Stereotypes of Norwegian social groups.

We present a pilot study and two main studies that address the nature of stereotypes of social groups in Norway within the framework of the Stereotype Content Model (SCM). The first study focused on stereotypes of a wide range of groups across categories such as gender, age, religious conviction, socioeconomic and health status. The second study focused on stereotypes of immigrant groups. Participants (n = 244 and n = 63, respectively) rated the groups on perceived warmth, competence, status, and competition. Results from both studies support the applicability of the SCM in Norway and provides a unique insight into stereotypes of Norwegian social groups.

Introduction

The Stereotype Content Model (SCM; Cuddy, Fiske & Glick, 2007, 2008; Cuddy, Glick & Beninger, 2011; Fiske, Cuddy & Glick, 2007; Fiske, Cuddy, Glick & Xu, 2002) has been proposed as a pan-cultural framework for understanding how structural relationships between groups (status and competition) shape the content of stereotypes associated with different social groups along two dimensions: warmth and competence (Cuddy et al., 2009; Fiske et al., 2002). Studies across a range of countries indicate that the basic principles of the SCM are universal, but the results also point to important cultural and societal variations (Cuddy et al., 2009; Durante, Fiske, Kervyn et al., 2012). We present two studies that test the SCM in Norwegian samples and employ the model as a framework for describing the Norwegian intergroup landscape. The first study focuses on stereotypes of a wide range of social groups across categories such as gender, age, religious conviction, and socioeconomic and health status. The second study focuses on stereotypes of immigrant groups in Norway. Beyond addressing the universality of the Stereotype Content Model, this provides a valuable foundation for further research into stereotyping, prejudice, and discrimination in Norway.

The Stereotype Content Model

The Stereotype Content Model (Cuddy et al., 2007, 2008, 2011; Fiske et al., 2002, 2007) focuses on shared stereotypes within a culture, organized along two dimensions: warmth and competence. Well-intentioned others are seen as warm (sincere, good-natured) and capable others are seen as competent (skillful, confident). Perceptions of groups' intent and capability originate from intergroup competition and status, respectively. Groups com-peting with the in-group for scarce resources are stereotyped as lacking warmth. Groups who enjoy high status in society, in the form of economic success, education, or prestigious jobs, are stereotyped as competent (Caprariello, Cuddy & Fiske, 2009; Fiske et al., 2002).

Combining the two dimensions provides four main types of stereotypes. High-status, noncompetitive groups (e.g., in-groups, societal reference groups) are stereotyped as both competent and warm. Low-status, competitive groups (e.g., welfare recipients) are stereotyped as incompetent and cold. Low-status, noncompetitive groups (e.g., housewives) are ambivalently stereotyped as warm, but incompetent. Finally, high-status, competitive groups (e.g., rich people) are ambivalently stereotyped as competent, but cold. The inclusion of such ambivalent stereotypes is one of the most central features of the SCM and separates the model from perspectives focusing more exclusively on univalent positive or negative stereotypes (Fiske et al., 2002).

Stereotypes of social groups in Norway

Prejudice, discrimination, and stigma associated with various so-cial groups are not uninvestigated phenomena in the Norwegian context. Examples are studies of how immigrants manage stigma associated with their national backgrounds (Valenta, 2009), levels of anti-Muslim and anti-immigrant attitudes (Strabac, Aalberg & Valenta, 2014), (lack of) sex discrimination of female managers (Storvik & Schone, 2008), discrimination of men, blue-collar workers, and Arabic immigrants in the housing marked (Andersson, Jakobsson & Kotsadam, 2012), negative images of individuals with overweight (Malterud & Ulriksen, 2010), and effects of religious stigma among Muslims (Kunst, Tajamal, Sam & Ulleberg, 2012).

These studies provide valuable insights into the social position and challenges faced by groups in Norwegian society. However, because each study focuses on only one or a few social groups, it is difficult to discern patterns in stereotypes, prejudice, and discrimination across groups. The advantage of describing the intergroup context in the SCM framework is that it provides a picture of stereotypes connected to several social groups, allowing inferences about which groups are likely to share experiences of prejudice and discrimination (Cuddy et al., 2007).

