Open access and citation rates
Making a research paper openly available on the internet enables it to be found and read by a wide audience - including researchers who may otherwise have no access to the work because, for example, it has been published in a journal to which they have no access. There is evidence of a relationship between the wide availability of research and its impact: specifically, papers which are openly available may be cited more frequently than those which are not.
The Open Citation Project (OpCit) maintains a bibliography of papers on the topic of open access and citation impact:The effect of open access and downloads ('hits') on citation impact: a bibliography of studies
Most papers in the bibliography - though not all - show a relationship between open access to research and an increased citation rate i.e. those papers which are freely available are cited more often than those which are only available to those with a subscription to the journal. The effect has been demonstrated across a number of subject disciplines. Look at the bibliography and make your own mind up!
Where possible, the self-archived version of a work will be identical to any formally published version. However, for the majority of journal papers, the "self-archived" copy in the repository is not identical to the published work. Often, though, it is similar- it has been peer reviewed and accepted for publication. Most discussion of citation impact assumes that it is the published version of the work which is cited. Making a full text version of the work available complements the published work. It makes the research visible and gives researchers a very good idea of the contents of the published work and its relevance to their own research. More work needs to be done on the relationship between open access to research papers and any subsequent increased citation rate but it seems likely there are several mechanisms at work: