Go to main content


Unboxing people: How to de-construct clichés and counter prejudice in reporting

Social categorisation, prejudice and the role of the media: Relevant research findings and scientifically solid tools and tips for editorial practice to responsibly counter social divisions and hate. Part 9 of the "Psychology for Journalists'' article series

By Katja Ehrenberg and Margarida Alpuim

Diverse Figuren in allen Formen und Größen repräsentieren die Diversität von Menschen

Introduction and overview

When I, Katja, ask my students in social psychology courses how prejudices emerge, they will often name “the media” as an important factor, along with upbringing and peers. Movies and TV shows especially old ones are often teeming with clichés. But journalism also plays a role: news reports are never just portrayals of “reality”, but are inevitably a selection and reflection of what the journalist or media company considers “newsworthy”. And the way media set and frame issues has a direct impact on the thoughts, feelings and actions of the audience (1). My students often also emphasise that news and documentaries contribute to perpetuating stereotypical ideas, e.g. through the labels they use or narratives that blame the victim.

The media co-create reality: They may inspire and empower, but their stories can also fuel social divisions, hate speech and crimes between and among social groups, up to and including genocide between social groups (2). The challenge for media professionals is to inform in a way that is comprehensive and nuanced, critical and complete, but never dehumanising, never implicitly feeding xenophobia, misogyny, racism or anti-religious hatred — let alone explicitly, as happens in propaganda or disinhibited social media channels (3).

This is not new. But while journalists often talk about fighting to counteract stereotypes and prejudices, they may not even be aware that they themselves (and their audiences) are being impacted by implicit biases. Deeper knowledge on the psychology of categorisation, stereotyping, prejudice, discriminatory behaviour; the consequences of these; and well-researched possible solutions could help. The topic is complex, as it involves many different psychological processes that are interconnected.

In this article…

  • We present an overview of the functions of social categorisation, and its “companions” (stereotypes, prejudice and discrimination), including multiple category memberships;
  • We explain different ways in which stereotypes and prejudice emerge and how everyday journalism may inadvertently contribute to each of these processes;
  • We review some of the consequences prejudice has on thoughts, feelings and behaviour — including self-protective mechanisms that affect journalistic research, interviewing and reporting;
  • Finally, we provide an overview of the primary established approaches for solutions from the social sciences based on this research, including tips and tools on how journalists can reduce pejorative, polarising and simplistic storytelling through more sensitive and nuanced reporting.

Structure and self-esteem: Why we categorise and what happens when we do

The world is an extremely complex place. Due to our limited mental capacity, we humans have a need to bring order to the continuous stream of sensory input, to reduce complexity and make sense of it all (4). One way we do that is by forming categories boxes that group similar objects together and efficiently reduce the amount of information we need to process. If I know that an animal is a bird, I “know” that it has feathers and is able to fly. These are characteristics I’ve learned and I don't have to check this again with each and every bird. Categorisation thus helps us to condense observations and structure the world. However, these characteristics don’t always apply: ostriches and penguins are birds, but can’t fly. And in some cases, categorisation might be misleading like in the case of bats, which are flying mammals, not birds. However, we still stick to categories in many cases, because in addition to ordering chaos, categorisation processes also contribute to self-esteem: living in a well-structured and thus somewhat predictable world fulfils our need for control and mastery. It makes us feel that we understand what is going on and are able to handle things (see our article on self-efficacy). In addition, as helpful and self-serving as they are, we should keep in mind that categories are often somewhat arbitrary or oversimplified (e.g. the idea of gender as binary).

Once we have category labels, cognitive research shows that differences within each category tend to be largely ignored and differences between different categories tend to be overemphasised. This so-called accentuation effect applies both to simple objects, as well as biological or social categories, and it gets stronger when judgements are made under uncertainty (5). Labelling is thus by itself a powerful tool to alter perception and judgement, even before it comes to whatever characteristics might be associated with belonging to a certain category (see below).

By mentioning some social category memberships (usually considered to be the “extraordinary” ones) and not others (usually those seen as the “default”, e.g. being heterosexual, employed, or physically able and healthy), media creators render some features more salient than others. Moreover, highlighting a person’s group membership in a particular context suggests that it matters for the topic reported about (e.g. a criminal act or special achievement). Finally, labelling suggests which trait dimensions might be considered relevant for the audiences’ own personal identity within a given society, and thereby can heavily affect social climates of community versus division.

Historic examples

In the former country of Yugoslavia, a federation that existed in different forms in the Balkans from 1918 to 1992, Croats, Serbs, and Bosnians — groups that mainly differ in religious tradition, not in ethnicity lived mostly peacefully as neighbours, colleagues and couples for decades under President Josip Broz Tito (6, 7). Then, within a few years after Tito’s death in 1980, these categories were re-established in populist political speeches and public debate to sow division and a social climate, which — among economic and other factors — prepared the ground for wars, mass rape and genocide in the region, which in turn of course manifested the divides considerably (8).

In the African country of Rwanda, regime-driven media spread explicit hate propaganda against the Tutsi minority ethnic group, similarly to the way that the Nazi regime used the German tabloid newspaper Der Stürmer to rally against Jewish citizens ahead of the Holocaust (2). After the shock of the 1994 genocide in Rwanda, in which more than 500,000 Tutsi and supporters from other ethnic groups are believed to have been killed in just 100 days, and a subsequent regime change, naming ethnic groups was largely banned from public discourse. A consequent framing of the common identity of citizens “all being Rwandan” was part of a holistic reconciliation programme that successfully incorporates, among others, many features of constructive journalism, such as solution focus and appreciative inquiry (9; see also article 6 of this series).