Cross-cultural research has supported the universality of the tenants of the stereotype content model regarding warmth and competence as the organizing dimensions of stereotype content, and group status and competition as the underlying mechanisms that shape stereotypes of groups (Cuddy et al., 2009; Durante et al., 2012). We therefore expect these principles to hold also in a Norwegian sample. At the same time, cultural and economic factors shape the intergroup context in a society. Reference group (i.e., in-group and societal reference group) favoritism is less pronounced in collectivistic countries relative to more individualistic countries (Cuddy et al., 2009). As Norway has been described as an individualistic country (Hofstede, 2001), we expect to see a clear pattern of reference group bias.

Consistent with the SCM framework, the most common finding across cultures is that many (often the majority) of social groups receive ambivalent stereotypes (i.e., high competence/low warmth or high warmth/low competence) (Cuddy et al., 2009; Durante et al., 2012). However, there is also a tendency that fewer groups receive ambivalent stereotypes in societies with a higher degree of income equality (Durante et al., 2012). Put differently, the correlation between warmth and competence varies across countries and a positive correlation between warmth and competence is more common in countries with more economic equality. The distribution of economic resources in Norway is among the most equal in the world (Central Intelligence Agency, 2014). Based on prior research, we expect that many social groups receive ambivalent stereotypes also in a Norwegian sample. At the same time, given the country's level of economic equality, we expect a positive correlation between warmth and competence.

The aim of the first study is to describe the Norwegian intergroup context within the SCM. In order to do this and to align our work with previous research, we tested the basic propositions of the SCM: (1) Warmth and competence will serve to distinguish stereotypes of groups in a Norwegian sample; (2) many groups receive ambivalent stereotypes; (3) perceived status predicts ratings of competence; and (4) perceived competition predicts ratings of warmth.

Study 1

Method

Pilot study

To identify salient social groups in the Norwegian context, we performed the same pilot study as Fiske et al. (2002; Pilot study for Study 2). The participants (N = 40, 50% male) were approached in public places (e.g., parks, bus stops) in the city center of Bergen, Norway, and asked to take part in a short survey about “groups in Norwegian society.” The mean age was 30.93 years (SD = 11.25, range 17–59). All of the participants were Norwegian-born; however, two respondents indicated having an immigrant background (one or both parents born abroad). The participants responded to a short questionnaire containing the three questions employed by Fiske et al. (2002, p. 890):

What types of people do you think are categorized into groups by most people in society (i.e., based on ability, age, ethnicity, gender, occupation, religion, etc.)? Write down the first groups you can think of.

Which groups do you think are considered to be low status in today's Norwegian society?

Employing the same criteria as in question 1, which groups do you see yourself as part of?

In response to question 1, 10% or more of the participants listed the following groups: Muslims (40%), immigrants (37.5%), elderly/retired people (37.5%), young people (25%), Roma people (20%), welfare recipients (17.5%), foreigners (17.5%), students (17.5%), rich people (17.5%), drug addicts (15%), Christians (12.5%), homosexuals (12.5%), disabled / handicapped people (10%), Norwegians (10%), educated people (10%), poor people (10%), unemployed people (10%), and Sami people (an indigenous people in Norway, 10%). Beyond the groups listed in response to question 1, question 2 generated the groups beggars (30%) and people with low education (10%). Question 3 did not generate any new groups mentioned by 10% or more of the participants, which we set as the inclusion criterion.

To create a final list of groups to be included in the main study, we combined immigrants and foreigners into one group and added the following groups based on previous research and theoretical interest: men, women, Jews, housewives, feminists, and middle class people. In total 25 groups were included.

Participants and procedure

The sample for Study 1 consisted of 244 individuals (45.5% male, 50.0% female, 4.5% did not report sex) with a mean age of 35.04 (SD = 16.19, range 16–80). Of the total sample, 231 reported their background. Of these, 89.6% were Norwegian-born, 3.9% had immigrated to Norway, 3.5% were Norwegian-born with one or two immigrant parents, 2.2% declared their background as “other,” and 0.9% stated that they did not wish to provide this information. Among the participants, 7.4% listed compulsory primary and secondary school at their highest level of completed education, 25.8% had completed high school, 36.5% had completed a lower university or college degree (1–4 years), 22.5% had completed a higher university or college degree (5–6 years), and 2.0% had obtained a Ph.D., whereas 5.7% indicated “other” as their educational level or did not provide this information.

