page logo
Home Evaluations Anti-Bias-Training

Anti-Bias-Training 

Mitigating bias—our commitment and your role 

(Unconscious) Biases can skew results and harm lecturers, students and long term the quality and fairness of the study program. We address this through how we ask which questions (avoiding bias from our side by e.g. asking non-suggestive questions, offering guidance in open feedbacks, asking before grading) and helping you avoid unconscious biases via training. 

In context of feedbacks we see e.g. how often likeableness and competence are showing up in strong correlation, a general gender bias (with male educators being perceived as more competent), higher competence-ratings for native-english speakers, and so on. In a typical classroom, students rate the likeableness and ability to engage correlating with how far they are usually seated (highest in the front-row, least in the back-row).  

Those biases often happen unconsciously, with some being linked to expectations based on e.g. gender-norms (e.g. reading the same behaviour in a man as good leadership and in a woman as aggressive and bossy), fields of expertise and others.  

We invite you to challenge your own biases with this video and the training in Canvas: 

Getting started- the Basics: 

https://youtu.be/Iaan2XSw6ho?si=0soIVDFo4zaSGxC1 

 

How to mitigate bias when giving feedback 

  • Ground yourself first 

    • Re-read the supervision agreement, intended learning outcomes, and any rubric. Rate against these, not personal preferences. 

    • Reflect on the whole supervision period—not just the most recent meeting (avoid recency bias). 

  • Focus on observable behaviors, not identities or traits 

    • Comment on what you saw: meeting frequency/punctuality, feedback timeliness, clarity of guidance, availability, ethical conduct, support for autonomy. 

    • Avoid references to gender, ethnicity, age, accent, disability, nationality, or personality labels. 

  • Keep items independent 

    • Rate each criterion on its own merits. Don’t let one strong/weak area color all ratings (avoid halo/horns effects). 

  • Separate roles and constraints 

    • Distinguish what was within the supervisor’s control from lab/resource or administrative constraints. 

    • Don’t conflate project difficulty or your expected grade with supervision quality. 

  • Use specific evidence and neutral language 

    • Prefer “I observed…” statements with examples and frequency. 

    • Be concrete and actionable. 

    • Examples:  

      • Helpful: “We had 30–45 minute meetings every two weeks as agreed; feedback on drafts arrived within 5–7 days with specific comments on methods.” 

      • Unhelpful/biased: “I didn’t vibe with their style” or “Her accent made it hard to follow.” 

  • Balance positives and improvements 

    • Note what supported your learning and one or two priority improvements. 

  • Acknowledge shared responsibility 

    • If you missed meetings or deadlines, factor that into how you interpret the supervision experience. 

  • Quick pre‑submission checklist 

    • Have I rated against the stated criteria? 

    • Did I provide at least one concrete example per major point? 

    • Did I avoid identity-based or personal-appearance comments? 

    • Did I keep each rating independent? 

    • Did I consider the entire supervision period? 

    • Would this feedback help a supervisor improve specific behaviors? 

  • Cooling‑off practice 

    • If you feel frustrated or elated after a key event (e.g., submission), draft comments and wait 24 hours before submitting. 

  • If in doubt 

    • Ask yourself: “Would this read as fair and professional to someone who doesn’t know me or my supervisor?” If unsure, revise for specificity and neutrality. 

 

An overview over relevant biases: 

Term 

Explanation 

Ableism, Disableism 

Ableism refers to judging people based on their based on abilities that are not in the respective domain - e.g. an ability to code well, vs an ability to hear unaided. 

Disableism is the discrimination against people seen as impaired. 

Anchoring Bias 

Relying too heavily on the first piece of information you see. 

Mini-Me-Effect, Similarity-Attraction Effect 

People are attracted to others who are similar to them in appearance and personality traits. 

We tend to like people who are like us in many ways. 

(Fundamental) Attribution Error 

We often think a person’s behaviour is because of their personality, even though it may be due to the situation they are in. 

Authority Bias 

We tend to believe and follow the opinions of authority figures, even though their opinions may not be correct. 

Confirmation Bias 

We tend to look for information that supports our existing beliefs and ignore information that contradicts them. 

Blind Spot Bias 

We often think we are unbiased and unaffected by biases, even though we are influenced by them just like everyone else. 

Cross-Race Effect 

People may have difficulty recognising and distinguishing faces of individuals from a different ethnic group than their own. This is linked to a tendency to more easily identify/recognise faces of people belonging to one’s own racial group. 

Distance Bias 

We tend to place more importance on things that are closer to us in space or time. 

Fading-Affect Bias 

Negative memories tend to fade faster than positive memories, which can sometimes lead to biased behaviour. 

Framing Effect 

Different ways of presenting information can influence how people respond, even if the content is the same. 

Gender Bias 

Biased perceptions and stereotypes based on gender can lead to unfair treatment or misinterpretation of situations. 

Illusory Correlation 

Stereotypes often lead to attributing certain characteristics to certain groups, even if they are not accurate or supported by evidence. 

Ingroup-Bias 

Favoring people from your own group over others or judging them individually, while judging outgroup-members based on the perceived group-characteristics. 

Judicial Bias 

Biases that can affect judges’ decision-making, including cognitive biases and external influences. 

AI Bias 

Biases that can occur in artificial intelligence systems due to skewed or incorrect data, leading to unequal treatment or discrimination. 

Contact Effect 

More contact with different people can help reduce prejudice and hostility between groups. 

Naïve Realism (Direct Realism) 

Believing that our own perception of reality is objective and unbiased, and assuming others should reach the same conclusions if they have the same information. 

Negativity Bias 

Negative experiences or thoughts have a stronger impact on us than neutral or positive ones. 

Primacy Effect 

The first information we receive often has a strong influence on our judgments and memories. 

Racial Bias 

Stereotypes and biases based on race can influence judgment and behaviour, even if they are unconscious. 

Recency Effect 

 Giving more weight to recent information than earlier information. 

Social Desirability 

People may give answers they think others want to hear, instead of expressing their true beliefs, to avoid social disapproval. 

Status Quo Bias 

A tendency to prefer the current situation over making changes, especially when there are limited alternatives and knowledge about them. 

Stereotypes 

Generalised beliefs about certain groups of people based on their gender, race, ethnicity, religion, sexual orientation, socioeconomic background, or education.