The accentuation effect is stronger for groups we ourselves do not belong to — so-called “outgroups” — as compared to our own so-called “ingroup”. This outgroup homogeneity effect (10) makes it harder for us to distinguish people who are not in our group (“they all look the same”), suggesting that “they are all the same” (see also article 3). One explanation for the effect is that we usually spend more time with ingroup rather than outgroup members, naturally making us more attune to their variety and individual differences. At the same time, perceived outgroup homogeneity seems to legitimise simplifications and derogating judgements about all its members.

Journalists can try to mitigate the outgroup homogeneity effect in at least two ways: First, especially when portraying social groups the audiences might be less familiar with, show a broad, diverse, and balanced spectrum, e.g. of migration biographies in order to counter dehumanisation (11, for further reading see a study about the European media discourse on migration and this series’ article 5 on Power of Images). Second, diversity and representation within newsroom teams are very likely to have corrective and sensitising effects on nuanced research and reporting about marginalised or minority groups.

Finally, all of us belong to not one but multiple social categories due to our gender, profession, native language, political stance, ethnic identity, family situation, certain music style, soccer club preferences, etc. Different people may identify strongly with some categories and not others (12). Whether a particular "box" seems meaningful and therefore applicable to oneself or to others depends very much on the context (13). Lab research shows that we are not only “cognitive misers” that categorise due to scarce mental resources (14), but also “motivated tacticians”'(15): We use certain boxes and labels when it makes us feel good to do so, and we actively suppress them when it doesn’t. For instance, if a doctor from a low status minority gave us positive feedback about our health, concepts associated with their minority group are temporarily less likely to come to mind than normal, and the person is primarily perceived as being a doctor. Conversely, if the feedback was negative, we suppress associations linked to the respectable profession and perceive the person primarily as a low status minority member, which allows us to downgrade their competence or credibility (16).

Clearly, social categories are hardly ever neutral, even if we focus on just one dimension at a time. The process of categorisation often automatically activates stereotyping, prejudice and even discriminatory behaviour, all of which we will elaborate on below.

In a nutshell (14, 17)

  • Categorisation: the process of grouping objects or people together according to certain features (e.g. young, old; carpenter, teacher, etc.) along a particular category dimension (e.g. age, gender, religion, profession, subculture, nationality, etc.).
  • Stereotype or cliché: a cognitive schema, a mental knowledge structure about features commonly associated with a category. It can be value-neutral (e.g. “people with this particular professional training typically do things according to a schedule”), and one can be familiar with a stereotype without necessarily believing in it.
  • Prejudice: a generalised positive or negative, i.e. judgmental, attitude based on a person’s social category membership alone, implying feelings of sympathy or aversion (e.g. “they are so wonderfully structured and conscientious” or “…so terribly compulsive and rigid”).
  • Discrimination: refers to the behavioural level: Do I treat a yet unknown person differently, solely based on their social category membership? Do I favour them, subtly devalue them or even openly deny them access to things like health care, housing, educational and job opportunities, freedom of speech, or other human rights (e.g. “I won’t invite her for a job interview, she won’t fit into our informal, creative start-up team culture anyway”)?

For journalists, the fact that all people belong to many different groups at the same time raises the question which of those should be mentioned in reporting, if any at all, and in what way. For example, we are more likely to present someone as belonging to a minority group than a majority group (as this appears literally more noteworthy), and thus this category is more likely to be linked with newsworthy and often negative events and behaviours, such as crime. Research shows that even a single exposure to a media report may suffice to establish negative attitudes (18).

Even the use of specific words makes a difference. One example: research has shown that the use of nouns to indicate group membership (“a male Roma”) had stronger negative effects than the use of adjectives (“a Roma man”). Adjectives make a feature appear less “essential”, while nouns imply stability and invite stronger stereotypic inferences and intergroup bias (18). At the same time, failing to note the group memberships of crime suspects may lead to accusations of “censorship”, and in Germany, the debate about according press codex is ongoing (19, 20, 21).

At the same time, mentioning more attributes may add nuance to a story, broadening and enriching perspectives and mental associations. When we describe people in a multi-faceted way, we give audiences more opportunity to connect with them. Acknowledging that we all have something in common on some level is a powerful tool to overcome prejudice (22). Moreover, while we mostly think of primary categories such as gender, ethnicity, or age when we talk about diversity and representation, we often omit aspects like living in an urban vs. rural area, regional differences within a country, health status, level of education, socioeconomic background, having kids or not, etc. Many of these “secondary” categories may have a stronger influence on the lifestyle and values that matter for a story than the “primary” categories, and may also help audiences relate.

Journalists need to be aware that they may be choosing categories based on their own associations. In the first step, categories “just” structure the world; they only gain meaning in the next step, when stereotypic ideas of further “typical” characteristics kick in and are applied to a person just because they belong to that group. In the next section, we take a look at how these associations are learned and what role the media play in establishing and reinforcing them.

Learn and repeat: Stereotypes, prejudice, and the role of the media

We have already seen that stereotypes may be defined as associations between a category label and what we believe to be typical of its members. Prejudice is the associated feeling towards them. But how do such associations evolve in the first place? The following table lists some classic psychological processes that lead stereotypes to emerge, though it is of course far from complete. For each process, we have a brief general description on the left and a spotlight on how they work in media contexts on the right.

Evaluative conditioning

… is a very basic, yet powerful process that refers to repeatedly pairing an initially neutral stimulus, such as a sign of social group membership, with a pleasant or unpleasant stimulus (23). For small children, the latter can be the disapproving or friendly face of a parent or peer; later it can be toys, book or movie characters that associate features such as gender or appearance with values like “power”, “weakness” or “danger”, being “good” or “bad”.

