Aren't US and European women likely to be equally aware and capable of doing self-exams in your opinion?
We already discussed this. Didn't we agree that European women are more likely to do self-exams because they aren't afforded mammograms as frequently?
In any case, the difference between them doesn't matter. What matters is that women being diagnosed from self-exams should have been but were not excluded from the data comparison between the US and Europe because we are only interested for the purpose of her argument in lives saved via mammogram for women in their forties. Women of whatever age who were diagnosed on the basis of self exams don't count. They are a confounding variable. Proper analysis isolates the significant variables and excludes confounding variables.
You haven't verified that she was off by a factor of ten.
No, I haven't. I figured it was enough to demonstrate that she was off by a considerable amount, although that wasn't critical to my point, just a bit of a ramble, so I regret having mentioned it. The fact is that you and I don't have the data to do do a computation any more than the author had the data to come up with eleven percent. I was just speculating that she might be off by a factor of ten or more and was trying to illustrate how one can tell intuitively that she was off by a whole lot. As is easily recognized because it is familiar, one designated hitter in a slump doesn't bring down the team's batting average very much. Likewise, half the forty-something women saved by mammograms wouldn't likely bring the survival rate down by very much.
Regardless, for my purposes, it doesn't matter if she was off by a lot or a little. That was just a peripheral observation. What matters is that her methodology for computing her percentage was totally wrong-headed due to apples and aircraft carriers.
I'm sorry but it sounds pretty close.
If it still does, then nothing I can say will disabuse you of that notion. I already explained as best I can why it wasn't remotely the same. I said:
"What she did was equate said forty-something US women with all European breast cancer victims regardless of age, mammogram, or any other variable."
That is not the same as, "if forty-something women in Europe got mammograms at the same rate as they do in Europe, then the European forty-somethings getting the mammograms would be saved at a similar rate, which is one in two thousand." Surely you can see the difference between ALL European breast cancer victims and SOME European women, namely the forty-somethings who would benefit from the mammograms they don't now get. "All" and "some" are not the same. It's like the difference between all SI members, on one hand, and conservaive SI members who post on this thread, on the other. The former group is huge in comparison with the latter group and they have different characteristics, aka lotsa variables.
I can't be any clearer than that.
But its clear you think the difference in survival rates is mostly due to things other than earlier average detection here. I don't know that. I rather think early detection is a primary reason for our higher survival rate.
It doesn't matter what either of us thinks that the difference might be. The point is that we don't know, exactly. Neither you nor I nor she could possibly have the data to figure it out. Had she simply speculated that the survival rate might go down by ten percent or so, I wouldn't have made an issue of it. But she didn't speculate. She claimed that the decrease in the survival rate would be exactly the same as the difference in overall survival rates between the US and Europe which makes about as much sense as saying that the decrease in survival rate would be exactly the same as the difference in exchange rates between the Canadian and US dollar.
you compare early detection due to more early mammograms to one player on a baseball team. I think its more significant than that.
I thought of that when I wrote the analogy but was trying to keep it simple. So, make it three or four players on the team if it helps you focus better on the punch line of the analogy. There are 24 players on a baseball team so the proportion of players in a slump would be closer to the percentage of over forty women who are in their fourth decade.
Now, do you really think that the White Sox batting average would go down by the exact same amount as the difference between the White Sox batting average and the Red Sox batting average? Or the difference in player salary between the White Sox and the Red Sox? Or the number of patrons their home stadiums hold? It doesn't matter if there are one or three or ten players in a slump in this analogy because the decrease in the White Sox batting average has nothing to do with the Red Sox batting average. It's like those trick questions where they ask, "if a train is heading east at 50 miles per hour and another train is heading west at 40 miles per hour, how much snow is falling in Duluth?".
It's easy to get sucked in by those things. That's why that
technique I complained about, entheymematic argumentation, works so well. The author gives you one sentence, then makes some point. You assume that the leading sentence has something to do with the point and fill in any gaps to make the two logically fit. You don't look critically at it and say, "huh, those two things don't go together." In this case, claiming that the survival rate would go down 11 percent makes as much sense as her claiming that there were 18 inches on the ground in Duluth, IOW, no sense.
Like I said, it's easy to get sucked in. But when the disconnect is pointed out, and over and over in a variety of ways, to boot, people usually react, "oh, yeah, I missed that." They don't continue to argue about the depth of the snow in Duluth. <g> |