Similar to the procedure in the pilot study, potential participants were approached in public places in the city center of Bergen, Norway, and asked to take part in a survey on attitudes to various groups in Norwegian society. They were told that “we are interested in how you think the groups are evaluated by most people, not in how you personally evaluate the groups.” This is consistent with the emphasis on culturally shared stereotypes in the SCM and also servers to mitigate socially desirable responding. Each group was to be evaluated on a total of 33 items (some items are not relevant to the present study). To avoid participant fatigue each participant rated 4–5 randomly selected groups.

Measures

All measures were adapted from previous research (Cuddy et al., 2007, 2009) and translated into Norwegian independently by two of the authors. We then compared translations and agreed on the final version of the questionnaire.

Warmth and competence

For each group the respondents were asked to “think about how [group] are viewed by people in Norway in general. To what extent is [group] considered by most people to be [list of items].” The competence items were: competent, confident, capable, and skillful. The warmth items were: friendly, warm, good-natured, and sincere. All items were responded to on a scale from 1 (Not at all) to 5 (To a very large extent). The average internal reliability coefficient was α = 0.80 (range 0.59 to 0.91) for the competence items and α = 0.86 (range 0.69 to 0.93) for the warmth items.

Status and competition

To measure group status, the respondents were instructed to “think about how [group] are viewed by people in Norway in general. To what extent is [group] considered by most people to have …”: prestigious jobs, economic success, and a good education.

To measure competition, respondents were asked to indicate the extent to which “[Group] gets resources so that other groups in society get less,” “When [group] gets more power, other groups in society get less power,” and “[Group] gets special treatment that makes things more difficult for other groups in Norway.” Again, the respondents were asked to indicate the views of most people in Norway and they rated both the status and the competition items from 1 (Not at all) to 5 (To a very large extent). The average internal reliability coefficient was α  =  0.77 (range 0.43 to 0.95) for the status items and α = 0.82 (range 0.66 to 0.96) for the competition items.

Results

To assess the distribution of Norwegian social groups in the SCM space, a hierarchical cluster analysis was performed employing Ward's method, minimizing within-cluster variance. The distance measure was squared Euclidian distance. This analysis was performed on group-level data; a data set in which each group represents a case (N = 25) and the score on each variable is the mean for that variable across the individual ratings of the group. There is no perfect criterion by which one can decide the optimal number of clusters. However, the coefficients in the agglomeration schedule can be inspected similar to the scree-plot in factor analysis. A marked “jump” in coefficients suggests that two dissimilar clusters have been merged and that the number of clusters prior to the “jump” is a better fit (Blashfield & Aldenderfer, 1988; Ketchen & Shook, 1996). In our data, the coefficients “jumped” when five clusters were merged to four, suggesting that five was the appropriate number of clusters.

Two in-group clusters emerged. The first cluster was rated as very high in warmth and moderate to high in competence (HHW, M/HC). This cluster consisted of the groups women, housewives, elderly, disabled people, Christians, people with little education, gay men, Sami people, and the middle class. The second cluster was rated as moderate to high in warmth and very high in competence (M/HW, HHC) and consisted of the groups Norwegians, men, students, Jews, highly educated people, youth and feminists. Both clusters were above the scale midpoints on warmth and competence (Cluster 1: twarmth(8) = 12.47, p < 0.001, tcompetence(8) = 3.13, p < 0.05; Cluster 2: twarmth(6) = 3.43, p < 0.05, tcompetence(6) = 6.02, p < 0.01).

However, groups in the first cluster were on average seen as more warm (M = 3.67, SD = 0.16) than competent (M = 3.32, SD = 0.30), t(9) = 3.30, p < 0.05, and groups in the second cluster were seen as more competent (M = 3.66, SD = 0.29) than warm (M = 3.12, SD = 0.09), t(7) = –6.27, p < 0.01. Importantly, these two clusters were distinct from each other. The groups in the first cluster were rated as warmer than the groups in the second cluster, t(14) = 7.98, p < 0.001 and as less competent, t(14) = –2.27, p < 0.05.