Media contribute to evaluative conditioning wherever news, public discourse or ironic prototypes repeatedly present certain attributes or issues in the context of a category, which helps to form attitudes without the person even being aware of it (24, for further reading see a recent study by the Global Strategy Group that shows massively biased wording, framing, and pictures in US media reports about criminal courtroom trials).

Operant conditioning

… refers to learning from reward or punishment as a direct consequence of our own behaviour. Stereotypes emerge when we generalise good or bad experiences with a single person to others who happen to share a feature with that person. Operant conditioning may also work in statements we make: When we get approval or disapproval for speaking positively or negatively about a certain group, it is likely to affect our future attitude (25).

In media contexts, such social reinforcement can also work indirectly. For instance, if pejorative views are presented by respected celebrities or as the predominant opinion, this may be “encouraging” to those who subtly sympathise, since consensus fulfils our fundamental need to belong and feels better than being alone with one’s — otherwise “strange” or “extremist”— views (26).

Social Learning Theory

… by Albert Bandura (27) stresses that we also learn from observing others as (role) models. Depending on whether these models are “rewarded” or “punished”, we can decide whether we wish to do as they did, or not, e.g. in stating a particular opinion.

In media contexts, guests in talk shows or protagonists in documentaries are likely to serve as role models and may inspire prejudiced or egalitarian attitudes by making these appear more or less acceptable and worth adopting for the audience, depending on the feedback they get. Similarly, an online post that gets lots of hate comments or “likes” can have strong effects on the audience's attitudes and uttered opinions (28).

The Theory of Intergroup Conflict and Cooperation

… by Muzafer Sherif (29) traces the origins of prejudice and discrimination back to constellation of interests: If two or more groups compete for scarce resources such as land or power in a win-lose-scenario, it is advantageous for one's own group to devalue or even attack the competitor. Suggesting that “less worthy” others are profiting at your own group’s expense is a classic way to drive social division, as populists of all stripes have done in all eras.

Media professionals should be careful not to reiterate narratives of polarising competition thoughtlessly or give them an (inappropriate) stage. The same line of research shows that intergroup hostility can be overcome by a meaningful common goal, even more so, if that goal is only achievable through cooperation between groups. We will talk more about that below.

Social Identity Theory

… provides an even simpler account for negative prejudice: We seek to feel valuable, and since we socially identify with the (in-)groups we belong to, we tend to see these more positively than the outgroup(s) we do not belong to (30, 10; see also article 3 of this series). Studies show that the mere assignment of people to different groups — even artificial and temporary, “minimal” ones — makes them prefer their ingroup over the outgroup (30).

Journalists should avoid framing in ways that invite ingroup-outgroup dynamics (like “us” vs. “them” or “group A” vs. “group B”). Remember that categories are not only arbitrary in themselves, but that it is also often arbitrary — or at least an editorial decision — which of the multiple available group memberships are mentioned how and where, and which are not, and that there are multiple possible axes of similarity (“us”) and dissimilarity (“them”).

Mind and (re-)materialisation: Real consequences in real life

Most people are well aware that categorisation and stereotypes simplify things inappropriately, and they seek to treat others fairly and without prejudice. Why do stereotypes have such an impact on our feelings and behaviour, even if we don’t want them to, and why are they so resistant to change? Decades of social cognition research shows that associations between a group and supposedly typical attributes of its members get activated automatically — that is, without intention or even awareness — when we encounter a person belonging to that social category (15). In fact, other experiments show that stereotypes even affect judgement and behaviour when participants are personally convinced that they aren’t true, especially when stereotype activation happens outside of conscious awareness (17).

In a now classic laboratory experiment that was run in the US in 2001 to explore the role of automatic cognition in police interventions (31), participants were asked to sort pictures according to whether they showed a tool (e.g. pliers, a drill, etc.) or a weapon (e.g. a firearm), and to do so as quickly and as accurately as possible. Before each of these pictures, the participants were briefly shown a photo of a face, so quickly that they could not actually consciously perceive it. The result was that participants were more likely to misclassify a tool as a weapon if they had previously seen a dark-skinned face than if they had seen a light-skinned face. They were also faster in correctly classifying weapons as weapons in these trials. Similar effects emerged in a video-game study with Black and White figures holding guns vs. harmless items: participants were instructed to shoot at those holding guns, and were faster and more likely to shoot Black avatars even if they held a harmless object; the bias was about equally strong in White and Black community samples (Study 4) (32). These kinds of experiments have been replicated in other places, with different stigmatised minority groups.

Apparently, concepts that are still associated with being non-White in many “Western” countries or the global North — such as hostility, crime and arms — are activated in a split-fraction of a second and can have immediate effects on perception, decision-making and action in ambiguous or stressful situations, where conscious control kicks in too late (33). Automatically activated associations can lead to all kinds of racial profiling cases with more or less severe consequences for those concerned, from arbitrary controls and arrests to deadly physical interactions (for further reading, see this VICE article, and references 34, 35 or 36, on the debate of racial profiling in Germany).

