This is an open access article distributed under the terms of the Creative Commons
Attribution License (One of the most prominent psychological perspectives assumes that when perceiving a person we naturally categorise them into a social group (Taylor, Fiske, Etcoff, & Ruderman, 1978). We next associate this person with the set of traits that belong to this group. More generally, stereotypes exist as cognitive structures, such as schemas (Fiske & Linville, 1980), prototypes (Brewer, Dull, & Lui, 1981), or exemplars (Smith & Zarate, 1990). For several decades psychology was mainly interested in the process of stereotyping, and few studies focused on the content of stereotypes – their specific traits (Fiske, Cuddy, Glick, & Xu, 2002). Some scholars, however, noted that aside from the individual perceptions of groups and group members, stereotypes are shared across communities or entire societies, and are thus a collective entity — a part of shared knowledge (Ashmore & Del Boca, 1981; Devine, 1989). Collective stereotypes and intergroup stereotyping processes – how ingroup members perceive other groups – largely shape relationships between groups, but these relationships also seem to be, at least in part, a source of stereotypical content.
The history of intergroup relations shows that collective perceptions have played an important role in shaping several events involving more than one social group. This can be seen throughout the entirety of human history for most conflicts, such as wars, revolutions or any other political upheavals. One of the most vivid examples is the history of anti-Jewish prejudice. For instance, in “In Flaccum,” Philo of Alexandria suggests that one of the reasons for the attack on the Jewish community of Alexandria in 38 CE was the perception of Jews as arrogant, antisocial and misanthropic (van der Horst, 2003). There have been several instances of the persecution of Jews in Europe from the middle ages to the twentieth century, triggered by the accusation that Jews were kidnapping Christian children, which is closely related to the stereotype of Jews as cunning, and of the Jewish religion as dark, occultist and related to the devil (Lambroza, 1981; Poliakov, 2003a; Weinberg, 2008).
Stereotypes can also be a result of the structural relationships between groups, as it was in the case of the Jews living in diaspora. In the Ancient Roman Empire, Jews were perceived as having low intellectual abilities (Daniel, 1979), according to Roman literature, as well as the scriptures of several anti-Jewish writers, such as Apollonius Molon. The fact is that Roman Jews were relatively poor, had low social status and were largely illiterate (Leon, 1930; Poliakov, 2003a). What is also interesting is that an analysis of depictions of Jews throughout history reveals that the stereotype of Jews was far from this picture. The stereotype of the greedy, moneylender Jew emerged in Europe during the Middle Ages. The Church believed that collecting interest from loaned money was a sin, and therefore Christians were not allowed to lend money. Many Jews occupied this economic niche and became moneylenders – bankers and tax collectors. Their involvement with money was to some extent responsible for their reputation for greed, but also successful in business. Competitive financial relations with gentiles also contributed to the idea that Jews are loyal only to their own group (Poliakov, 2003a). In modern America there is a cultural stereotype of Jews as characterised by three major sets of traits: the first related to being powerful, controlling and devious; the second related to uncertain loyalty (USA vs. Israel); and the third to materialistic goals and abilities (Wuthnow, 1982). One of the more prominent modern Jewish stereotypes in American culture, for example, is that of the Jewish lawyer – greedy, exploitative and dishonest, but with a high intelligence (Asimow & Mader, 2004).
The content of the stereotype of Jews changed substantially across time and culture and it was shaped through the relationships of majority and minority groups. Other aspects of the Jewish stereotype across the ages suggest a completely different picture. Many ancient Roman and Greek texts (especially of a satirical nature) refer to Jewish sexuality in connection with circumcision. Jews are depicted as overly sexualised and having animal potency (Schäfer, 2009). This aspect of Jewish stereotyping is extremely interesting because the same contemptuous motives are apparent in medieval stereotypes, especially in association with the devil (Poliakov, 2003b) or Nazi propaganda (Herf, 2006). This suggests that some aspects of stereotyping are stable throughout time and space (Prothro & Melikian, 1954).
The results of social (e.g., psychological or sociological) studies of stereotypes are also not very consistent in describing the stability of the stereotype content. Some of the early works on stereotypes in social sciences were focused on stereotype content and its consistency and stability over time. The results suggested that stereotypes are relatively stable over time (Gilbert, 1951; Karlins, Coffman, & Walters, 1969; Katz & Braly, 1933). Furthermore, results suggest that depending on the context of the relationship with the target group (contact, geographical proximity and familiarity) some of the stereotypes are relatively consistent across both time and space (Prothro & Melikian, 1954).
Stereotypes and Social Structure
More recently, researchers have gone beyond simplistic positive-negative perceptions of social groups and have started to analyse specific stereotype content and its functions (Alexander, Brewer, & Hermann, 1999; Fiske et al., 2002; Fiske, Xu, Cuddy, & Glick, 1999). This approach has not only refocused the interest of psychologists from stereotyping to stereotype content, but also addressed the relationship between the structure of intergroup relations between in- and outgroups, and how groups are perceived (Caprariello, Cuddy, & Fiske, 2009; Cuddy et al., 2009).