The third and especially the fourth cluster encompassed typical out-groups. Cluster three contained groups rated as moderate in warmth and moderate to low in competence (MW, M/LC): poor people, unemployed, immigrants, Muslims and welfare recipients. The score on warmth did not differ significantly from the scale midpoint, t(4) = –2.70, p = 0.054, whereas the score on competence was below the scale midpoint, t(4) = –3.07, p < 0.05. The cluster means for warmth (M = 2.86, SD = 0.12) and competence (M = 2.51, SD = 0.40) did not differ significantly, t(4) = 1.94, p = 0.12. Cluster four consisted of groups rated as cold and incompetent (LW, LC): drug addicts, beggars and Roma people. Both warmth and competence scores were below the scale midpoint; twarmth(2) = –7.75, p < 0.05, tcompetence(2) = –14.11, p < 0.01. The groups in this cluster were on average seen as equally low in both warmth (M = 2.11, SD = 0.20) and competence (M = 1.80, SD = 0.15), t(2) = 2.42, p = 0.14. Also the two out-group clusters were distinct from each other; the groups in cluster 3 were seen as both warmer, t(5.65) = 3.84, p < 0.05, and more competent, t(6) = 2.83, p < 0.001, than the groups in cluster 4.

Rich people made up a fifth, single group cluster (LW, HHC) with warmth scores below the scale midpoint, t(38) = –5.70, p < 0.001, and competence scores above the scale midpoint, t(38) = 8.65, p < 0.001. This group was rated as significantly more competent (M = 3.89, SD = 0.64) than warm (M = 2.38, SD = 0.68), t(38) = –12.98, p < 0.001. As can be seen in Fig.1, the clusters/groups spread out across the SCM space, indicating that the warmth and competence judgments served to distinguish group stereotypes also in our Norwegian sample.

Figure 1

Clusters of social groups, Study 1.

Consistent with the SCM, paired samples t-tests revealed that 18 out of 25 individual social groups received ambivalent stereotypes (72%). Means, standard deviations and t-test results are presented in Table1. The correlation between warmth and competence was r = 0.67, p < 0.001 at the individual level1 and r = 0.58, p < 0.01 at the group level.

Table 1

Means, standard deviations and paired samples t-test results, Study 1

	Warmth		Competence
Group	M	SD	M	SD	df	t
Women	3.76	0.64	3.81	0.58	37	–0.60
Housewives	3.88	0.45	3.29	0.51	39	6.64***
Elderly	3.86	0.67	3.28	0.46	38	4.54***
Disabled people	3.77	0.72	2.87	0.50	41	8.63***
Christians	3.60	0.82	3.25	0.67	39	2.85**
People with little education	3.41	0.63	2.88	0.66	39	4.70***
Gay men	3.67	0.63	3.52	0.55	38	1.46
Sami people	3.59	0.85	3.50	0.65	42	0.98
Middle class	3.52	0.49	3.48	0.55	40	0.52
Norwegians	3.23	0.52	3.68	0.50	39	–5.20***
Men	3.18	0.56	3.82	0.57	41	–7.25***
Students	3.09	0.53	3.47	0.48	39	–4.60***
Jews	3.16	0.76	3.75	0.78	41	–4.31***
Highly-educated people	3.18	0.54	4.14	0.54	37	–9.205***
Youth	2.99	0.64	3.24	0.48	38	–2.49*
Feminists	3.01	0.68	3.51	0.75	38	–4.703***
Poor people	3.00	0.76	2.24	0.80	42	6.58***
Unemployed	2.93	0.72	2.38	0.58	37	6.52***
Immigrants	2.89	0.70	2.86	0.52	37	0.46
Muslims	2.73	0.74	2.92	0.61	36	–1.876
Welfare recipients	2.75	0.57	2.13	0.64	39	5.88***
Drug addicts	2.33	0.76	1.90	0.65	38	4.41***
Beggars	2.07	0.72	1.64	0.56	39	4.87***
Roma people	1.93	0.77	1.88	0.71	37	0.59
Rich people	2.38	0.68	3.89	0.64	38	–12.98***

Note:* p < 0.05. ** p < 0.01. *** p < 0.001.

