I am not sure if this question is appropriate for the main site or not. A while back I asked about the Value of light-to-none peer reviewed pay-to-publish articles when evaluating a potential PhD student. The answers all suggest I should treat them as any other non-peer reviewed article (which is essentially the same thing you do with peer reviewed articles). This recent question by a PhD applicant asks Should I list my papers which are published in less known journals in my CV? and the answers seem to suggest that if the "less known" journals are predatory that you should avoid listing them on your CV. This seems in contradiction to the answers I got. Is there a contradiction, or am I missing something?
Pay to publish ≠ Lesser known journals
There are thousands of perfectly legitimate but low-impact journals. Publishing in those is by no means a negative reflection on one's character, skill, or whatever. There are also thousands of predatory, pay-to-publish journals. Publishing in those is almost always a bad idea, and may reflect negatively on an individual.
I don't think it's necessarily a contradiction.
Presumably, listing an article you published in a predatory journal shows a lack of experience or understanding of the publication process.
The advice to you, as the person evaluating these CVs, is not to punish the student for this lack of understanding. The answers there suggest giving the student the benefit of the doubt, since he/she may have had an inadequate advisor who didn't train them in this aspect of academia.
But students can't count on everyone who evaluates their CV to be so understanding of their naïveté. Thus, the advice not to list it on the CV - since the reader's negative impression of those who publish in predatory journals can outweigh any favorable impact the content of the paper may have.
Aside from whether this is a contradiction or not, I think it is ambitious to assume that the answers given here are necessarily internally consistent. Different persons answer different questions, and the underlying assumptions etc. are not necessarily the same. That is, I am pretty sure that one would be able to find two related questions where the most upvoted / accepted answers are indeed contradictory, but I see absolutely no way to prevent this.