Considered by whom?
In medical research publishing, all authors must declare whether they have a conflict of interest in the study they want to publish. That's because the introduction of a money motive does influence the conduct and interpretation of research. In the case of Guttmacher Institute, the money motive is so strong that barely any reliable research comes out of that place. Their studies are designed to give pro-abortion results.
Whenever I look at studies involving any aspect of abortion, the first thing I look at is the source of the study. Most of these studies originate from some abortion associated group; those studies can be discounted without bothering to read past the abstract.
Much of the information people "think" they know is inaccurate: for instance, the idea that the majority of abortion clients are young teens. Not so. Only eighteen percent of U.S. women obtaining abortions are teenagers; those aged 15–17 obtain 6% of all abortions, teens aged 18–19 obtain 11%, and teens younger than age 15 obtain 0.4%.
I was not sure whether to include young 20 somethings among the frequent abortion customers. The salient point is that the rates of abortion are highest among those who are *not* in steady relationships. Women in long-term relationships are more likely to use contraceptives; it is not unusual for a woman to go decades without getting pregnant after she has had as many children as she wants (which is less than 2, these days). If we were to take various estimates of contraceptive effectiveness at face value, we'd expect a much higher pregnancy rate than is actually seen.
Anecdotal information is worth -- well, whatever it's worth. Every woman I know who has had an unexpected pregnancy --- that is of those whom I know well enough to know such an intimate thing about them --- was a contraceptive user. Whether she was a careful, methodical, defense-in-depth, doubled-down jellies-jams-plugs-rubbers user or not, I must leave unanswered.
I do not know why, but it seems like every woman I've known who has had abortions feels a need to tell me about it. I don't even have to know them all that well--they just tell me. Also, I read a lot--it is not difficult to find some hyper-defensive woman writing a blog or article about why her abortions were perfectly justified. Among probably dozens of these accounts, I have *never* seen anyone volunteer that they were using contraceptives at the time they got pregnant, even if they give a fairly detailed account of how they got pregnant. That is a pretty clear indication that they were not. You don't need a study to figure this out: you just need to understand human nature. There is a stigma associated with abortion, which is even greater if the woman didn't bother to try to avoid pregnancy. Therefore, women who are asked point-blank if they were using contraceptives will very often lie and say they were, especially if they sense that the person asking will judge them negatively if they tell the truth.
You don't need to have intrusive medical tests to verify whether a woman is using contraceptives or not, in order to design a study that can accurately determine contraceptive effectiveness. The study does need to be prospective. If I were designing such a study is to include a survey on attitudes towards abortion. I think a prospective study on contraceptive effectiveness, coupled with attitudes towards abortion, would be extremely powerful. There is every reason to think that contraceptive efficacy would be highest among those who have a strong desire to not get pregnant and who will NOT use abortion as a backup under ANY circumstance.
Richard Doerflinger of the USCCB Pro-Life activities committee (and a way-back-when colleague of mine) points out that Guttmacher is always more accurate than the Centers for Disease Control Abortion Surveillance System, since AGI depends on detailed interviews and questionnaires of the abortion industry and the contraception-abaortion population, whereas CDC relies on legal reporting. There is no uniform abortion reporting requirement across all states in the U.S., and California hasn't reported to the CDC for years. That's why CDC's figures are always significantly less accurate than AGI's --- because CA, the #1 abortion state in the USA, does not report to CDC.
It is factually baseless to say that abortion-associated group' studies "can be discounted without bothering to read past the abstract." I don't know of a single pro-life, profamily or conservative group that takes that approach. In fact, exactly the opposite: they would mine the study for data, and forget the "abstract", which is likely to be a summary of their own tendentious "conclusions" cooked for abortion-friendly editorial writers and policy makers.
Abortion rates are comparatively low among married women, because, by far, married women's #1 form of "contraception" is in fact surgical sterilization, i.e. tubal ligation. Their #2 form is long-term endocrine disruption, i.e. hormone-based patches, inserts, and injections which impair normal physiological function (Ovarian follicle and corpus luteum).
Naturally anybody who's achieved long-term surgical or chemical sterilization is going to have fewer "omigod" surprise pregnancies, than some first-year college student who's still fumbling around with foams, sprays and rubbers. You can't can get past the slip, rip and drip factor, plus "I was away for the weekend and I left my Trojans in my other purse."
I hope this information is useful to you.