I recall a paper that sought to explore how willing authors were to release data when they had published open access and stated 'data available on request'. Iirc it was below 50% that responded with data. See related here; https://osf.io/preprints/psyarxiv/jbu9r/

That being said, there are also issues with opening the gates to anyone to get your data and use it to publish; https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

My data is OA, but I'd appreciate any one wanting to use to first reach out and discuss any nuances that may influence their analysis.