As an editor, I reject quite a lot (based on out of scope submissions, or too limited value to readership (e.g. lack of newness)).
If I commit to review for a conference, I usually get assigned a number of papers for which I did not see the title/abstract beforehand. Furthermore, I know there will only be limited time (and usually willingness) to make revisions. Hence, I may again have a substantial rejection rate.
When I am invited to review a journal submission, I only commit if the paper seems like something I should read anyway. (so part of the papers I might reject if I were to review them, I simply do not accept the review invitation for). Also for revising journal papers there is usually more time and willingness, and I'm also likely the get the revision back to judge again. Hence, with journal articles I'd be more inclined to select 'major revision', or 'revise and resubmit'. Those being the most common assessments.
Finally the standing of a journal may play a role. For a conference or a no or low impact factor journal, you mainly assess the soundness of the work. For a high IF journal (even though this should be done by the editor and not the reviewers), I would be inclined to also judge size of the work and size of the contribution to the literature, and the general interest to the readership. Hence, higher change of reject for high IF journal.
I have never rejected a research paper and always ask for minor revision or major ones, only once when I had been sure it was a copy one and fake data in article, I asked the editor to check it !
I do not think that all the research during the review will be accepted because it is illogical, as there must be a certain proportion of this research will not be acceptable. According to my practice as a research reviewer, the rejection rate was about 30%,it is depending on the type and severity of the research presented for evaluation.
I am always positive in my assessment procedure, and at the same time when I start reviewing, I seeking for standard steps in writing papers, thesis, dissertation, article including plagiarism level, so it depends on many factors. I do not remember the exact number.
Rejection is the last decision for me, I try to review the research more than once to give a greater chance of admission taking into account the scientific level of research.
I have never rejected any empirical studies reviewed. What i do in most cases is to ask for major review with enough useful suggestions to help the study better. In case of reviews, rejection is recommended without thinking twice once the works are not thorough.
Based on my experience as a reviewer of papers for journals I've rejected something like 2 thirds of the overall amount. This is due, mainly, to lack of time to do a thorough review or the fact that the paper is not totally within the scope of my expertise.
In the case of conference papers, the rejection rate is smaller, because I've already committed to participate as a reviewer in that time frame.
I reviewed many papers for various journals, and I rejected a paper only once because it was not in the journal's range of covered fields of science. In many cases, though, I demanded major revisions.