The stereotype content model (SCM) proposed by Susan Fiske and colleagues (2002) provides a framework for understanding the relationship between social structural variables and stereotype formation. The model postulates that stereotype content varies along two primary dimensions—warmth and competence—and that the specific content associated with social groups can be ascribed to four possible combinations of warmth (high vs. low) and competence (high vs. low). This four-fold typology further translates into different emotions and discriminatory behaviours (Cuddy, Fiske, & Glick, 2007).
The model further specifies that two structural variables – group status and competition – will predict competence and warmth assessments, respectively, leading to a particular pattern of stereotype content for a given social group or category.
Outgroups are perceived as more competent when they are perceived as powerful and of high status. There are several mechanisms that might be responsible for this connection. Gilbert and Malone (1995) described a correspondence bias, a tendency to infer people’s traits based on their situation and behaviour. Another explanation of this relationship might be related to the process of developing and maintaining (self-reinforcing) power-status hierarchies (Berger, Rosenholtz, & Zelditch Jr., 1980). The connection between status and competence might be an outcome of perceiving the world as a fair place where individuals (Lerner & Miller, 1978) and groups (Jost & Banaji, 1994) receive what they deserve.
The second characteristic of social structure that shapes stereotypes is competition. Outgroups are perceived as warm and friendly versus cold and hostile depending on whether there is a perceived or actual competition between groups. Several studies have described mechanisms leading to negative stereotyping, developing negative emotions and prejudice or discrimination toward groups that have competing goals (Esses, Dovidio, Jackson, & Armstrong, 2001; Falomir-Pichastor, Muñoz-Rojas, Invernizzi, & Mugny, 2004; Sherif, Harvey, White, Hood, & Sherif, 1961). A major source of negative affect toward outgroups is a result of the perceived incompatibility of their goals with in-group goals (Fiske & Ruscher, 1993). On the other hand, an outgroup will be perceived as relatively warm if there is no competition. Moreover, subordinate groups, as long as they do not compete with the ingroup, will be perceived as not competent, but warm (Glick & Fiske, 2001).
These predicted relationships were observed in several correlational studies across a wide range of social group stereotypes (Cuddy et al., 2007; Cuddy et al., 2009; Fiske et al., 2002; Fiske et al., 1999). Causal relationships were also established in an experimental study in which manipulation of the social structure in specific scenarios about unfamiliar ethnic groups had an impact on evaluations of warmth and competence (Caprariello et al., 2009).
Stereotype Content and Collective Memory of Historical Relations
As noted above, studies on stereotypes initially focused on cognitive processes. There is a large body of literature on how cognitive representations are developed and which cognitive mechanisms sustain stereotypes (McGarty, 2002; Stangor & Lange, 1994). We learn to categorise as children (Bigler, 1995) and the content of stereotypes comes from our surroundings – parents, media, peers and culture (Stangor & Schaller, 1996). Several mechanisms explain and reinforce the process of learning stereotypes from social and cultural surroundings, such as social learning, conformity or taking on social roles (Mackie, Hamilton, Susskind, & Rosselli, 1996).
Some theorists have underlined the importance of both the context of interaction, perceiving behaviour, and theories, understood as higher-order knowledge structures that are much broader than categories or schemas (Brown & Turner, 2002). These theory level perceptions might be especially important in the context of ethnic stereotypes where broader contexts related to ideology, world view or knowledge of history might play a crucial role.
When thinking about historical relations we have to acknowledge that the ingroup’s perception of history is a social construct. Collective memory might involve events, places, situations, and as we hypothesise, descriptions of outgroups, that happened or existed before our own lives (Paez, Besabe, & Gonzalez, 2013). Collective memories are especially strong in cases of strong emotional, traumatic events, significant changes or when they are identity related (Conway, 2013; Liu & Hilton, 2005; Pennebaker & Banasik, 2013). What seems especially important in the context of intergroup relations and stereotype content is that collective memory is biased in favour of the ingroup. Memories related to ingroup history are stronger and perceived as more important than those related to outgroups (Liu, Wilson, McClure, & Higgins, 1999). Oliver Klein and colleagues (Klein, Licata, Van der Linden, Mercy, & Luminet, 2012) proposed a theoretical model connecting collective memories with justice principles, differently applied to the ingroup and outgroup, depending on the context. These principles in turn predict dimensions of stereotype content. The authors proposed a complex model in which collective memories of past conflict (violations of justice principles) validate stereotypes of the outgroup by presenting them as anchored in the past. Additionally, collective memories and stereotypes shape the social context influencing current collective understandings.
