How do you know it _wasn't_? We basically have a _whole bunch_ of possible hypotheses here that the data/analysis here does not allow us to distinguish.