The numbers vary a lot.  Here are a few credible research stats

• The National Center for Educational Statistics reported in 2011 that 9% of students in grades 6-12 experienced cyberbullying.
• The Centers for Disease Control found in 2011 that 16.2% of students had been bullied via email, chat rooms, instant messaging, websites or texting — compared to 20.1% who had been bullied on school property (traditional bullying) — during the 12 months prior to the survey.
• The Cyberbullying Research Center reports that “on average, about 24% of the students who have been a part of our last six studies have said they have been the victim of cyberbullying at some point in their lifetime.”
• Dan Olweus, who the editor of the European Journal of Development Psychologyreferred to as the “father of bullying research” wrote a 2012 article for that journal where he said that “claims about cyberbullying made in the media and elsewhere are greatly exaggerated and have little empirical scientific support.” Based on a three-year survey of more than 440,000 U.S. children (between 3rd and 12th grade), 4.5% of kids had been cyberbullied compared to 17.6% from that same sample who had experienced traditional bullying.  An even more interesting statistic from that study is that only 2.8% of kids had bullied others
• A 2011 Pew study found that 15% of teens say they have been the “target of online meanness.” When you include in-person encounters, 19% say they’ve been “bullied” in the past year.
• An article in the July, 2013 Journal of Adoloscent Health by Robin Kowalski and Susan Limber reviews several studies to show a vast range of cyberbullying from 4% to 72%* and concludes “that there appears to be a substantial, although not perfect, overlap between involvement in traditional bullying and cyberbullying. Additionally, the physical, psychological, and academic correlates ofthe two types of bullying resembled one another.

*The 72% figure comes from this study. If you go to the methodology section you’ll see it’s an opt-in survey from a single website — Bolt.com, which has closed down after allegations of copyright infringements. As a former university instructor in educational research and survey design, I find this to be a highly unorthodox and unreliable sampling methodology.

View this slideshow from Pew for more