The project aims to assess teachers’ implicit cognitive biases in evaluating the performance and educational trajectories of primary education pupils as a function of students’ ascribed characteristics such as socioeconomic status (SES), ethnic origin, and gender. Individual assessment biases according to students’ backgrounds are challenging to capture with standard observational data due to social desirability bias and the impossibility of measuring all (un)observable student characteristics. Thus, to address these issues, we will implement a factorial survey that experimentally manipulates fictitious students’ profiles with different combinations of characteristics, including student performance and behaviour, to isolate the causal effect of students’ SES, ethnic background, and gender on teachers’ assessments. We will recruit pre-service teachers–university students enrolled in the BA Degree in Primary Education–from a representative sample of public and private Spanish universities. We are unable to contact our target population directly to protect participants’ privacy and data. Instead, we will contact the administrative staff at each sampled university so that they distribute the invitation to participate in the study through a mailing list comprising all enrolled students. Upon voluntary participation, which will be incentivised with a lottery of vouchers that is communicated in the invitation e-mail, the subjects will go over an online survey where we will randomly allocate a fictitious student profile per participant to evaluate. Finally, we will ask the usual individual sociodemographic questions. In this pre-analysis plan, drawing from three previous pre-tests reaching 600 observations among in-service and pre-service teachers, we define the study background and objectives, the research hypotheses, and the study design–including methods, measurements, models, power analysis, sampling, and data collection protocols–before conducting the fieldwork and data analysis.