Setting cases of determined hate crime aside, discriminatory bias is often subtle and unconscious, which makes it harder — but not impossible — to overcome. Of course, it’s not just about race or ethnicity: a person’s age, gender, occupation, housing area, etc., can lead to unconscious bias and subtle discrimination in journalistic contexts. For example, you could ask yourself:

  • Am I more or less willing to adjust to the other person’s calendar when scheduling an interview simply based on his or her category features (e.g. expert or celebrity status)?
  • Do I ask some people more critical questions than others, and why?
  • Am I speaking more slowly or in simpler terms, suggesting the other person is less intelligent or less educated (37)?
  • Do I introduce or address an expert with their full professional title and affiliation or not (38, 39), thereby reflecting as well as further fostering associations between expert status and other group memberships, such as gender (40)?
  • Do I interrupt some people more often than others, do I generally behave in a more accommodating or more dismissive way through eye contact, body posture, mimics or tone of speaking (41, 42)?

Such microaggressions and unconscious biases are of course not only relevant when interacting with interviewees or guests, but also within newsroom teams. This aspect is beyond our scope in this article. Nevertheless it is important for a healthy and constructive working environment, which, in turn, is likely to have a ripple effect on reporting.


Before we get to the ways in which you can overcome stereotypical biases and discrimination, let us take a look at two clusters of effects of stereotypes that either a) protect them against being proven wrong or b) even create stereotypic realities through vicious cycles in social dynamics. For both aspects, we will focus on how these phenomena affect journalistic research, interview design, and storytelling.

Bulletproof: How we deal with exceptions to the rule

One could think it should be easy to overcome stereotypes by simply presenting data or showing people they are wrong through observation or experience. However, the human mind seems to be designed in a way that such critical pieces of information are dealt with in a, well, “flexible way” again, mainly because it feels better to think that you have known something all along instead of critically scrutinising one’s beliefs and integrating new knowledge, which can be hard to do.

In article 1 of this series, we elaborated on biases in information processing, including the so-called confirmation bias – the tendency to prefer evidence that confirms rather than questions one’s beliefs. Evidence that could prove stereotypes wrong and help overcome them might simply be ignored, actively avoided by confirmatory research, or forgotten. In journalistic contexts, that might mean:

  • in research: googling certain keywords or word combinations rather than others when first approaching a topic, e.g. searching for group labels and “crime statistics”; in addition to this human bias, filters and algorithms in social media further contribute to highly selective content “bubbles”;
  • in interviews: choosing stakeholders or experts that are likely to share one’s own (preliminary) views on a subject or asking leading questions during the interview;
  • in editing and storytelling: cutting out “confusing” or “less interesting” parts or choosing headline, teaser or pictures in a biased way (see article 2 on Power of Language and article 5 on Power of Images);
  • starting and/or ending with the preferred arguments (see article 7 on Primacy Effects).

Given that some counterevidence or counterviews cannot be blended out entirely, people are very creative in explaining other people’s statements or behaviour in a way that allows them to maintain their prior beliefs or clichés. For example, if someone from one’s ingroup founded a successful small start-up company, it is usually attributed to their inner virtues (e.g. hard work, creativity, or entrepreneurial spirit). If a member of an outgroup is similarly successful, it is often attributed to favourable (external) circumstances. This so-called ultimate attribution error (43) perpetuates negative prejudices about an outgroup, even if they have just been refuted. Conversely, if an ingroup member fails, this is attributed to external factors (e.g. market volatility), while outgroup failure is typically blamed on personal characteristics (such as laziness or low management competence). See article 4 on Causal Attribution).


  • should take care not to imply causal attributions that perpetuate prejudices (e.g. by framing some groups as active, others as passive; or as “lucky” rather than having “earned it” when introducing interviewees or reporting about them);
  • must be aware that their own memberships are likely to subtly affect such framing. When journalists from marginalised groups report on issues they are “involved in”, they are often accused of a lack of objectivity. From a scientific point of view, this is just equally likely to affect reporting by privileged majority members, as the strong dominance of views from the global North in reporting illustrates (44);
  • are likely to profit from more diversity, equity and inclusion in newsrooms when it comes to overcoming or at least balancing out such attribution biases in interpretation of counterstereotypic information.

Other ways of reinterpreting evidence that contradicts one’s prejudices are subtyping and recategorization: For instance, men who are sensitive (contradicting the stereotype of “manly men”) might be dubbed not “real men” or even called slurs. Similarly, a professionally successful female who is rational and analytic may be denied her femininity, and be predominantly perceived according to professional background (e.g. being a scientist or engineer), thus shifting the focal category dimension from gender to professional education. Both processes allow one to maintain stereotypic beliefs (in this case about men or women), even in the face of clearly contradictory evidence (45).

Journalists should be careful…

  • In research: to be aware of their own category schemes when searching for protagonists meant to represent a particular group;
  • In interviews: not to unwittingly invite subtyping or re-categorisation processes by representing counterstereotypic protagonists as “interesting” or “special”, i.e. atypical;
  • Prior to and in editing and storytelling: reflect the personal relevance of each feature for the person’s identity and make clear that whatever makes them interesting does not make them less “typical”.

Ironically, people often remember so-called “counter-stereotypic” evidence better thanks to the elaborate processes of “explaining it away” (46). While it may seem strange, this is a most efficient way of protecting the initial belief from being questioned (47).

All the above phenomena “just” happen in the perceiver’s mind. But what if stereotypes and prejudices also affect the real, objective behaviour of the counterpart in social interaction, say in an interview or a scientific intelligence test? What if those expectations produced confirmatory “facts”, true behavioural differences that “prove” the prejudice, yet would never have emerged, if there had been no prejudice in the first place?

Vicious circles in co-creating stereotypic realities

What we think and feel affects our behaviour, and our behaviour affects the behaviour of others around us. Therefore, it does not really come as a surprise that stereotypes and prejudice have the power not only to filter or reinterpret realities, but also to create them in social interaction.

