.
Tummers, Lars G., Victor Bekkers, Evelien Vink, and Michael Musheno. 2015. “Coping during Public Service Delivery : A Conceptualization and Systematic Review of the Literature.”
Journal of Public Administration, Research and Theory (JPART), 1–43.
Wagenaar, Hendrik. 2004. “Knowing the Rules: Administrative Work as Practice.” Public Administration Review 64 (6): 643–55.
Wenger, Jeffrey B., and Vicky M. Wilkins. 2009. “At the Discretion of Rogue Agents: How Automation Improves Women’s Outcomes in Unemployment Insurance.” Journal of Public Administration Research and Theory 19 (2): 313–33.
Yanow, Dvora. 2003. Constructing Race and Ethnicity in America: Category-Making in Public Policy and Administration. New York: M.E.Sharpe.
Table 1: Overview of vignettes
Vignette MC:
Middle class Vignette UC:
Upper class Vignette LC:
Lower class Health Nurses Name: Mads.
Occupation of parents:
Pre-school teacher and teacher.
Problem: Crying in the evening, regurgitation, problems breastfeeding and possible lack of weight gain. Parents are insecure. parents are new to the neighborhood.
Name: Vitus.
Occupation of parents:
Medical doctor and engineer.
Problem:
Lack of weight gain.
Parents are not focused, possible problems with contact to the child. The living room is messy. The mother refuses to
participate in ‘mothers group’.
Name: Mike.
Occupation of parents:
Nursing assistant (unemployed) and truck- driver (on sick leave).
Problem:
Lack of weight gain.
Parents are not focused, possible problems with contact to the child. The living room is messy. The mother refuses to
participate in ‘mothers group’.
Pre-school teachers Name: Mads.
Occupation of parents:
Pre-school teacher and teacher.
Problem:
Aggression and conflict towards other children.
Refuses to eat lunch with the other children and leaves the lunch room.
Difficulties with getting friends. Motor skill problems. Parents new to neighborhood.
Name: Vitus.
Occupation of parents:
Medical doctor and engineer.
Problem:
Child is fragile, quit and keeps to himself during the day. Problems with concentration. Refuses to eat the food in the child care center. Problems with language development.
Name: Mike.
Occupation of parents:
Nursing assistant (unemployed) and truck- driver (on sick leave).
Problem:
Child is fragile, quit and keeps to himself during the day. Problems with concentration. Refuses to eat the food in the child care center. Problems with language development.
Teachers Name: Mads.
Occupation of parents:
Pre-school teacher and teacher.
Problem:
Tells inappropriate jokes and creates disturbance.
Has problems finding friends. Problems with math skills.
Name: Vitus.
Occupation of parents:
Medical doctor and engineer.
Problem:
Disciplinary problems in classroom, but popular among other boys. Has problems with reading skills.
Name: Mike.
Occupation of parents:
Nursing assistant (unemployed) and truck- driver (on sick leave).
Problem:
Disciplinary problems in classroom, but popular among other boys. Has problems with reading skills.
Note: See table A.2 in appendix for full length vignettes.
Table 2: Coding Frame
Code Description
No.
of cases
No. of code references 1. Dimensions in use of stereotypes
1.1 Use of marked
class categories Use of any category based on class, e.g., reference to the
occupation, education, resources or lifestyle of parents 42 93 1.2 Use of non-marked
class categories Use of categories categories based on age or universal categories of e.g. ‘families’, ‘children’, and categories emphasizing normality, e.g. ‘normal boy’.
48 90
1.3 Closed interpretations Interpretations presenting simplified assumptions based on
group membership 52 175
1.4 Open interpretations Interpretations rejecting simplified assumption, suggesting multiple interpretation or insisting on the collection of moreinformation.
50 163
2. Identification of concern in vignette experiment 2.1 Expression of concern
in case stories
Any instance in the judgment of the case stories where SLB expresses concern or specifically expresses no concern
58 206
- SLB expresses no concern 24 37
(+)
SLB expresses some concern, but it is explained as within a
“normal range” and thus something that they should merely
“keep an eye on.”
53 111
+ SLB expresses explicit concern 36 63
2.2 Specific reasoning with regard to increased worry from vignette A to vignette B1/B2
Any instance where SLBs justify why they are more worried about the family in vignette B1/B2 compared to vignette A.