Following the suggestion of Susan Fiske and colleagues (Fiske et al., 1999; Fiske et al., 2002), we hypothesise that part of the variance in the content of cultural stereotypes is shaped by the collective memory of intergroup relations. To our knowledge, no studies have tested the link between stereotype content and history. However, there is some direct evidence that the history of intergroup relations, and specifically the length of intergroup conflict, has a substantial impact on perceptions of outgroups. Studies on social threat have shown that groups that have a long history of conflict are most prone to perceiving intergroup threats (Stephan et al., 2002). Some studies have connected historical roles in conflict and historical narratives to perceptions of outgroups who were other actors in these events (Bilewicz, 2007; Bilewicz & Jaworska, 2013; Imhoff et al., 2017).
Polish Historical Context
Modern Polish history is rich in dramatic events. Firstly, 123 years of partition between three great European powers took place, which ended during World War I. There were then twenty years of an independent state filled with political instability and violence, and eventually a military coup. Poland was the first country attacked in World War II, and suffered substantial human and material losses. Most of the Holocaust took place on Polish soil (Hilberg, 2003). After the war, Poland became a communist country and one of several states dependent on the Soviet Union, which lasted until a peaceful transition in 1989 (Davies, 2005b). These events all had a substantial impact on Polish narratives of history, but the biggest impact on intergroup and interethnic relations was WWII. There are two major reasons for this. First, one of the biggest Polish minorities – Jews – were almost entirely wiped out during the war. Accompanied by large waves of migration in the 1940s, 1950s and 1960s, the Polish Jewish population fell from around 10% (3 million people) to less than 0.001 (seven thousand people) (Gudaszewski, 2015; Tomaszewski, 1990). Secondly, due to post-war treaties, Poland lost around one third of its eastern pre-war territory, and as compensation, parts of the west and north German state were incorporated. These territorial changes triggered a mass relocation of people, mainly Germans, out of Poland, and Poles from the eastern parts to the so-called “regained lands” in the west. The change of borders also moved the majority of other minorities, mainly Ukrainians, Byelorussians and Lithuanians, out of the Polish territories. These two factors shaped contemporary Poland into a very ethnically homogenous country; in a recent census, over 94% of the Polish population declared their ethnic identity as only Polish (Gudaszewski, 2015).
Historically, interethnic relations in Poland were very diverse. Regions of Poland were highly diverse in terms of ethnic composition and the status of the groups (Davies, 2005a, 2005b). For instance, almost all Ukrainians lived in four south eastern voivodeships (basic administrative areas in Poland), including the Wolyn and Stanisalawow regions, where they constituted over two-thirds of the population (Szturm de Sztrem, 1938). A Jewish minority lived mainly in cities and small towns in central and eastern Poland, but there were almost no Jews in the north-west counties, where almost all the German speaking population lived (Szturm de Sztrem, 1938). In addition to a different distribution of the minority population in the Polish regions, the relative status of those minorities substantially differed. For example, Ukrainians were of a relatively low status and were discriminated against – for example, through land reform (Snyder, 2003). The economic and social status of Jewish communities varied depending on several factors, including the level of urbanisation, which partition the region belonged to, and its legal status (Eisenbach, 1988; Szturm de Sztrem, 1938).
Trace of the Past – Regional Differences in Stereotype Content
World War II split and entirely transformed interethnic relations in Poland. These changes made the country ethnically homogenous, and so made it difficult to change stereotypes via contact with outgroup members (Pettigrew & Tropp, 2008). Currently, the content of stereotypes is almost entirely dependent on associations and cultural stereotypes. The arguments presented in the previous sections lead us to hypothesise that 1) the contemporary content of ethnic stereotypes varies regionally, and that 2) the regional variance of content is related to historical intergroup relations, that is, the status of those groups throughout the history of specific regions of Poland and the role they played in the life of the community.
Method
Study Data
In order to test our hypothesis we decided to reanalyse existing data. The major problem with testing a regional hypothesis using nationwide representative data is that samples are not representative of regions (Dorofeev & Grant, 2006), and that due to large differences in population sizes between regions, some subsamples are very small. In most of the studies on nationwide representative samples, the problem lies in the procedure that ensures representativeness for several variables independently (e.g., education is independent from region). However, all of the studies that we combined were conducted using random sampling where the problem of representativeness for regions concerns small subsamples that produce unstable estimators. Therefore, combining several studies is, to some extent, a solution that allows for regional comparison. We found only two representative datasets that utilised questions directly measuring stereotype content in the SCM framework. Therefore, we decided to analyse coded data from open-ended questions. In order to test our predictions, we re-analysed the data from five consecutive nationwide survey studies, conducted on a representative sample of adult Poles. The first author was involved in planning four of the five studies. Using this data we created a joint database and conducted a substantial part of the analysis on the merged data file. For that reason, we do not go into a great amount of detail when describing the five separate studies – especially since their methodology was very similar – and we differentiate between them only when necessary. We did not use weights, and therefore some characteristics of individual datasets might differ from the original reports (Bilewicz, Bukowski, Cichocka, Winiewski, & Wójcik, 2009; Krzemiński, 2004; Winiewski, 2010).