Correlations at the group and individual levels showed that perceived status correlated positively with ratings of competence (r = 0.95, p < 0.001 and r = 0.90, p < 0.001, respectively). Competition correlated negatively with perceived warmth at both the group (r = –0.28, p = 0.177) and individual levels (r = –0.18, p < 0.01), but the correlation was only significant at the individual level.

Discussion

The results from Study 1 support the applicability of the Stereotype Content Model in the Norwegian context. Consistent with the model's predictions, the dimensions of warmth and competence served to distinguish stereotypes of 25 social groups cutting across gender, age, religious, socioeconomic and health categories. Groups' perceived competence was strongly related to their perceived status and groups' perceived warmth was negatively related to their perceived competitiveness, as expected. In line with previous research (Durante et al., 2012; Fiske et al., 2002), the correlations between status and perceived competence were very high. This raises the issue of whether or not status and competence are separate constructs. Although the constructs are strongly related on a statistical level, they are arguably not identical at the conceptual level. Having a prestigious job, economic success, and a good education is not the same as possessing competence-related traits (e.g., being skillful and confident) (Fiske et al., 2002). The competition-warmth relationship was weak, but in line with previous findings (Cuddy et al., 2009; Durante et al., 2012).

Consistent with the SCM framework, three clusters and the majority of individual groups were characterized by ambivalent stereotype content and rated as higher in either warmth or competence. At the same time, warmth and competence ratings were positively correlated, which is in line with findings that the degree of ambivalence tends to be lower in countries with more economic equality (Durante et al., 2012). Moreover, the cluster analysis revealed a clear reference group bias. In-groups and societal reference groups were assessed as both warm and competent, as compared to more moderate assessments of in-groups in more collectivistic cultures (Cuddy et al., 2009).

What is interesting is the lack of a clear high-warmth, low-competence cluster. A look at Fig.1 reveals that groups commonly rated as warm and incompetent (e.g., housewives, elderly and people who are disabled) are located quite close to the theoretical in-group quadrant. We believe that this finding is related to the high level of economic equality in Norwegian society. Based on the GINI index, Norway is ranked as 133 out of 139 (Central Intelligence Agency, 2014), which suggests that it can be considered an extreme case. Durante et al. (2012) found that in more economically equal societies the degree of ambivalence is not only lower, but that the lower levels of ambivalence is primarily due to fewer groups located in the high-warmth, low-competence quadrant. Our result of no high-warmth, low-competence cluster may therefore be a reflection of the extreme case that Norway in this context represents. Thus, our findings fit both the original formulation of the model (Fiske et al., 2002) and later research on cultural and economic influences on stereotype content (Cuddy et al., 2009; Durante et al., 2012).

An interesting finding is that men (M = 3.82) and women (M = 3.81) were rated equally high in competence in our sample. Other studies in Western countries have found that women tend to be stereotyped as lower in competence than men (Germany; Asbrock, 2010; Australia; Durante et al., 2012; USA; Fiske et al., 2002, Study 2). Our findings may be a reflection of the high level of gender equality in Norway (World Economic Forum, 2013) and are consistent with research showing no discrimination in the promotion rates of male and female managers in the Norwegian state bureaucracy (Storvik & Schone, 2008). The stereotypes of traditional women (housewives) and non-traditional women (feminists), however, seem to match previous findings (Asbrock, 2010; Fiske et al., 2002): housewives are less competent and feminists are colder than women in general. The stereotype of men as competent, but only moderately warm, is similar to results from other countries (Asbrock, 2010; Cuddy et al., 2009; Durante et al., 2012; Fiske et al., 2002).