Self-fulfilling prophecies emerge in many everyday situations (42), such as between teachers and pupils (48, 49) or recruiters and job applicants (41). They can also affect interactions between journalists and their interviewees or guests. In one study (41), real-life job interviews were observed. It was found that the Caucasian recruiters treated candidates of different ethnicities in systematically different ways, and that the non-White candidates on average presented themselves in a less professional manner. In a second study, actors were trained to present the two interview styles in the role of the recruiter. Only Caucasian candidates were randomly assigned to one of these two experimental conditions. This study revealed that successful self-presentation of the candidates was a direct result of the atmosphere created by the recruiter: less immediacy (e.g. more distant sitting arrangements and body position) and shorter interview time led to more insecure behaviour on the part of the applicants (e.g. more speech errors per minute). To sum up: In real life, if a recruiter’s interaction style varies based on candidates’ ethnicity or other stereotypically stigmatising factors, an independent observer is likely to conclude that those candidates indeed behaved less competently and were passed over for “objective” reasons, i.e. that the prejudice is actually true.

In journalism, self-fulfilling prophecies are most likely to emerge during interaction phases in research and reporting, so one should take great care…

  • to create fair, non-discriminatory general conditions for interviews and conversations (e.g. give them equal amounts of time, similar room arrangements, exhibit equally appreciative non-verbal behaviour), in particular when hosting more than one guest at a time;
  • to pose interview questions fairly, e.g. not to “interrogate” members of minority or low status groups with closed or leading questions vs. asking open-ended questions of privileged group members, providing them with the power to co-direct the course of the conversation;
  • to become more sensitive to one’s own unconscious biases, including for instance whether celebrities or experts are treated differently than “normal people”, who tend to be underrepresented in public debate and therefore less media savvy, which might make them in turn appear less suitable for future invitations and further promote underrepresentation.

Last but not least, stereotypes can even affect performance without any discriminatory behaviour. Simply knowing that a negative stereotype exists about one's own group (e.g. “girls aren’t good at maths”, “white men can’t jump”) can be enough to give someone anxiety that they could confirm the stereotype and thereby discredit their ingroup. The more important that group membership is to them, the more important it will be for their self-esteem (e.g. to perform well at an intelligence test, to give a crisp presentation). This feeling of stereotype threat — similar to test anxiety — can induce stress and restrict cognitive capacity, thereby leading to objectively worse task performance… which then ironically confirms the stereotype (50). Please note: once participants are told that what they are going to do is not relevant for the stereotype (e.g. a test that only measures “some skills”, not “mathematical intelligence”), people perform just as good as the other, non-stereotyped group. Conversely, when group labels are made very visible, performance of the stereotyped group suffers more strongly. Thus, worse performance is likely a product of feeling social threat, not of any “real” deficit whatsoever (51).

Journalists may mitigate stereotype threat by…

  • framing questions in a way that does not highlight the person’s social group membership(s) when it is not relevant for the topic at hand (e.g. “What do you, as a single mom…”);
  • doing a “pre-interview” in order to reduce pressure.

Rethink and get in touch: Solutions from the social sciences

A major goal of all the research we have reviewed so far is to find strategies to prevent and/or overcome stereotypes, prejudice and discrimination.

If stereotypic associations affect behaviour regardless of people’s personal attitudes and best intentions, the safest way to overcome negative effects is to systematically unlearn them and/or establish new associations. The human brain adjusts more flexibly than we often expect and unlearning stereotypes actually works quite quickly and sustainably in the lab (52). Eliminating prejudiced associations from everyday life, however, is a job for society as a whole for parents, teachers, the authors of (children’s) books, films and series, etc. Considerable change has taken place over the last decade for instance with regard to normalising representation of diversity, non-conventional, non-heteronormative gender roles and relationships (53). Documentaries that portray people in ways usually not associated with their social category will support those changes. The more often counter- or non-stereotypic associations are established under different conditions, the less that confirmation or attribution biases, re-categorisation or subtyping “protect” from a change in mental knowledge structures. 

Nuanced representation of multiple social category membership provides potential for social inclusion, empathy, and more individualised person perception (see 22 for a comprehensive review). Or put another way: the more diverse, nuanced representations we have, the harder it will be to keep people, places, and ideas locked in “boxes”. Journalists can support this by reporting about people’s intersectionality. This is especially true in the case of “gateway groups” in conflicts that identify with both of two conflicting parties.

Sherif’s work on realistic conflict theory (29) has shown that competing groups reconcile when they pursue a common goal they can only achieve in full cooperation (see also 54). Focussing on eventual common goals (“we’re in this together”) when reporting about conflict might help to shift focus in narratives and even inspire polarised communities to organise events (especially powerful in local journalism; 55).

After all, one of the eldest and most powerful approaches against prejudice is contact. Decades of empirical studies show that contact reduces prejudice by increasing knowledge, increasing empathy, and reducing anxiety (56). Journalism has many ways to create and to report about events that encourage informal personal exchange between people of different backgrounds and viewpoints in a constructive manner (see below).

Guidelines and training for fair nonverbal behaviour, wording of questions, length and overall arrangement in interviews (as common for professional applicant selection procedures) are another efficient remedy that may also prevent unconscious bias and self-fulfilling prophecies in journalistic settings.