Participants
The five studies were conducted on representative samples of adult Poles. In Study 1, N = 1085 adults participated, amongst them 608 men and 477 women, aged between 18 and 94 (M = 45.10, SD = 17.49). In Study 2, N = 891; 442 men and 449 women, aged between 18 and 90 (M = 46.62, SD = 18.38) took part. In Study 3, N = 931 Poles participated; 437 men and 494 women, aged between 19 and 91 (M = 48.74, SD = 18.04). The sample of Study 4 consisted of N = 948, 448 men and 500 women, aged between 19 and 92 (M = 48.93, SD = 18.46). Lastly, in Study 5, N = 979 Poles participated, amongst them 445 men and 534 women, aged between 18 and 89 (M = 48.21, SD = 18.03). The overall sample consisted of N = 4834, however due to missing answers, the analyses were conducted on a smaller sample.
Procedure
All studies were large, nationwide survey studies conducted on random samples based on PESEL numbers (Powszechny Elektroniczny System Ewidencji Ludności – national identification number), and their procedure was to a great extent similar. Study 1 was conducted in 2002 by the team of Ireneusz Krzemiński (2004) and devoted to studying antisemitism in Poland. Fieldwork was conducted by the Partner in Business Strategies polling agency. Studies 2-4 were dedicated to the topic of social dilemmas and were a part of the “Omnibus Survey” - survey studies conducted monthly by the Polish Public Opinion Research Centre. Study 2 was conducted in December 2004, Study 3 in February 2007 and Study 4 in March 2007. The participants were interviewed personally by a skilled pollster and their answers were noted on a paper version of the questionnaire. Study 5 was the first edition of the Polish Prejudice Survey – a study planned by the Centre for Research on Prejudice and executed by the Public Opinion Research Centre (Bilewicz et al., 2009). It was conducted in May 2009, using the computer-assisted personal interview (CAPI) method. In all five studies, after being presented with a number of questions about various issues, the participants were asked a series of open-ended questions regarding the characteristics that, in their opinion, are typical of representatives of several concrete ethnic groups. In Study 1 the groups were Roma, Russians, Germans, Ukrainians, Poles and Jews, in Study 2 Germans, Russians, Poles, Vietnamese and Jews, in Studies 3-4 the groups were Germans, Poles, Russians, Vietnamese and Jews and in Study 5 the groups were Germans, Poles, Russians, Belarusians, Ukrainians, Vietnamese, Jews and Roma. In Study 5, thanks to the use of computer-assisted interviews, the order of the groups was randomised.
Measures
As all five studies were nationwide surveys measuring the attitudes of Poles regarding several social issues, they included a list of measures that go beyond the scope of this article. We describe only the questions used for the analyses presented in this paper.
Characteristics of various ethnic groups were measured with an open-ended question. In Studies 1-4 the question read: “Many people of a different nationality than Polish live in our country. Those people often have different customs, languages, and rules of behaviour. We would like to know which traits, in your opinion, characterise the typical…” In Study 5 the question had a slightly different format: “Individual nationalities have various customs, languages and rules of behaviour. We would like to know what traits, in your opinion, characterise the typical…” The participants were asked to provide up to eight characteristics for each ethnic group. The answers to these questions served as a basis for creating an index of warmth and competence for each of the groups. The process of creating the indexes is described below.
Region of residence was established based on the PESEL population list that was used for drawing the sample. The unit used in the analysis was the voivodeship, which is an elementary administrative area in Poland. Currently there are 16 Polish voivodeships.
Construction of the Indexes of Warmth, Competence and Ambivalence
The process of constructing the indexes of warmth and competence involved several steps. Firstly, an extensive data file was created by merging the answers to the open-ended questions about the characteristics of various ethnic groups. The extensive database of initial entries was then organised into a list of 629 preliminary semantic categories. Synonymous categories that were mentioned less than 20 times were combined using the Polish Dictionary (Szymczak, 1995). For combined categories label was taken from the most frequent of the combined categories. This resulted in a list of 339 unique traits. We asked five competent judges (three men, two women, age range 25-35, three people with higher education and two with secondary education) to assess the 339 traits regarding the dimensions of warmth and competence, using a scale from -2 (not competent/cold) through 0 (neutral) to 2 (competent/warm). The judges were given a definition of both dimensions based on a review paper by Fiske, Cuddy, and Glick (2007). Inter-rater reliability, measured by the intra-class correlation coefficient (ICC) (McGraw & Wong, 1996; Shrout & Fleiss, 1979), was very good for both the dimensions, especially for the warmth dimensions: warmth ρI = 0.917; competence ρI = 0.810. Lastly, we created the indexes of warmth and competence by averaging all the judge’s assessments. Using the indexes of warmth and competence, we computed an index of ambivalence of the stereotype, by subtracting the index for competence from the index for warmth. A positive value for this coefficient indicates a more paternalistic stereotype of a group, whereas a negative value indicates a more envious stereotype (Fiske et al., 2002). If the index equals zero, it means that the stereotype is consistently positive or consistently negative.
