This tells me if a patient dropped out, they are no longer "evaluable". I'd say drop outs were left out of the data as they dropped. A patient could be responding very positively and dropped out due to countless reasons unrelated to the drug, as well as patients dropping which were not responding well to the drug.
I'll start with the fact that if you need to make "assumptions" to arrive at your conclusion, well.... that's exactly my point. You shouldn't have to do that.
Next lets "assume" your "assumption" is correct. The only place I see an actual number of patients is on slide 24 and that says the trial was for 32 patients. So if this graph doesn't represent 32 patients shouldn't they clearly say so?
Are they simply reluctant to say: "this is not really a 32 person trial it's a 29/27/25 (pick one) trial."
That graph is not a scatter plot of all scores it is an average plus or minus the SEM. If you tell people you ran a 32 person trial but just show them a graph of 25 patients, is that kosher? It is, only if you clearly state that's what you did imho.
Taking Anavex completely out of the equation, if a biotech company (any "other" biotech company) leaves off critical information from their presentations, would it be wise for investors to fill in the blanks with the most positive method possible?
If several kids drop out of 1st grade during the year and at the end of the year the class average grades increased did the teacher do a good job teaching or get lucky that the bad students dropped out or unlucky became the really smart kids dropped out and the average would have been even better if they haven't.
Market Data 
Markets 
AI