Tools and tips

  • When researching, mind the keywords you use in your searches and how they set the stage for confirmation bias and selective facts, viewpoints or potential experts, protagonists, etc.
  • When doing interviews or hosting guests, make sure to create equal, fair, appreciative conditions for all and not fall into traps of unconscious bias, self-fulfilling prophecy dynamics or create stereotype threat (find more strategies in the according text section above).
  • In interviews as well as storytelling, use sensitive language and avoid metaphors, labels or visualisations that dehumanise groups of people in any way, (such as “waves of x group flooding in”, etc.). When showing victims of war, catastrophe or hardship, avoid framing that only evokes pity, as pity implies a status difference and denies them agency or the ability to cope. Report about individual experiences to give protagonists more substance and show variety (57). When reporting about people from the LGBTQIA* community, make sure you have their consent to indicate their gender identity or sexual orientation, and double-check with them or consult professional guidelines to use the proper terms. If you are looking for solid tips on how to amplify women’s voices at each stage of the news value chain check out this comprehensive resource.
  • In storytelling, think hard about which social group memberships you need to mention at all and at what point (headline, lead, text; see article 7 on Order Effects). Use adjectives rather than nouns (see above), as they imply less stable “essence” of category labels and allow more flexibility for “un-boxing” people.
  • In storytelling, consider that each person belongs to many different groups at the same time — which ones can you mention to add nuance, invite a shift in perspective or create a relatable similarity for your audience, if appropriate? How can you use different angles or layers to break or at least open boxes and invite audiences to see the person as an individual at the intersection of all those layers, without creating an exotic subtype that “does not count”? How can you, in a similar way, report differently about ingroup issues and members, and eventually add nuances?
  • When you tell stories about interesting counter-stereotypic protagonists, avoid suggesting causal attributions that “protect” general prejudices because the inconsistency can be blamed on very special conditions. The more numerous and the more various the “exceptions from the rule”, the lower the chance that they are explained away and end up reinforcing the bias. The platform “Wir sind der Osten”, for instance, is an initiative founded to give voice to (progressive) Eastern Germans with many different biographies, lifestyles and viewpoints.
  • When you research and report about conflict between groups, try to look out for any possible common goals or even a superordinate common category membership they share. In reporting, dig deeper to explore to what extent seemingly competitive situations might be reframed so that a cooperative approach could allow for win-win instead of win-lose results. What else can you add (data, perspectives) that might make people come up with win-win perspectives on the issue? (see also our upcoming article on mediation and negotiation)
  • At the same time, journalistic events can be created to acknowledge that conflicting positions exist and still do not necessarily need to polarise. Contact among individuals is one of the most powerful ways to understand different positions, bridge divides instead of ignoring them and foster empathy. The initiative “The World Talks” is an outstanding example of a journalistic initiative pairing up people with different views on controversial issues. It is very successful, scientifically evaluated, and has gone global. Local and community journalism provides lots of opportunities to create contact within neighbourhoods (for Bonn Institute’s suggestions on “Lösungsjournalismus im Lokalen”, 55)
  • Reflect on your own social category affiliations: Which ones do you consider relevant to your personal identity, which ones not so much? Which ones are associated with privilege in the society you live in, which ones subjected you to (unconscious) bias or even open discrimination? How might this affect your research and reporting, the way you interact with different people and frame issues, maybe even how you tackle research, the keywords you use? What might you have come across if you used other terms or portals?

Stereotypes and prejudice create social divides and fuel polarising, extremist forces within societies. Frustration due to lack of opportunities, not feeling seen, heard or respected as a result demotivates people to participate in democratic processes and breeds hatred and violence. Journalists have a very high responsibility when it comes to shaping the climate of public discourse in news, talk shows, and social media. Subtle reiteration of stereotypes of ingroup vs. outgroup in storytelling, of only slightly different treatment of guests from different backgrounds, etc., can be overcome with better knowledge about these phenomena and their impact. The decisive factor here is the will to change century-old narratives and embrace the treasures of diversity and to enable more conciliatory and respectful journalism in editorial practice with constructive know-how, patience and daily training.

Some additional resources on the topic

  • Elliot Aronson: Nobody left to hate. Teaching compassion after Columbine.
  • Jane Elliot: Blue Eyed. Documentary on anti-racism workshops aiming for experience-based learning.
  • Irshad Manji: Don't label me: an incredible conversation for divided times.
  • Trevor Noah: Born a crime. Stories From a South African Childhood.

About the authors

Margarida Alpuim is a Portuguese psychologist and journalist. She completed her master's in Community Psychology at the University of Miami, where she focused on issues of collective well-being. As a journalist, Margarida wants to explore more constructive ways to tell stories that consider both the audience and media professionals. Margarida currently works from Lisbon on innovative projects that unite psychology and journalism.

Katja Ehrenberg holds a PhD in psychology and is professor at Fresenius University of Applied Sciences in Cologne. For almost 25 years, she has conducted research, taught and published on topics of social, communication, organisational and health psychology. As a freelance systemic consultant, she advises teams and individuals and enjoys applying social science findings to the various challenges of everyday human (work)life.