The created indexes were then copied back to the original five data files. The final database was created by combining information about the source (no. of the study), participants’ ID, the demographic variables and the indexes of warmth, competence and ambivalence. All of the analyses were conducted on the merged database. In all subsequent models the source of the data was controlled (dummy coded) and did not yield any significant results. No other harmonisation procedure was used.
The List of Characteristics
Some additional analyses were performed in order to obtain a more in-depth understanding of the differences in the content of stereotypes between the particular areas of Poland, and we additionally semi-qualitatively analysed the frequencies for characteristics provided by the participants. The 339 traits were collapsed using a thesaurus (Dabrowka, Geller, & Turczyn, 1993). This resulted in 176 unique traits.
Results
Although the database contained data regarding the content of stereotypes for several ethnic groups, we focus on two specific groups in this paper: the Jews and the Germans. This choice is primarily related to the fact that only these two groups have been present in Polish society throughout its history, and their role, status and function varied between specific regions of the country.
Descriptive Statistics
As can be seen in Table 1, in general the Germans were perceived as slightly warm and rather competent, whereas the Jews were perceived as slightly cold and rather competent. This was consistently observed throughout all of the five studies. The negative value of the index of ambivalence for the Germans and the Jews shows that the stereotype of both of these groups is relatively envious and that the disparity between warmth and competence is greater when it comes to the Jews, t(5790) = -5.74; p > .001. It is noteworthy that the standard deviations for all the indexes were always greater than their mean, suggesting that there is a great variety in the way individuals perceive representatives of the two ethnic groups.
Table 1
Descriptive Statistics for the Indexes of Warmth, Competence and Ambivalence for Jews and Germans
| Index | Study 1
|
Study 2
|
Study 3
|
Study 4
|
Study 5
|
Overall
|
||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | |
| Warmth Germans | 0.38 | 0.88 | 0.21 | 0.89 | 0.32 | 0.87 | 0.46 | 0.75 | 0.31 | 0.83 | 0.35 | 0.84 |
| Competence Germans | 0.71 | 0.66 | 0.56 | 0.62 | 0.64 | 0.63 | 0.79 | 0.55 | 0.67 | 0.63 | 0.68 | 0.62 |
| Ambivalence Germans | -0.33 | 0.50 | -0.35 | 0.53 | -0.32 | 0.50 | -0.33 | 0.45 | -0.36 | 0.45 | -0.34 | 0.48 |
| Warmth Jews | -0.06 | 0.87 | -0.02 | 0.88 | -0.01 | 0.85 | -0.04 | 0.77 | 0.09 | 0.73 | -0.01 | 0.81 |
| Competence Jews | 0.47 | 0.63 | 0.45 | 0.60 | 0.48 | 0.60 | 0.47 | 0.57 | 0.52 | 0.54 | 0.48 | 0.58 |
| Ambivalence Jews | -0.53 | 0.55 | -0.47 | 0.59 | -0.49 | 0.57 | -0.52 | 0.46 | -0.43 | 0.50 | -0.49 | 0.53 |
It is notable that the indexes of warmth and competence for both the groups were highly correlated. The correlation for the overall sample equalled r = .76, p < .001 for the Jews and r = .80 p < .001 for the Germans.
Analysis of Indexes of Warmth, Competence and Ambivalence by Region
In order to analyse the differences between the inhabitants of the 16 Polish voivodeships in regard to the perceived stereotypes of the two ethnic groups, we conducted a series of regression analyses, consecutively treating the indexes of warmth, competence and ambivalence as dependent variables and the voivodeship of residence as a set of effects-coded independent variables. We used effects coding instead of simple dummy coding because it allows for a comparison of the mean for a given level of the variable to the grand mean for all of the categories. We were therefore able to firstly avoid the problem of arbitrarily preselecting the reference category and, secondly, to make comparisons for every voivodeship. The properties of consecutive models are presented in Table 2, and the results of the regional analysis are presented in Figures 1, 2, 3, 4, 5, and 6.
Table 2
The Properties of Regression Models for the Indexes of Warmth, Competence and Ambivalence for Jews and Germans
| Statistic | Germans
|
Jews
|
||||
|---|---|---|---|---|---|---|
| Warmth | Competence | Ambivalence | Warmth | Competence | Ambivalence | |
| R2 | 0.02 | 0.04 | 0.01 | 0.03 | 0.03 | 0.02 |
| F | 3.77*** | 7.39*** | 1.19 | 8.24*** | 5.53*** | 4.71*** |
***p < .001.
Figure 1
Perceived warmth of Germans by region.
Note. Grey colour indicates regions with low average but not significant coefficients.
As shown in Figure 1, the Germans were perceived as more warm than average in four voivodeships: Lubelskie (B = 0.19, SE = .07; p < .004), Łódzkie (B = 0.17, SE = 0.06; p = .002), Świętokrzyskie (B = .31, SE = .10; p < .001) and Kujawsko-pomorskie (B = 0.13, SE = 0.05; p = .021), and as less warm in the following three regions: Lubuskie (B = -.24, SE = .09; p = .048), Wielkopolskie (B = -.16, SE = .06; p < .001) and Zachodniopomorskie (B = -.31, SE = .08; p < .001). The perceived warmth of this group in every other region did not differ significantly from the grand mean. The average in Pomorskie and Warmińsko- Mazurskie was relatively low.
