The study said about half of the girls surveyed reported having had sex before. 40% of those were found to have at least one sexually transmitted infection, versus 26% overall.
Now let's see what a little High School-level probability theory gets us:
Let P(S) be the probability a randomly selected girl from the sample reports having had sex before.
Let P(I) be the probability a similarly selected girl has at least one sexually transmitted infection.
Bayes' theorem tells us:
P(I,S) = P(S)P(I|S) = P(I)P(S|I)
Now P(S) = 0.5 (about half the girls reported having had sex before)
P(I) = 0.26
and P(I|S) = 0.4 (40% who claimed having had sex before had an infection, vs. 26% overall)
Therefore P(S|I) = P(S)P(I|S) / P(I) = (0.5)(0.4)/(0.26) = 0.77
i.e. 77% of those who had at least one STD reported having had sex before. That means 23% of those girls currently infected with an STD reported never having had sex.
If we accept the proposition that the incidence of STDs in girls who have had sex does not depend on whether they lied about having sex, if follows that roughly
(0.23)(0.26)/(0.4) = 15% of girls in the study are not being completely honest about their sexual history.
This looks to me like a pretty clear indication these girls have learned the religious conservative message very well: "Do whatever you want, but be ashamed and lie about it."