  1. McIntyre, K., & Gyldensted, C. (2018). Constructive journalism: An introduction and practical guide for applying positive psychology techniques to news production. The Journal of Media Innovations, 4(2), 20-34.
  2. Kellow, C. L., & Steeves, H. L. (1998). The Role of Radio in the Rwandan Genocide. Journal of Communication, 48(3), 107–128.
  3. The UN Office on Genocide Prevention (o. J.): Countering and Addressing Online Hate Speech: A Guide for policy makers and practitioners. Available at https://www.un.org/en/genocideprevention/documents/publications-and-resources/Countering_Online_Hate_Speech_Guide_policy_makers_practitioners_July_2023.pdf
  4. Taylor, S. E. (1981). A categorization approach to stereotyping. In D. H. Hamilton (ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 83-114). London: Psychology Press.
  5. Corneille, O., Klein, O., Lambert, S., & Judd, C. M. (2002). On the role of familiarity with units of measurement in categorical accentuation: Tajfel and Wilkes (1963) revisited and replicated. Psychological Science, 13(4), 380-383.
  6. Holbrooke, R. (1999). To end a war. The Conflict in Yugoslavia - America’s Inside Story - Negotiating with Milosevic. Penguin.
  7. Stanišić, S. (2020). Herkunft [Where you come from]. München: Luchterhand.
  8. Iskakov, Y. (2010). Stereotypes and Prejudice in Conflict: The Case of Bosnia and Herzegovina. Journal for the Study of Peace and Conflict, 2012, 122-29.
  9. McIntyre, K., & Sobel, M. (2018). Reconstructing Rwanda: How Rwandan reporters use constructive journalism to promote peace. Journalism Studies, 19(14), 2126-2147.
  10. Ostrom, T. M., & Sedikides, C. (1992). Out-group homogeneity effects in natural and minimal groups. Psychological Bulletin, 112(3), 536–552.
  11. Bain, P. G., Vaes, J., & Leyens, J. P. (Eds.). (2013). Humanness and Dehumanization. New York: Psychology Press.
  12. Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel, & W. G. Austin (Eds.), Psychology of Intergroup Relations (pp. 7–24). Chicago: Nelson-Hall.
  13. Macrae, C. N., Bodenhausen, G. V., & Milne, A. B. (1995). The dissection of selection in person perception: Inhibitory processes in social stereotyping. Journal of Personality and Social Psychology, 69(3), 397–407.
  14. Fiske, S. T., & Taylor, S. E. (1991). Social Cognition. McGraw-Hill Book Company.
  15. Fiske, S. T. (2004). Intent and ordinary bias: Unintended thought and social motivation create casual prejudice. Social Justice Research, 17, 117-127.
  16. Sinclair, L., & Kunda, Z. (1999). Reactions to a black professional: motivated inhibition and activation of conflicting stereotypes. Journal of Personality and Social Psychology, 77(5), 885-904.
  17. Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56(1), 5-18.
  18. Graf, S., Linhartova, P., & Sczesny, S. (2020). The effects of news report valence and linguistic labels on prejudice against social minorities. Media Psychology, 23(2), 215-243.
  19. Norddeutscher Rundfunk (2016). Minderheitenschutz: Pressekodex unverändert. [Protection of minorities: Press code unchanged.] Available online at https://www.ndr.de/fernsehen/sendungen/zapp/Minderheitenschutz-Pressekodex-unveraendert,presserat122.html
  20. Deutschlandfunk (2019). Herkunftsnennung von Straftätern - Keine Lückenpresse. [Naming the origin of criminals - No sketchy press]. Available online at https://www.deutschlandfunk.de/herkunftsnennung-von-straftaetern-keine-lueckenpresse-100.html
  21. Mediendienst Integration (2019). Sollten Medien die Herkunft von Strafverdächtigen nennen? [Shall media indicate the origin of crime suspects?]. Available online at https://mediendienst-integration.de/artikel/sollten-medien-die-herkunft-von-tatverdaechtigen-nennen.html
  22. Prati, F., Crisp, R. J., & Rubini, M. (2021). 40 years of multiple social categorization: A tool for social inclusivity. European Review of Social Psychology, 32(1), 47-87.
  23. Hofmann, W., De Houwer, J., Perugini, M., Baeyens, F., & Crombez, G. (2010). Evaluative conditioning in humans: a meta-analysis. Psychological Bulletin, 136(3), 390-421
  24. Sweldens, S., Corneille, O., & Yzerbyt, V. (2014). The role of awareness in attitude formation through evaluative conditioning. Personality and Social Psychology Review, 18(2), 187-209.
  25. Conger, A. J., Dygdon, J. A., & Rollock, D. (2012). Conditioned emotional responses in racial prejudice. Ethnic and Racial Studies, 35(2), 298-319.
  26. Morrison, K. R., & Matthes, J. (2011). Socially motivated projection: Need to belong increases perceived opinion consensus on important issues. European Journal of Social Psychology, 41(6), 707-719.
  27. Bandura, A., & Walters, R. H. (1977). Social learning theory (Vol. 1). Prentice Hall: Englewood cliffs.
  28. Bak, M. P., & Keßler, T. (2012). Mir gefällt’s, wenn’s euch gefällt! Konformitätseffekte bei Facebook. [I like it if you like it. Conformity effects on Facebook]. Journal of Business and Media Psychology, 3(2), 23-30.
  29. Sherif, M. (1958). Superordinate goals in the reduction of intergroup conflict. American Journal of Sociology, 63(4), 349-356.
  30. Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behaviours. European Journal of Social Psychology, 1, 149 –178.
  31. Payne, B.K. (2001). Prejudice and perception: The role of automatic and controlled processes in misperceiving a weapon. Journal of Personality Social Psychology, 81, 181–192.
  32. Correll, J., Park, B., Judd, C. M., & Wittenbrink, B. (2002). The police officer's dilemma: Using ethnicity to disambiguate potentially threatening individuals. Journal of Personality and Social Psychology, 83(6), 1314–1329.
