Food for thought.

A paper published stated “Limited special studies have indicated that fire reports are subject to a substantial but unbiased error rate. That means there is a high rate of errors, but the errors favor no particular characterization, so the errors tend to cancel out when reports are aggregated for statistics.” Are we being told that the statistics are basically correct because the errors cancel themselves out, so the substantial rate of error does not mean anything?

I understand that these reports are mostly completed by individual with little to no training in fire investigation and this is the reason for the substantial error rate. What I do not understand is the methodology that allows one to ignore this error rate by saying, “errors tend to cancel out when reports are aggregated for statistics.” Where do they get the statistical support for this conclusion? If anyone knows I would appreciate the information.