32 56
3. Population composition of local communities
3.1 Area description Any instances of SLBs describing the area within which they
work 58 118
3.1.1 Descriptions of areas as
homogeneous Descriptions of the area as inhabited by “normal”/ordinary people and the lack of diversity, e.g., noting certain groups of people that the SLB never meets
44 62
3.1.2 Descriptions of areas as heterogeneous
Descriptions of the area as inhabited by “all kinds of people,”
and the rejection of not meeting certain groups
37 56
Table 3: Vignette responses across professions and vignette class cues
Concern in
vignette MC Concern in
vignette UC or LC Total number of interviewees
p-value, difference in concern between vignettes (chi-sq)
Health nurses 0.0 % (0) 62.5 % (10) 16 0.000 ***
Pre-school teachers
in childcare 10.0 % (2) 60.0 % (12) 20 0.001 ***
Teachers 13.6 % (3) 50.0 % (11) 22 0.01 **
Total 8.6 % (5) 56.9 % (33) 58 0.000 ***
Note: The analysis compares the level of concern between the two vignettes in each. The p-value is two-tailed, based on the N-1 Chi-square test (Campbell 2007). When expected cell counts fall below 1, the Fisher Exact test is used. ***) p ≤ 0.001; **) p ≤ 0.01; *) p ≤ 0.05
Concern in
vignette MC Concern in
vignette UC or LC Total number of interviewees
p-value, difference in concern between vignettes (chi-sq) Interview given vignette
B1 (Upper class Vitus)
10.7 % (3) 64.3 % (18) 28 0.000 ***
Interviews given vignette
B2 (Lower class Mike) 6.7 % (2) 50.0 % (15) 30 0.000 ***
Total 8.6 % (5) 56.9 % (33) 58 0.000 ***
Note: The analysis compares the level of concern between the two vignettes in each. The p-value is two-tailed, based on the N-1 Chi-square test (Campbell 2007). When expected cell counts fall below 1, the Fisher Exact test is used. ***) p ≤ 0.001; **) p ≤ 0.01; *) p ≤ 0.05
Table 4: Examples of dimensions of class stereotypes
Use of marked and non-marked class categories
Teacher(TC17) talking about lower-class Mike (use of marked class cateogories):
This is a completely different story […]. First, because these parents do not have the same resources as Mads’
parents did. […] And it’s something about education, but primarily it’s because the mother is unemployed and the father is on sick leave. So, their resources are not top level, so to say.
Nurse (NB06) talking about upper-class Vitus (use of marked class categories):
Again, these are well educated people, and so on. But often times with careers like that, the mother is a doctor, so career people, sometimes that thing about just turning it down a knot and finding out what it’s all about … . (NB06, on Vitus).
Teacher (TB10) talking about middle-class Mads (use of marked class categories):
I am wondering a bit about them being teacher and pedagogue. I would wonder about that. Because, I don’t know, maybe I would think that they ought to have the same standards as I do.
Pre-school teacher (CD16) talking about middle-class Mads (use of non-marked class categories):
I think Mads is a very ordinary boy. And I think that I would, uh … I would ask him about his old kindergarten. I think I’d talk to him about that. (CD16, on Mads).
Use of closed interpretations (based on simplified class assumptions) or open interpretations Health nurse (HA01) talking about upper-clas Vitus (closed interpretation):
Well, I am thinking … this family needs to relax and stop reading in 110 different books, but try to see the child the have. And ehm … then I am thinking, I will … the reason that this child is not gaining weight is that the mother has so many projects. […] I don’t think she sees the child. […] Maybe she is very rational, also based on the fact that she is so theoretically knowledgable and wants to know everything. This child is a project.
Teacher (TA03) talking about lower-class Mike (closed interpretation):
Well, I also think about … I mean the parents, where … how can I say this … they don’t have the resources to ehm, for example help the children with homework. […] Many parents don’t have very good experiences with doing their homework. And ehm, maybe they don’t have very good experiences with being a student in school.
Pre-school teacher (CB08) talking about middle-class Mads (open interpretation):
I don’t know, I would have to find out […] So I would, before that conference with the parents, I’d make some observations. Where I would probably get some observations showing, where he ehm … I’d follow him with a notebook, right? Demonstrating that he can’t, what it is … now Mads is sitting in the sandbox, with these children, right? What happens then?
Teacher (TA01) talking about lower-class Mike (open interpretation):
When I read it, I also thought: “Maybe he’s too clever”. Right? Because sometimes we have children who are no stimulated in the right way, and they can be difficult to identify. (TA01, on Mike after first suggesting a different interpretation of the situation).
Note: This table shows examples from the interpretive analysis of dimensions of class stereoypes. The analysis is based on codes 1.1-1.4 as defined in table 2.
