If a dog gave a positive result 93% of the time, but it was a false positive 40% of the time, does that mean 63% of suspects were carrying? Maybe I'm square, but that seems absurdly high to me unless the dogs are only used when there is already suspicion otherwise.