Why Do the Ethanol Paper’s Correlation Numbers Conflict With Its Conclusion?
Why Do the Ethanol Paper’s Correlation Numbers Conflict With Its Conclusion?
Ioan and colleagues describe vitreous ethanol as more strongly related to blood ethanol than urine is. Yet the results section reports a urine–blood Pearson correlation of .905 in 36 paired cases and a vitreous–blood correlation of .887 in 85 paired cases. The printed numbers are ordered the other way. The discrepancy should be preserved, not silently “corrected.”
Study design
The correlations came from different subsets with different sample sizes, so even a valid numerical comparison would not be a simple head-to-head test on identical cases. The full study involved 202 forensic cases, but not every case contributed both comparison pairs. Correlation also measures association, not prediction error for an individual blood value.
Limits
The paper’s conclusion may reflect other considerations or a reporting error; the supplied text does not resolve which. It would be unsafe to use this inconsistency to declare urine universally superior or vitreous universally inferior. Collection condition, postmortem change and the actual specimen pair still matter.
Application
In a report, quote each coefficient with its denominator and explicitly note that the prose ranking conflicts. Request underlying paired data before claiming a definitive comparison of the two matrices.
Scientific source
Ioan BG, Jitaru V, Damian R, Damian SI. Study on the relationship between the concentration of ethanol in the blood, urine and the vitreous humour. Romanian Journal of Legal Medicine. 2015;23:211–216. doi:10.4323/rjlm.2015.211.