  33. Payne, B. K. (2006). Weapon bias: Split-second decisions and unintended stereotyping. Current Directions in Psychological Science, 15(6), 287-291.
  34. Die Beauftragte der Bundesregierung für Migration, Flüchtlinge und Integration / Die Beauftragte der Bundesregierung für Antirassismus (2023). Rassismus in Deutschland. Ausgangslage, Handlungsfelder, Maßnahmen. [The Federal Government Commissioner for Migration, Refugees and Integration / The Federal Government Commisioner for Antiracism (2023). Racism in Germany. Initial situation, fields of action, measures.] Available online at https://www.integrationsbeauftragte.de/resource/blob/1864320/2157012/77c8d1dddeea760bc13dbd87ee9a415f/lagebericht-rassismus-komplett-data.pdf?download=1
  35. Antidiskriminierungsstelle des Bundes (o. J.). Ausweiskontrolle! Rein zufällig nur bei Ihnen. [Identity check! By chance, only with you.]. Available online at https://www.antidiskriminierungsstelle.de/SharedDocs/aktuelle-faelle/DE/Ethnie_Rassismus/Ethnie_Racial_Profiling_Bahn.html
  36. Scharlau, M. (2021). Racial profiling: Positionspapier zu menschenrechtwidrigen Personenkontrollen. [Racial profiling: Position paper on identity checks that violate human rights.] Berlin: amnesty internation. Available online at https://www.amnesty.de/sites/default/files/2021-10/Amnesty-Positionspapier-zu-Racial-Profiling-Oktober2021.pdf
  37. Nelson, T. D. (2005). Ageism: Prejudice against our feared future self. Journal of Social Issues, 61(2), 207-221.
  38. Duma, N., Durani, U., Woods, C. B., Kankeu Fonkoua, L. A., Cook, J. M., Wee, C., ... & Knoll, M. A. (2019). Evaluating unconscious bias: speaker introductions at an international oncology conference. Journal of Clinical Oncology, 37(36), 3538-3545.
  39. Files, J. A., Mayer, A. P., Ko, M. G., Friedrich, P., Jenkins, M., Bryan, M. J., ... & Hayes, S. N. (2017). Speaker introductions at internal medicine grand rounds: forms of address reveal gender bias. Journal of Women's Health, 26(5), 413-419.
  40. Dupree, C. H., Torrez, B., Obioha, O., & Fiske, S. T. (2021). Race–status associations: Distinct effects of three novel measures among White and Black perceivers. Journal of Personality and Social Psychology, 120(3), 601-625.
  41. Word, C. O., Zanna, M. P., & Cooper, J. (1974). The nonverbal mediation of self-fulfilling prophecies in interracial interaction. Journal of Experimental Social Psychology, 10(2), 109-120.
  42. Banaji, M. R., Fiske, S. T., & Massey, D. S. (2021). Systemic racism: individuals and interactions, institutions and society. Cognitive Research: Principles and Implications, 6, 1-21.
  43. Pettigrew, T. F. (1979). The ultimate attribution error: Extending Allport's cognitive analysis of prejudice. Personality and Social Psychology Bulletin, 5(4), 461-476.
  44. Granger, J. (2021). Decolonising journalism: what does it mean, why does it matter and where do you start? Available at https://www.journalism.co.uk/news/decolonising-journalism-what-does-it-mean-and-where-do-you-start-/s2/a856450/
  45. Richards, Z., & Hewstone, M. (2001). Subtyping and subgrouping: Processes for the prevention and promotion of stereotype change. Personality and Social Psychology Review, 5(1), 52-73.
  46. Hastie, R., & Kumar, P. A. (1979). Person memory: Personality traits as organizing principles in memory for behaviors. Journal of Personality and Social Psychology, 37(1), 25-38.
  47. Ehrenberg, K., & Klauer, K. C. (2005). Flexible use of source information: Processing components of the inconsistency effect in person memory. Journal of Experimental Social Psychology, 41(4), 369-387.
  48. Rosenthal, R., & Jacobson, L. (1968). Pygmalion in the classroom. The Urban Review, 3(1), 16-20.
  49. Okonofua, J. A., Walton, G. M., & Eberhardt, J. L. (2016). A vicious cycle: A social–psychological account of extreme racial disparities in school discipline. Perspectives on Psychological Science, 11(3), 381-398.
  50. Spencer, S. J., Logel, C., & Davies, P. G. (2016). Stereotype threat. Annual Review of Psychology, 67, 415-437.
  51. Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 69(5), 797–811.
  52. Kawakami, K., Dovidio, J. F., Moll, J., Hermsen, S., & Russin, A. (2000). Just say no (to stereotyping): effects of training in the negation of stereotypic associations on stereotype activation. Journal of Personality and Social Psychology, 78(5), 871-888.
  53. Poole, J. (2014). Queer representations of gay males and masculinities in the media. Sexuality & Culture, 18, 279-290.
  54. Aronson, E., & Patnoe, S. (1997). Cooperation in the classroom: The jigsaw method. New York: Longman
  55. Urlbauer, L. (2023). Lösungsjournalismus im Lokalen. Ein Werkstattbericht des Bonn Institute in Kooperation mit der Rheinischen Post und dem Journalismus Lab der Landesanstalt für Medien NRW. Bonn: Bonn Institute. Available at https://www.bonn-institute.org/cms-assets/files/bi_rp_werkstattbericht_09.pdf?v=1680245184
  56. Pettigrew, T. F., & Tropp, L. R. (2008). How does intergroup contact reduce prejudice? Meta‐analytic tests of three mediators. European Journal of Social Psychology, 38(6), 922-934.
  57. Heinrichs, E., Ehrenberg, K., & Tillmanns, P. (2022). Constructive Opportunities and Challenges. Journalism in times of war. Bonn: Bonn Institute. Available at https://www.bonn-institute.org/cms-assets/files/bi_study_war_reporting_web.pdf?v=1671126854

Stay informed...

Knowledge transfer, debates, events – our b° newsletter shows you what constructive journalism can do for the industry and society. Stay informed by subscribing to our newsletter.

Sign up