this is probably field dependent, but I would say: - don't focus on individual studies but try to find meta-analyses. - if not available, then review articles

this is probably field dependent, but I would say: - don't focus on individual studies but try to find meta-analyses. - if not available, then review articles. - if not available, then individual papers. # of citations is probably a closer proxy for influence rather than "good" (if you mean quality). Of course, one hopes that there is a strong correlation between the two. More in the weeds would then assess the impact factor of the journal that an article is published in as well as other 'altmetrics'.