First, we applied our sampling methodology for estimating the taxes avoided for the middle 300 Internet Retailer firms on a state-by-state basis.42 The reason for applying this state-by-state method was that it allowed for variation in each state's ratio of sample avoided taxes to sample total taxes, creating a more accurate portrayal of the each state's estimated avoided taxes. Adding the estimated avoided taxes for the middle 300 firms to the avoided taxes for the top 150 and bottom 50 firms within each state yielded the total avoided tax for the top 500 internet retailers in each state. Second, we then distributed the avoided taxes attributable to firms in the “tail” by allocating the total estimated avoided taxes for firms in the tail on a pro-rata basis according to each state's proportion of taxes avoided by the top 500 internet retailers.