Figure 2
Perceived competence of Germans by region.
The perceived competence of Germans was higher than average in the Lubelskie voivodeship (B = 0.10, SE = 0.04; p = .006), Łódzkie voivodeship (B = .11, SE = .03; p < .001), Małopolskie voivodeship (B = 0.09, SE = .04; p = .041), Świętokrzyskie voivodeship (B = .30, SE = .06; p < .001) and Świętokrzyskie voivodeship (B = .30, SE = .06; p < .001). Perceived competence was lower than average in the following regions: Lubuskie (B = -.31, SE = .07; p < .001), Pomorskie (B = -.11, SE = .05; p = .011), Warmińsko-mazurskie (B = -.14, SE = .06; p = .049), Wielkopolskie (B = -.09, SE = .04; p = .035), and Zachodniopomorskie (B = -.20, SE = .05; p < .001).
Finally, although generally the stereotype of Germans is rather envious (which is indicated by the grand mean), the regional analysis for the “ambivalence” index did not yield significant results and we did not observe any significant differences.
Figure 3
Perceived warmth of Jews by region.
Note. Grey colour indicates regions with low average but not significant coefficients.
As presented in Figure 3, Jews were perceived as warmer than average only in Małopolskie (B = 0.07, SE = .04; p = .043) and Lubuskie (B = 0.14, SE = .06; p = .032), and less than average in the Łódzkie voivodeship (B = -0.17, SE = 0.04; p < .001), Lubelskie (B = -0.10, SE = 0.04; p = .042), Swiętokrzyskie (B = -0.13, SE = 0.06; p = .030) and Podkarpackie (B = -0.10, SE = 0.05; p = .041). None of the other regions differed significantly from the grand mean with regard to the perceived warmth of Jews, but in Podkarpackie the mean was relatively low.
As depicted in Figure 4, he competence of Jews was perceived as higher than average in the Dolnośląskie voivodeship (B = 0.08, SE = 0.02, p = .003), Łódzkie (B = 0.06, SE = 0.02; p = .049), Śląskie (B = 0.05, SE = 0.02; p = .015) and the Małopolskie voivodeship (B = 0.10, SE = 0.02; p < .001). There were stereotypes of Jews characterised by lower competence in Podkarpackie (B = -0.08, SE = 0.03; p = .007) and Swiętokrzyskie (B = -0.10, SE = 0.04; p = .010). None of the other regions differed significantly from the grand mean, however in Lubelskie the mean was relatively low.
Figure 4
Perceived competence of Jews by region.
Note. Grey colour indicates regions with low average but not significant coefficients.
Lastly, in general the stereotype of Jews in Poland is envious (which is indicated by the grand mean). Jews are perceived as more envious than average in Łódzkie (B = -0.18, SE = 0.03; p < .001) and significantly less envious in Lubuskie (B = 0.18, SE = 0.05; p = .001) and Zachodniopomorskie (B = 0.09, SE = 0.06; p = .030) as shown in Figure 5.
Figure 5
Enviousness of the stereotype of Jews by region.
Analysis of the Most Popular Characteristics
In order to see the differences in the actual content of the stereotypes of Jews and Germans, we conducted a semi-qualitative analysis of characteristics, looking at the traits that were most frequently provided by the participants for both ethnic groups. As the general stereotype of Jews and Germans in Poland is relatively envious, meaning that they are perceived as more competent than warm (in the case of Germans) and competent and rather cold (in the case of Jews), we wanted to see the specific characteristics that were responsible for the creation of this particular image of the groups. The overall results for both groups are presented in Figures 6 and 7.
Figure 6
Word cloud for the 35 most frequent traits of the Jewish stereotype.
Note. Size of the trait represents its relative frequency.
Figure 7
Word cloud for the most frequent traits – Germans.
Note. Size of the trait represents its relative frequency.
As can be seen in the figure above, the five most frequently provided traits for Jews were: greedy/stingy (9.34% of the participants mentioned this characteristic at least once), religious (8.88%), involved in trade (8.02%), resourceful (6.88%), clever/sly (6.76%) and good in business (6.63%).
The most frequently provided traits for Germans were: reliable and punctual (21.3%), hardworking (20.3%), tidy (18.7%), frugal (11.7%) and proud (11.7%).
Characteristics by Region
We looked at specific characteristics by region. There are 16 voivodeships and more than 100 traits can be analysed on multiple levels, such as the frequencies or proportions within a region, comparisons across regions, or in comparison to overall proportions. In the analysis we focused only on those regions that deviate significantly from the others in previous analyses. We report only those traits that had a higher frequency compared to other regions or a substantially higher frequency in a region.
