Those transactions pay fees. You think people just move money around for no reason? If you don't like the metric, then claim that it needs to be weighted, not discounted. I would conservatively guess transaction volume about 20% as actual transactions because businesses report actual sales in bitcoins.
Currently, transaction fees are negligible (~13 BTC total per day; on average, less than 0.08 USD per transaction, or less than 0.01% of the BTC volume excluding change-backs). For many kinds of non-payment transactions (tumbling, moving between hot and cold wallets, depositing and withdrawing from exchanges and other "bitcoin banks", over-the counter bitcoin purchases, etc.) those fees are not a deterrent.
And fees are not yet mandatory, is that correct?
Moreover, there are many people (such as fund employees) with motivation to generate "fake" traffic in order to give the impression of usage.
My guess is that payments for goods and services are no more than 5% of the blockchain transaction volume. The justification is that the latter does not vary with BTC price as one would expect. If that is the case, then one cannot use the traffic as a measure of adoption, even with a 0.05 weight, because the proportion of payment to non-payment traffic may vary a lot.
You call yourself an academic, yet you do nothing but criticize and offer nothing constructive. Your criticisms are weak and add little to the discussion. If you can't contribute something constructive, then you aren't putting in much effort and are resting on your laurels.
The whole basis of science and the scientific community is to question and criticise; the idea that someone who criticises a hypothesis has to produce a contrary, better or "constructive" (a highly subjective term in itself) thesis is incorrect.
You've clearly never gone through peer-review but I imagine you'd find the process pretty harrowing given statements like the one above.