Tenopir & King‘s confirmation of the finding (by Kurtz and others) — that as more articles become accessible, more articles are indeed accessed (and read), but fewer articles are cited (and those are cited more) — is best explained by the increased selectivity made possible by that increased accessibility:
The Seglen “skewness” effect is that the top 20% of articles receive 80% of all citations. It is probably safe to say that although there are no doubt some bandwagon and copycat effects contributing to the Seglen effect, overall the 20/80 rule probably reflects the fact that the best work gets cited most (skewing citations toward the top of the quality distribution).
So when more researchers have access to more (or, conversely, are denied access to less), they are more likely to access the best work, and the best work thereby increases its likelihood of being cited, whereas the rest correspondingly decreases its likelihood of being cited. Another way to put it is that there is a levelling of the playing field: Any advantage that the lower 80% had enjoyed from mere accessibility in the toll-access lottery is eliminated, and with it any handicap the top 20% suffered from inaccessibility in the toll-access lottery is eliminated too. Open Access (OA) allows all the cream to rise to the top; accessibility is no longer a constraint on what to cite, one way or the other.
(I would like to point out also that this “quality selectivity” on the part of users — rather than self-selection on the part of authors — is likely to be the main contributor to the citation advantage of Open Access articles over Toll Access articles. It follows from the 20/80 rule that whatever quality-selectivity there is on the part of users will be enjoyed mostly by the top 20% of articles. There is no doubt at all that the top authors are more likely to make their articles OA, and that the top articles are more likely to be made OA, but one should ask oneself why that should be the case, if there were no benefits [or the only benefit were more readers, but fewer citations!]: One of the reasons the top articles are more likely to be made OA is precisely that they are also the most likely to be used, applied and cited more if they are made OA!)