SummaryIn determining dose limiting toxicities in phase I studies, it is necessary to attribute adverse events to being drug related or not. Such determination is subjective and may introduce bias. We develop methods for removing or at least diminishing the effect of this bias on the estimation of the maximum tolerated dose. The approach that we suggest takes into account the subjectivity in the attribution of adverse events by using model-based dose escalation designs. The results show that gains can be achieved in terms of accuracy by recovering information lost to biases. These biases are a result of ignoring the errors in toxicity attribution.