Design an experiment that would determine the percentage of paper currency that has cocaine traces on it.?

Question

backgrond information: It is known that US paper currency in the general circulation is contaminated with cocaine. Several mechanisms have been offered to explain this finding, including contamination due to handling during drug deals and the use of rolled up bills for snorting. Drug is then transferred from one contaminated bill to others during counting in financial institutions.

the question continued: give some ideas on how one would be able to sample all the currency of canada and be representative, consider such things as: sampling, the method of analysis to use, what level of cocaine in money indicates criminal activity, and how reliable are the results you obtain from this experiment

Answer 1

I presume that bills are tested by swiping them with a swab that is then tested for cocaine content in a non-quantitative colorimetric test.  The test probably already has well defined limits for what level of the drug would give a positive result, and how often it gives false positive and false negative readings.  I don't think there is any legal way that a bill could pick up cocaine.

Therefore, I guess:  any cocaine on the bills would be evidence of criminal activity, and the reliability would be determined by the already known rates of false positive and negative tests.  Confirmation of the false positive rate could be made by testing a sample of uncirculated bills.  These could also be marked and sorted through circulated currency as handled by a financial institution to see if cocaine would be transfered to them without direct exposure in drug deals or by users.

For the sampling, I would sort the bills by denomination, and do a preliminary test to get an idea of what percent show evidence of exposure.  From this, I would decide on how large a sample to work with:  a low percent would need larger sample sizes to detect.  And then, I would take several samples for each denomination, perhaps 5 sets of 100 bills/denomination from 1 - 100, and test them.  This would give you a large enough sample size to allow statistics that would ensure the validity of your tests.  Of course, a statistician could tell you ahead of time what sampling approach to use in order to be most effective.

I would also test bills from several locales to see if there seemed to be a geographic gradient of exposure that might correlate with already known patterns of cocaine use.  If there was such a pattern, it might provide corroboration of your results.