We focused on Lodzkie when looking at specific traits for stereotypes of Jews, where the stereotype was the most envious; Lubuskie and Zachodniopomorskie, where the stereotype was relatively positive and relatively non-envious; and Lubelskie and Podkarpackie with the most contemptuous stereotypes – negative for both competence and warmth. Although "stingy" dominated in all provinces, it was more often mentioned in Łódzkie – 12.76% (more than in every other voivodeship, 8.91%). The situation with “good in business” and “greed” was similar. These two traits were mentioned more in Łódzkie than in any other region. Two specific traits, “money” and “usury”, were particularly mentioned in Łódzkie; 48.94% and 20.69% of all mentions, respectively. Łódzkie was unique because respondents did not describe the Jewish minority as religious, with 4.73%, in contrast to the rest of the regions, where on average 9.25% of respondents mentioned this trait. As Zachodniopomoirskie and Lubuskie were very similar, they were merged for the purposes of this comparison. In general, the traits mentioned by the residents of these regions were similar to others, however some positive traits were mentioned slightly more often, especially “hardworking”, “resourceful”, “honest” and “thrifty”. There were also large similarities between Podkarpackie and Lubelskie. The traits that were relatively specific to these regions were “exploitation”, “stingy” and “fraud”, and the terms “intelligence” and “education” were relatively rare.
In order to explore the stereotypes of Germans, we picked three regions with stereotypes that were relatively low in both competence and warmth: Wielkopolskie, Zachodnipomorskie and Lubuskie, and three regions with relatively positive stereotypes: Łódzkie, Lubelskie and Świętokrzyskie. In the three negative regions, Germans were less often characterised as “wealthy”, “hardworking”, “solid”, “clean” or “orderly”, the latter especially in Lubuskie and Zachodnipomorskie. The same traits were mentioned relatively often in Łódzkie, Lubelskie and Świętokrzyskie. What seems clear is that in all three western regions references to “past intergroup relations” were more often listed, and directly negative traits like “vengeful“ and “bad/evil” were also more often used, especially in Zachodnipomorskie and Lubuskie.
Discussion
The general results show that both stereotypes can be characterised to some extent as envious – having relatively higher competence than warmth (Fiske et al., 2002). Although we could not refer the stereotypes to a broader map based on previous studies in our analysis, we can infer that in fact the stereotype of these two groups in Poland is envious (Bilewicz et al., 2009; Cichocka, Winiewski, Bilewicz, Bukowski, & Jost, 2015; Winiewski, 2009). Similarly to previous studies, the stereotype of Germans seems on average to be more positive – based mainly on warmth and to a lesser extent on competence. This general result in itself might be interesting. The history of Polish- German relations has been turbulent and over the last two hundred years Poland and the Poles have been oppressed (e.g., two of three partitions were Germanic - Austria and Prussia - in World War II). The envious Jewish stereotype could also be related to historical structures. Due to medieval laws specifying where and how many Jews could settle, they occupied a social and economic niche that placed them in direct competition with serfs and free peasants and also in competition with the non-Jewish city population. In the sixteenth and seventeenth centuries, Jews dominated interstate trade, especially between villages and cities, and they were tax collectors, tenants and artisans (Eisenbach, 1988). The dynamic social and economic changes of the industrial revolution and emancipation changed the structure of Jewish communities. Jerzy Tomaszewski (1990) describes the interwar Jewish community employment structure as concentrated in small trade, small manufactory and workers, however these three major branches were all concentrated in specific areas of the economy. This specificity of the Jewish situation was an outcome of several factors, including partitions, geographical factors and traditional ways of living. Although the actual economic situation of the Jewish community was very difficult (it was vulnerable to crisis and very conservative) this described asymmetry and specificity was a strong basis for economical antisemitism (Rudnicki, 1985; Tomaszewski, 1990). Another aspect of the Jewish community worth underlining is that even at the end of sixteenth century the majority of the Jewish population was literate (Eisenbach, 1988). It therefore seems that in the case of both groups there is a historical root for the envious stereotype that is in line with the competence-status and warmth-competition hypothesis. The difference between the overall stereotypes of Jews and Germans also seems to correspond to differences in both historical and contemporary relations. The Jews were only a minority in contrast to the Germans, who were concentrated in cities, and were also associated with a large and successful neighbouring state or states. Contemporary relations also explain the differences. As we mentioned, intergroup relations with Jews virtually stopped during WWII, while relations with neighbouring Germany were ongoing and underwent serious changes, from open hostility during war to very warm, at least at the state level. Now both Poland and Germany are members of the EU and allies within NATO. Most of the major traits attributed by Poles to Germans are related to high competence and good organisation, which might explain their economic success and previous successes in dominating Poland. Some of the similar traits in the Jewish stereotype have slightly more negative connotations, however, and there are many associations with trade, business and education.
