Chemical Spillage as a Model for Accident Causation

Chemical spillage has been measured for undergraduate students undertaking a laboratory practical designed to help develop their chemical handling skills. There was a low correlation found between the amount of chemical spilled and the ability to accurately carry out the primary goal of the experiment, which was to measure the concentrations of samples of unknown dilution. There was also a low correlation between the amount spilled by individual students carrying out consecutive experiments, which is consistent with the proportion of spillage incidents due to personal factors (e.g., accident-prone individuals) being low. Quantitatively, only ≈6% of the spillages could be accounted for by factors relating to the individual, with a (95%) confidence interval of 0.2–23%. These observations therefore indicate that a blame culture should be avoided when dealing with chemical spillages, and also that a more effective focus for efforts to improve laboratory safety should be on organizational or workplace factors, through, for example, improving teaching methods for all students, rather than targeting specific individuals for retraining who may have had a noticeable chemical mishap. These observations also suggest that the volume of chemical spilled by a student should not be used as a measure for summative assessment, as it is a poor predictor of future spillage.