"a third of sexism complaints are exaggerations" Was not how he phrased it. It said there were multiple people experiencing similar activity. So that could be

"a third of sexism complaints are exaggerations" Was not how he phrased it. It said there were multiple people experiencing similar activity. So that could be read as not everyone considers XYZ to be sexism. While anecdotal, the comment that this "should always be backed up by some sort of data if you don't want to look unreasonable" doesn't make any sense. The data quality, by definition, are what is being questioned. You don't solve such a problem by quoting similar data, and you most likely won't find "proper" data given the nature of the observation.