Experimental artefacts occur during atom probe tomography (APT) analysis of oxide dispersion strengthened (ODS) steel particles. They skew the chemical composition measurements and the observed morphology of the particles. This study is mainly focused on chemical bias since it is one of the most debated questions about these particles. The chemical biases that occur during APT analysis are corrected by two distinct approaches: a numerical one using field evaporation simulation and an analytical one with a chemical composition correction (CCC) model. These two approaches provide a convergent solution: in the studied ODS steel, the particles are pure in Y, Ti and O and all the atoms of Fe and Cr observed in the particles are artificially introduced by APT artefacts. The particle stoichiometry evolves from YTiO2 for the smallest (<2 nm) to Y2Ti2O5 for the biggest (>8 nm). Based on simulation results, the particles morphological biases are attributed to cross-over trajectories of ions coming from the particles.