At least one other coauthor also had the data, which would have made it more awkward to withhold it. Still, this argument overall does point towards none of the authors knowing that the data was made up.