Our findings also reveal which groups are on the margins of Norwegian society: people who are poor, on welfare or unemployed, and people with drug problems. According to Cuddy et al. (2007), groups stereotyped as low in both competence and warmth tend to face both active (e.g., harassment) and passive (e.g., neglect, exclusion) forms of harming behaviors (discrimination). In Norway, low socioeconomic status is associated with discrimination in the housing market (Andersson et al., 2012) and poverty is associated with exclusion from civic organizations (Dahl, Flotten & Lorentzen, 2008). Dahl et al. (2008) suggest that social exclusion may not only result from poverty per se, but be related to subtle barriers such as stigma, being treated with disrespect, and considered inferior. Our findings lend support to this proposition. The “lowest-of-the-low” in terms of stereotype content in Norway is the Roma people. This finding appears to be consistent across many European countries (Durante et al., 2012) and suggests that Roma people also in Norway are at risk of substantial discrimination.

Immigrants as a generic category tend to be perceived as fairly cold and incompetent (Asbrock, 2010; Cuddy et al., 2009; Durante et al., 2012). To some extent this tendency is reflected in our findings, in which “immigrants” were seen as moderate to low in both warmth (M = 2.89) and competence (M = 2.86). When a group falls in the middle of the SCM space the possibility exists that respondents have averaged their perceptions across subgroups of the superordinate category (Fiske et al., 2002). In fact, Lee and Fiske (2006) found that stereotypes of specific immigrant groups in the US varied along the warmth and competence dimensions and reflected both immigrants' nations of origin and the groups' status in the US. In Study 2 we investigate whether Norwegians' stereotypes of immigrant groups are similarly differentiated.

Study 2

Stereotypes of immigrant groups in Norway

Since 1970 the percentage of immigrants in the Norwegian population has risen from 1.5% to 14.1% in 2013. Among the largest groups are immigrants from Pakistan, Iraq, Somalia, Poland and Sweden (Østby & Henriksen, 2013; Statistics Norway, 2013). Immigrants from Pakistan came as “guest workers” in the 1960s and 1970s, whereas immigrants from Iraq and Somalia came to Norway mainly as refugees in the 1990s and 2000s. Immigrants from Poland are typically labor migrants who immigrated after Poland joined the European Union in 2004. Similarly, most immigrants from Sweden come to Norway to work (Østby & Henriksen, 2013; Statistics Norway, 2014a). Status indicators, such as employment status, vary considerably across these groups (Statistics Norway, 2014b). This should lead to differences in perceived competence. Crime rates and reliance on welfare benefits also vary across the groups (Statistics Norway, 2010, 2011) and are common topics in Norwegian media. These are factors which may anchor majority members' perceptions of competition and perceived harmful intent, and thereby influence perceptions of warmth.

Swedes, a group associated with high employment, low crime and low reliance on welfare benefits, should be stereotyped as both warm and competent. Similarly, Polish immigrants have a high level of employment, levels of crime are similar to Nordic immigrants, and their reliance on welfare benefits is low. However, news stories of criminal groups from Eastern Europe (Rud, 2012), and public debates about whether it is fair that Polish migrant workers receive child care benefits for children living in Poland (Meland, 2013; Rønning, 2014), may lead to perceptions of competition and harmful intent. Thus, Polish immigrants are likely to receive ambivalent stereotypes as competent, but cold.

Pakistani and Iraqi immigrants are likely to be stereotyped as moderate to low in both warmth and competence. Employment rates tend to be lower and reliance on welfare benefits and crime rates somewhat higher than in the general population. Moreover, many of the immigrants from these two countries are Muslims and it is likely that the stereotypes associated with them are close to the stereotype of Muslims in Study 1.

Somali immigrants have been shown to receive a disproportionate amount of negative media attention as the immigrant group with “the poorest outcomes across all statistics” (Simonsen, 2007). As a group, Somali immigrants do face substantial unemployment, a higher reliance on welfare benefits, and higher crime rates. Official reports and statistics provide nuanced interpretations of variations across groups (e.g., the role of population composition in the case of crime rates). However, it is reasonable to assume that many people's images of Somali immigrants are shaped by un-nuanced and typically negative media coverage. We therefore expect Somali immigrants to be stereotyped as cold and incompetent.

