First of all, as mentioned by others, any comparison is meaningless unless from the same master and high res file, with the 16/44 version simply derived from that. For a fair comparison of different resolutions there used to be a great resource of high quality files in various resolutions from the same source, specifically to enable real comparison (see DSD, where and why do we use it? - #11 by Innocent_Bystander), but having just checked I note it is currently unavailable, with return uncertain.
From reading numerous descriptions over several years, the consensus of opinion where files are from the same high resolution master seems to be that there is a subtle difference, in terms of more “air” or better ambience to recordings, rather than anything more distinct like increased detail or altered soundstage. My own impression varies from something consistent with this to, sometimes, no discernible difference at all. As I have yet to run out of storage space I do tend to buy the highest resolution files that are available, unless a silly premium price, on the basis that they could sound better and won’t sound worse. However if the mastering is different, which seems not to be uncommon, either could sound better.
The same of course is true when comparing, say, vinyl with CD- and the futility of direct comparisons unless mastering is known to be the same was brought home to me some years ago when a friend and I had apparently identical CD releases of the same album, which sounded distinctly different: scrutinising the discs and liner the only discernible physical difference was that one said made in UK, the other made in Germany. They must have been mastered differently.
N.B. In terms of serious listening and sound quality I only stream from my own local store, online streaming only being for sampling new material to decide whether or not to buy, when sound quality is not a major consideration, the free services being quite adequate.