I may be reading too much into your statement, but are you using
"presacral length" as a proxy for body size in your calculations?
Yes -- as one of five characters which are all analysed separately.
If I remove the sauropods from the sample for this character, the highly
significant increase that I get for the whole sample does not change much.
If I remove all sauropodomorphs, the magnitude of the increase rises
drastically... Quite a counterintuitive result!
*Erketu* is not in my analysis; it was published too late and is far too
incomplete.