Determining openness of a dataset using PID Graph API


I’m currently doing some research into whether papers that use openly available datasets are more/less likely to be retracted than those that don’t. However, do do this, I need to be able to separate the DOIs (that reference DOIs of datasets) which do use them/don’t use them.

Is there any way I can do this? I’ve scoured the schema for some such property but perhaps I’ve missed something, or perhaps it’s not possible at all.

Many thanks.