Combatting Unconscious Bias in Education

The intervention consists of Unconscious Bias Training (UBT) designed specifically for primary school teachers. The training builds on rich evidence from an ever-growing interdisciplinary literature and has been fine-tuned to the primary school setting via two pilot studies in both the UK and Italy. Our theory of change is that making teachers aware of the existence and pervasiveness of unconscious bias and suggesting concrete actions to counter it (targeting classroom environments, modes of communication and feedback, and teaching Unconscious Bias (UB) to children) can improve teacher practices, learning environments, and, in the long run, pupils’ outcomes.

2021-11-30

2022-06-30

teacher’s ability to recognize bias, changes in teaching practices, and disadvantaged pupils’ performance

Teachers' ability to recognize bias is measured by using a database of 100 images coming from Pearson textbooks. These images have been previously selected and rated by Pearson experts. The measurement consists of the teachers evaluating how stereotypical an image is on a 1-5 Likert scale. They are presented with a series of 20 images and they will do this exercise twice, one at the registration step and the other after the evaluation is done.

Changes in teaching practices are measured with a survey that will be repeated post-evaluation. We ask about the frequency of group work, role games and simulation, debates and Socratic seminars, presentation of students' research, vote to decide between activities, use of ICT (information and communication technology), and students' auto-evaluation.

Pupils' performance is assessed via a simple test done twice (before and after the intervention). The test is divided into four thematic categories: Math, Language, Verbal reasoning, and Non-verbal reasoning. Each theme is divided into two questions with a box for the final result and another one for details and explanations about how they reach the final result. We then identify disadvantaged pupils via a survey conducted at the same time asking for characteristics such as gender, disability, or migrant status. The test is then double-marked by teachers, first openly, then blindly (with the pupil's information hidden).

teachers’ enjoyment as well as pupils’ learning environment, well-being, and aspirations

Teachers' enjoyment is measured with a survey conducted twice (before and after the intervention). We ask about their satisfaction with several aspects of teaching on a 1-7 Likert scale: teaching in general, life in general, their work, their school, their colleagues, their relationship with students, their relationship with the students' families.

The pupils' learning environments are measured through a survey conducted twice (before and after the intervention). The pupils are asked about their overall well-being in the classroom, how much they like school, how easily they can get help for schoolwork, the extent of anti-social behaviours in their classroom and their school, and reactions from their teacher and school. All questions are based on Likert scales.

We measure the pupils' aspirations and expectations in a survey conducted twice (before and after the intervention). We first ask pupils about their dream job via an open question, then how confident their parents and teacher are in them achieving this goal on a 1-5 Likert scale.

The study consists of a two-arm randomized controlled trial. The treatment arm is an unconscious bias training tailored for primary school teachers. The training consists of three sessions. The control arm is a training programme on digital skills and inclusive teaching, and will have exactly the same amount of hours of intervention and at the same time as the treated teachers.
The research population will include grade 3-4-5 teachers and pupils in Italian primary schools. The schools are recruited by Pearson.

Not available

individual-level randomization of teacher done by the computer once the teacher signs up on the website

Teacher

No

300 teachers. 100 in the first year of the study and 200 in the following year

300 teachers and roughly 1000 students

150 teachers per arm

the minimum detectable effect size with 100 teachers individually and equally randomized between treatment and control arm is 0.28 of a standard deviation. Once we will reach 300 teachers we will have an MDE of 0.16