We combine the Critical Cost Efficiency Index and the Houtman Maks Index to evaluate the consistency of subjects in a set of choice data from budgets. We show that by simply allowing subjects for one significant mistake (by removing the worst observation), the mean Critical Cost Efficiency Index in a set of choices from an online experiment with the general population increases by 6 percentage points.
Furthermore, we find that by excluding the worst observation per subject, the fraction of subjects wasting 5% or less of their budget increases from 45% to 64%. The highest improvement in terms of efficiency can be seen among retired and 65+ aged subjects which indicates that their low level of effciency is largely due to a single mistake.