Assuming your assertion is true, you'd expect the percentage to be higher if women are as likely as men to commit abuse, which implies that your assertion is not true or the percentage is wrong.
Here's where the % came from. Say girls go to college for an average of 4 years. If 27.7 girls of the 1000 are raped or almost raped in 6.9 months, then 192.7 (approx.) girls of the 1000 will be raped over the course of 4 years, giving us a raped or almost-raped percentage of about 19%. Clearly this is some fucking impeccable statistics.
For all intents and purposes, this could include me kissing a girl until she felt uncomfortable. Pretty sure the penetration part disqualifies your hypothetical.