I thought you might be interested in this.
Here is the full paper, available as a PDF: http://minus.com/mqFALh
Here are links to some figures, if you don't want to read through the whole paper. (But there are many interesting facts that aren't in the charts, so I recommend that you download the full paper.)
Incidence of neglect and abuse in borderlines: http://i.min.us/ijs9Mo.png
Incidence of specific behaviors (sorted by abused vs. non-abused and BPD vs. non-BPD): http://i.min.us/ijtb3U.png
Relationship between IQ and abuse: http://i.min.us/ijo6ta.png
Note that the p-value of a statistical test is the probability that the result could have occurred by chance (or more specifically, the probability that the result could have occurred if there were actually no differences). Therefore, a p-value of .001 means the researcher is very sure of his prediction. This is helpful for analyzing the tables.