Some of the points from Coyne's blog bear repeating:
Coleman's article came out shortly before the expected release of the full report on the mental health consequences of abortions by the Royal College of Psychiatrists. Fortunately, anyone interested can access a draft copy of that report from the Internet. The long document is quite clear and excruciatingly detailed in describing the methods leading up to its conclusions, with a precise description of the search strategy used to locate studies so that anyone could replicate it; identification of which studies were included and which were excluded; how studies were evaluated; and the strength of the evidence for the final conclusions. In striking contrast, the Coleman article has no such transparency and we are left either to accept her conclusions or puzzle how she reads them.
The Steering Group for the Royal College report consisted of 11 members, mostly physicians, and all were required to declare before each meeting any conflicts of interest, not only financial, but also whether they had made public statements for or against abortion or held any membership in organizations that had public views or direct interest in abortion.
Back to Coleman's review, the Royal College of Psychiatrists report indicates that in their evaluation of the literature, they considered a number of studies conducted by Coleman herself and included in her review. They considered all of the Coleman studies that they reviewed to be methodologically poor. One of her own him him papers that she included in her own review was rejected as having no usable data. A number of her papers were rejected because they had inappropriate control/comparison groups and still other papers were rejected because they used inappropriate measures of mental health after the abortion. In addition to its other limitations, still another Coleman paper was rejected because of a lack of statistical control for mental health prior to the abortion. The Royal College of Psychiatrists report did include one of Coleman's papers, but with serious criticism. You can check all this out for yourself by examining Appendix 6 of the Royal College of Psychiatrists report.
What is the issue with Coleman's choice of comparisons/control groups? She relied heavily on comparisons between women with received an abortion for unwanted pregnancy and women who had completed a pregnancy that was planned or wanted. Of course, this strategy cannot get at the effects of abortion because presumably the circumstances of having an unwanted or unplanned pregnancy are different from the circumstances of completing a wanted pregnancy. It is not surprising that women having a wanted pregnancy have better mental health than women who did not want to be pregnant. Do you think the Coleman knew what she was doing and proceeded anyway? Even before the Royal College of Psychiatrists report, Coleman's work had received serious criticism and summary rejection, so that she had to be aware of criticism before including the studies in her review....
Coleman either deliberately did reckless things with her statistical calculations or did not understand basic statistics. For instance, near the end of the results section of her review, she states that the population attributable risk (PAR) associated with abortion is 10% of all mental health problems and 34.9% of all suicides in women of reproductive age. PAR is intended to represent the proportion of such problems that can be causally attributed to abortion. Coleman's interpretations are ridiculed in the E-letter responses to her article, but I think you can see that they are patently absurd. If you need assistance in interpreting PAR, please consult the the E-letter responses for excellent, succinct explanations of what is wrong with what Coleman did.
Let's pause there to let that claim sink in: Coleman is saying that 35% of
all suicides by women of reproductive age are caused by abortion. How such a ludicrous claim got by a reviewer is beyond me. Coyne continues:
If you examine Figures 1 and 2 in Coleman's review, you can see that she counts each of her own studies multiple times in her calculation of the effects attributable to abortion. This practice was also roundly criticized in the E-letter responses to her article because each study should only be entered once, if the conditions are met for integrating results of studies in a meta-analysis and providing a test of the statistical significance of the resulting effect size. This may sound like a technical point, but it is something quite basic and taught in any Meta-Analysis 101.
Coleman's calculation of overall effect sizes for the negative mental health effects of abortion involve integrating multiple effects obtained from the same flawed studies into a single effect size that cannot accurately characterize any of the individual effects - anxiety, depression, substance abuse, and suicide - that went into it. Again we are encountering a nonsensical statistic. In meta-analysis workshops and basic courses, it has become standard practice to include a slide allegedly of a road sign announcing that you are entering New Cuyama, California on which a population of 562, an elevation of 2150 feet, and a date of establishment of 1951 are summed up to a total of a total of 4663. This is meant to be a spoof of the meaninglessness of combining diverse outcomes into a single summary effect size.
Absolutely shameful.
by 
