I made a list of things I hate:

• Studies that assume people are clueless wrt their online identity and social data management
• Studies that assume people are totally aware and make active conscious decisions about their online identity and social data management
• Studies that generalise to all social networks from facebook or twitter
• Studies that generalise to all people from one demographic
• When that demographic is western college students
• ..because that just means they couldn't make time/effort to step off campus
• Studies that assume people have a straightforward or one-to-one mapping with their social profiles
• Mining profile information to improve recommendations/advertising/marketing blah
• Mining profile information to infer attributes
• Studies that act like their results will still be meaningful in 6 months
• Anything that reduces humans to dots and lines (all SNA lol)
• Papers that assume there are technical solutions to all privacy issues
• Studies that totally ignore imagined audience or expectation of privacy (it's public so we can use it!)

I'm not saying I can necessarily do better, but I will at least refer back to this list and try...