[quote]pat wrote:
Both show a reduction. One shows a reduction in teen birth the other shows a reduction in pp clinics. There being a wiggly line on the pp graph does still show that the trend is downwards.
You cannot disassemble the correlation when you display data that shows both trending down. That is more indicative of correlation than it is not.[/quote]
They do not. Look at years of concurrence and then come back. The PP Clinic graph shows effectively zero net change over the years of concurrence. This is not a reduction. It is a spike and then a drop back to the level that it began with. Over this time, the teen pregnancy rate does not spike and then drop, and does not show a net change of zero. It simply falls.
To make a larger point, the “look at what both graphs are doing–falling” argument–which, as I’ve proved again and again, is factually inaccurate for the years of concurrence between the two graphs–is not how correlation is suggested. Correlation implies that when one data set trends in a particular, statistically-significant direction over a number of time-intervals, another data set will follow. This is absolutely, unequivocally not borne out in the data that I presented. The teen pregnancy rate never rises between 1989 and 2005, whereas the number of PP clinics absolutely does on a number of occasions. This, again, is absolutely not evidence of correlation. It is precisely the opposite.
Let me try it this way. The following information is going to be presented for each year of overlap between the two graphs, beginning in 1988 and ending in 2006 (which represents the entirety of their concurrence), in the following manner: [rise/fall in number of CC clinics], [rise/fall in teen pregnancy rate], [X for identical trend, O for unidentical trend]
RISE, RISE x
FALL, RISE o
RISE, FALL o
RISE, FALL o
NO CHANGE, FALL o
FALL, FALL x
RISE, FALL o
FALL, FALL x
NO CHANGE, FALL o
FALL, FALL x
NO CHANGE, FALL o
RISE, FALL o
NO CHANGE, FALL o
NO CHANGE, FALL o
FALL, FALL x
FALL, FALL x
RISE, FALL o
Overall, there is effectively no change in PP clinic number and a large change in the teen pregnancy rate. If you want to get into specifics–which is problematic in its own right–there are six year-long intervals during which the two data sets trend in the same direction, and eleven during which they trend in opposite directions (6 years) or unidentical directions (5 years).
Now, there is absolutely no way to draw correlation from a juxtaposition of these sets of data. None whatsoever.