With a sample of 10 that makes sense. We have done over 150 PHP tests over the last couple of years and the average number of issues is higher than other technologies. What would account for that bias?