The data seems questionable:
The figure was found by asking an online, representative sample of 779 women whether they themselves or whether their best friends had ever tried to self-induce an abortion. Of the Texas women surveyed, 1.7 percent said they had performed an abortion on themselves, but 4.1 percent of them said their best friend had or they suspected she had.
So, we're throwing out a figure based on a sampling of 779 women AND going off the fact that some of them 'suspected' their best friend had tried to induce an abortion? Doesn't really seem like sound research.
http://www.theatlantic.com/health/archive/2015/11/texas-self-abort/416229/