Regional differences, although not very strong, form a pattern that can be linked to historical differences. Jews are perceived substantially more negatively in eastern Poland, especially in the south-east. As we mentioned before, interethnic relations were very diverse and different in various parts of Poland. In the eastern part of contemporary Poland (together with the part that now belongs to Ukraine) the Jewish community was large, poor, very traditional and relatively uneducated (Eisenbach, 1988; Tomaszewski, 1990). These parts were also the scene of strong ethnic tensions, and among the Polish population there was relatively strong support for Narodowa Demokracja, a radical nationalistic movement. Most of the local government and church officials have supported extreme right wing political parties, and the press, which has the biggest impact on a community, have had a nationalistic focus (Kopstein & Wittenberg, 2018; Rudnicki, 1985). This strong Polish-Jewish conflict and the relatively low status of the Jewish community is true for all of eastern Poland (including contemporary Podlaskie). There was a substantial difference between south-east and north-east Jewish communities. In the northern parts – including contemporary Lithuania – the majority of Jewish community were Litvaks. This is a specific Jewish community who, in contrast to other eastern Jews, retained a strong connection to rabbinical Judaism, cherishing the traditional intellectual character of the religion (Ozer, 2009). This might explain why this stereotype is low in warmth in the entire eastern area, but it is relatively high regarding competence in north-eastern parts. The high competence in Łódzkie, Małopolskie and Śląskie is related to the high urbanisation and industrialisation of these regions, and that the Jewish communities there were relatively well off, successful and educated (Eisenbach, 1988; Trojański, 2005). The most interesting area in this context seems to be Łódzkie, however, which differs most from the other regions. Here the content of the Jewish stereotype is both low in warmth and high in competence, and additionally the most envious. One possible explanation for this phenomenon is that before WWII Łodzkie was a developing industrial area (textile industry)- the “promised land”. The dynamically developing economy equalized the chances of success for various social groups and nationalities. The Jewish community in Łodź was relatively rich and successful (Landau & Tomaszewski, 1989).
Analysis of the results for regional differences in the stereotyping of Germans shows that northern and north-western Poland differ the most from the other regions. The relatively negative stereotype in both dimensions is specific to this region. There are two aspects in the history of Polish-German relations that can explain the findings for these regions. The first is related to historical struggles, especially in Pomorze and Wielkopolska. These lands historically belonged to Poland and experienced strong Germanisation during partition. Wielkopolska was the scene of several uprisings against Prussia, with the last triggered by the end of WWI that led to the incorporation of Wielkopolska in Poland. This strong historical competition would be in line with our initial hypothesis. The second explanation deals more with the so-called regained lands – the regions that were incorporated into Poland based on post-WWII treaties. These lands were settled by a mixture of people from central Poland and the parts incorporated into the USSR. Migrations and forced resettlements changed almost the entire population of these regions (Dominiczak, 1975; Machałek, 2016). Marcin Zaremba (2012) notes that in the immediate post-war era, the major emotional theme of these lands was a fear of Germans coming back and reclaiming their land and possessions. The issue of restitutions for German possessions is an important part of the local and national debate that still elicits strong emotions.
We adopt a historical explanation as most probable due to the clear overlap of regional differences in stereotypes with a historical pattern and lack of obvious alternatives in contemporary intergroup relations. The case of the Jewish minority is very specific and, as we mentioned, is virtually non-existent. Therefore, individual interactions between Poles and Jews are very rare. Most of the contacts of Poles with Jews and Jewish themes are mediated by the media and related to political debates and therefore are not dependent on the region. In the case of Germans, the situation is different. The minority is slightly bigger, but still under one percent and unequally distributed across Poland (Adamczuk & Lodzinski, 2006). Germany is a neighbouring country, and therefore the proximity of its border increases chances for interaction. Also, many Poles have migration experience and Germany has, historically, been a highly frequent destination for migration (Gudaszewski, 2015; Nowosielski, 2012). Importantly, the migration is differentiated regionally (Adamczuk & Lodzinski, 2006). However, none of the mentioned factors, that is, proximity of border, percent of German minority or relative frequency of migration to Germany, overlap with the stereotype pattern.
Limitations and Conclusions
The results of this study should be treated with caution, primarily because the data was collected as a representative sample of the entire population of Poland and was not meant to be used for regional comparisons. As we mentioned earlier, this results in problems for the overall inferences. The models presented account for some key demographic variables, however the representativeness is on the level of country, not region. The second problem is that all our models are significant, but the explained variance is relatively small. We do not expect very strong effects of historical variables on contemporary attitudes, however the low explained variance might also be an outcome of the procedure in which we transformed qualitative questions in two steps.
We think that the historical relations offer the most obvious and coherent interpretation of regional differences in the stereotype of the Jews and Germans. However, one needs to acknowledge that alternative explanations are possible. Due to the archival nature of the data and the correlational analysis, we can only speculate about the meaning of the differences.
The results show that the stereotype content of ethnic groups is not uniform across a whole society. Physical difference is one element differentiating the content of stereotypes, and it seems that the local historical experience of a population shapes the perceptions of the outgroup. These past historical intergroup relations can be seen today in the framework of stereotypes.