Just saw your comment on my talk page about the Harvard study article (I've been on jury duty the last couple of days). I actually think there are multiple issues with Andy's interpretation of the findings, but the overarching issue is this: he is convinced abortion causes breast cancer, and no amount of contradictory data will convince him otherwise. He (and he's hardly alone) will latch onto any issues with the study to suggest the whole thing is invalid. I do understand the stats pretty well, as I do health-related research and am well-trained in stats. I agree that it's silly to think that nurses wouldn't understand that spontaneous abortion = miscarriage. I'm not looking at the study right now but I'm not even sure if we know which term was used in the questionnaire. It's possible they used miscarriage but use spontaneous abortion in the article because it's medical terminology for a medical journal. Some of the issues that Andy raises are reasonable objections. The sample was not very ethnically diverse, though to accuse them of trying to hide that is also pretty silly. There are 2 more reasonable explanations - most likely they were conserving space to stay within the journal's page limit; also the details on the % of specific minorities don't change the interpretation substantially because there weren't very many of them. The issue of a quarter of the sample not responding fully to the abortion question is also a reasonable objection - it's unfortunate they weren't able to get this information. It is notable that they re-analyzed the data after excluding women who didn't respond to either of the questions and found the same results (last sentence of your quote), and that they found the same if they coded only some of those missing answers as non-abortions. Andy's ignoring these efforts by the authors to address the problem, presumably because they suggest that any confusion or embarrassment that caused people to leave the question blank didn't change the results. I'm not sure how well I addressed your question - if there's anything I didn't get to I'd be happy to try to clarify it. [[User:Murray|Murray]] 20:37, 3 May 2007 (EDT) | Just saw your comment on my talk page about the Harvard study article (I've been on jury duty the last couple of days). I actually think there are multiple issues with Andy's interpretation of the findings, but the overarching issue is this: he is convinced abortion causes breast cancer, and no amount of contradictory data will convince him otherwise. He (and he's hardly alone) will latch onto any issues with the study to suggest the whole thing is invalid. I do understand the stats pretty well, as I do health-related research and am well-trained in stats. I agree that it's silly to think that nurses wouldn't understand that spontaneous abortion = miscarriage. I'm not looking at the study right now but I'm not even sure if we know which term was used in the questionnaire. It's possible they used miscarriage but use spontaneous abortion in the article because it's medical terminology for a medical journal. Some of the issues that Andy raises are reasonable objections. The sample was not very ethnically diverse, though to accuse them of trying to hide that is also pretty silly. There are 2 more reasonable explanations - most likely they were conserving space to stay within the journal's page limit; also the details on the % of specific minorities don't change the interpretation substantially because there weren't very many of them. The issue of a quarter of the sample not responding fully to the abortion question is also a reasonable objection - it's unfortunate they weren't able to get this information. It is notable that they re-analyzed the data after excluding women who didn't respond to either of the questions and found the same results (last sentence of your quote), and that they found the same if they coded only some of those missing answers as non-abortions. Andy's ignoring these efforts by the authors to address the problem, presumably because they suggest that any confusion or embarrassment that caused people to leave the question blank didn't change the results. I'm not sure how well I addressed your question - if there's anything I didn't get to I'd be happy to try to clarify it. [[User:Murray|Murray]] 20:37, 3 May 2007 (EDT) |