The data was from four counties. The data shows tens of thousands of switched votes. The data shows a distribution slope proving a pattern, except in Wayne county where manufactured ballots were the means to cheat the totals so there is no discernible distribution slope (think little blue imogis dropping from high left to lower right).
Did he say whether these were the only 4 counties he studied or did he study the whole state and cherry pick these counties to make his case?