ResearchGate was started by 3 academics in 2008 and is now used by 3 million---Bill Gates has invested heavily so he clearly sees its potential. And yet we'd like to advise you to proceed with caution. Here are a few criticisms of these services: caveat emptor (let the buyer beware).
1. ResearchGate automatically emails invitations to your coauthors on your behalf. These invitations are made to look as if they were sent by you but are emailed without your consent.
2. Academics can claim work authored by someone with a similar name, even if it doesn't belong to them. All three sites use algorithms to generate publications that look like they might be yours based on similar bibliographic information. If you have a common name, be particularly careful that no-one else has erroneously claimed your publication on their profile. (Or that you don't accidentally claim theirs.)
3. It's hard to tell what the purpose of these sites is, but assume they have their own agenda. Since 2001, only educational institutions have been allowed to use the prestigious .edu. domain (Academia.edu was registered before then). Mendeley is owned by Elsevier, a scholarly publisher with a profit margin in the billions. And ResearchGate's founder wants a Nobel Prize.
4. Unintentional copyright infringement is really, really easy. ResearchGate has a stated mission to give science back to the people who make it happen. This is a reference to the stranglehold that large academic publishers have over the scholarly communication process. The reality is that in most cases, the copyright in your published journal article no longer belongs to you because you have signed it away in the publisher agreement.
These sites rely on you uploading your papers to build their content base, so they don't provide you with much information on what you can and can't upload. They rely on you reading and understanding your publisher agreement to know which version (if any) you can post. And lots of people get this wrong. Elsevier has recently started sending mass takedown notices not just to Academia.edu and ResearchGate, but also to individual researchers and even universities that are hosting infringing Elsevier content on their websites. Even if you do the right thing, you can't always stop your coauthor from uploading an infringing copy of a work with your name on it. So be vigilant.
If you'd like to post your work online, contact the Swinburne Research Bank team for assistance. We're experts in interpreting publisher agreements and can help you avoid the pitfalls. And we don't have a hidden agenda: we just want to make more publicly-funded research available online, legitimately.
Image credit: flickr/Frank Kovalchek