The aim of Study 2 is to describe and compare stereotypes connected with immigrant groups in Norway within the framework of the Stereotype Content Model. We expect immigrant stereotypes to vary along the warmth and competence dimensions, consistent with competition and status indicators as outlined above.

Methods

A convenience sample from the same geographical area as in Study 1 was gathered for Study 2. The participants filled in questionnaires in their place of work or at home and returned them to the researchers in a sealed envelope. In total each respondent rated 11 groups. Five of these were immigrant groups (Swedish, Pakistani, Iraqi, Polish, and Somali immigrants). We also included Muslims and Roma people and four other groups from Study 1 (Norwegians, elderly, drug addicts and rich people) to provide points of reference and comparison.

The sample (N = 63) consisted of 44.4% males and the average age was 27.10 years (SD = 7.68, range 18–61). Among the participants, 1.6% listed compulsory primary and secondary school as their highest level of completed education, 31.7% had completed high school, 36.5% had completed a lower university or college degree (1–4 years), 23.8% had completed a higher university or college degree (5–6 years), and 1.6% had obtained a Ph.D., whereas 4.8% indicated “other” as their educational level. All of the participants were ethnic Norwegians.

Similar to the procedure in Lee and Fiske (2006) one question measured perceived warmth and competence, respectively. For each group the respondents were asked to “think about how [group] are viewed by people in Norway in general. To what extent is [group] considered by most people to be (a) warm (friendly, good-natured, and sincere) and (b) competent (confident, capable, and skillful)? The items were responded to on a scale from 1 (Not at all) to 5 (To a very large extent).

The status and competition items were identical to Study 1. The average Cronbach's alpha was 0.77 (range 0.52 – 0.90) for the status items and 0.84 (range 0.73 – 0.93) for the competition items.

Similar to Study 1, a group-level data set was created and cluster analyzed (hierarchical cluster analysis, Ward's method). The coefficients in the agglomeration schedule suggested that a four cluster solution provided the most appropriate fit to the data (Fig.2). The first cluster (HHW, HC) was rated as very high in warmth (M = 4.08, SD = 0.20) and high in competence (M = 3.68, SD = 0.12) and consisted of Swedish immigrants and elderly. The second cluster (M/HW, HHC) was rated as moderate to high in warmth (M = 3.17, SD = 0.47) and very high in competence (M = 4.17, SD = 0.15) and included Norwegians and rich people. The third cluster was located in the middle of the SCM space (MW, MC) with moderate scores on warmth (M = 2.74, SD = 0.08) and competence (M = 2.98, SD = 0.25), and consisted of Muslims and Polish, Pakistani and Iraqi immigrants. The fourth cluster (LW, LC) consisted of groups rated as low in warmth (M = 2.20, SD = 0.37) and competence (M = 1.85, SD = 0.06): Somali immigrants, drug addicts, and Roma people.

Figure 2

Clusters of immigrant groups and comparison groups, Study 2.

A more focused comparison of the warmth and competence ratings of the five immigrant groups was made through repeated measures ANOVAs on the individual-level data. The first ANOVA assessing differences across the five groups in ratings of warmth was significant; Wilks' lambda = 0.285, F(4,55) = 34.42, p < 0.001, multivariate ηp2 = 0.72. Pairwise comparisons showed that Swedes were rated as higher in warmth than all other groups and Somali immigrants were rated as lower in warmth than all other groups. The means on warmth for Polish, Pakistani and Iraqi immigrants did not differ significantly (see Table2).

Table 2

Pairwise comparisons of ratings of warmth and competence, Study 2

	Warmth		Competence
Group	M	SD	M	SD
Swedish immigrants	3.94a	0.74	3.76a	0.73
Polish immigrants	2.72b	0.72	3.33b	0.77
Iraqi immigrants	2.74b	0.68	2.79c	0.63
Pakistani immigrants	2.84b	0.72	2.90c	0.69
Somali immigrants	2.16c	0.85	1.92d	0.75

Note:Means within a column that do not share subscripts differ significantly at p < 0.001.

The second ANOVA assessing differences across the five groups in ratings of competence was also significant; Wilks' lambda = 0.251, F(4,57) = 42.61, p < 0.001, multivariate ηp2 = 0.75. Pairwise comparisons showed that Swedes were rated as higher in competence than all other groups. Polish immigrants were rated as more competent than Pakistani, Iraqi and Somali immigrants. The ratings on competence for Pakistani and Iraqi immigrants did not differ from each other, and Somali immigrants were rated as significantly lower in competence than all other groups.

Additional analyses in the form of paired sample t-tests2  showed that 8 out of 11 (73%) of the groups rated in Study 2 were rated as significantly higher in either warmth or competence (i.e., ambivalently). At the same time, perceived warmth correlated positively with perceived competence (rindividual level = 0.63, p < 0.001; rgroup level = 0.72, p < 0.05). Status was positively correlated with ratings of competence (rindividual level = 0.81, p < 0.001; rgroup level = 0.93, p < 0.001) and competition was negatively correlated with perceived warmth (rindividual level =–0.24, p = 0.06; rgroup level = –0.40, p = 0.23). However, competition-warmth correlations were not significant.

The results from Study 2 generally conformed to our expectations. The cluster analysis and the additional analyses (t-tests, correlations) mirrored the results from Study 1 and further validated the applicability of the SCM in the Norwegian context. Stereotypes of specific immigrant groups varied along both the competence and warmth dimensions. As expected, Swedes were rated as warm and competent and Somali immigrants were rated as incompetent and cold. Pakistani and Iraqi immigrants were rated as moderate in both competence and warmth and were located in the same area of the SCM space as Muslims and generic immigrants in Study 1. Also as expected, Polish immigrants were rated as more competent than warm. They were seen as more competent than Pakistani and Iraqi immigrants, but equally low in warmth. Taken together, these results suggest an ethnic hierarchy in Norwegian society, with Norwegians and Swedes at the top, followed by Poles, and then Pakistani and Iraqi (Muslim) immigrants, with Somali immigrants and Roma people at the bottom.

Similar to other studies, this hierarchy reflects cultural similarity and group status (Schalk-Soekar, van de Vijver & Hoogsteder, 2004). The hierarchy implied by our findings also match the groups' position in official statistics on employment, crime and welfare benefits (Statistics Norway, 2010, 2011, 2014b). Moreover, the findings fit with reports showing that experiences of discrimination are generally more prevalent among Somali than Pakistani and Iraqi immigrants across arenas (Statistics Norway, 2008), lending additional support to the idea that stereotype content matters because it is associated with different forms of discrimination (Becker & Asbrock, 2012; Cuddy et al., 2007; Sibley, 2011).

Stereotype content of immigrant groups also matters because it is related to majority members' attitudes to integration. In a recent study in Norway, Phelps, Ommundsen, Turken and Ulleberg (2013) found that majority members become more positive towards making an effort to integrate immigrants when they perceive them to have positive intentions (akin to perceived warmth) and as having the ability to integrate in Norway (akin to perceived competence). Combined with our findings, it seems likely that majority members' attitudes to their own active efforts in the integration of immigrants will vary across immigrant groups. This is an interesting avenue for future research.

General discussion

We have presented two studies that describe the Norwegian intergroup landscape within the framework of the stereotype content model. Results from both studies supported the applicability of the SCM in the Norway and fit well with findings from other studies assessing the model outside the US (Cuddy et al., 2009; Durante et al., 2012). The two studies also provide a detailed picture of the stereotypes associated with a range of social groups in Norway. Beyond addressing the universality of the stereotype content model, mapping the intergroup context in different societies is important in understanding prejudice and discrimination against the same social groups (e.g., gay men, Muslims, Roma people) across countries (Asbrock, 2010). Thus, describing the Norwegian intergroup context contributes to building the knowledge base on how the same social groups are perceived across cultural contexts. It also gives a unique insight into intergroup relations in Norway and may be a valuable foundation for further research addressing prejudice and discrimination in this specific national context.

The present research was funded by the Faculty of Psychology at the University of Bergen, Norway through a summer scholarship granted to Henrik Herrebrøden. We are grateful to Julia Linn Nævdal who contributed in the collection of data for Study 2.