Tuesday, June 12, 2012

On Reading a Graph

Long ago, I taught a course on statistics, and math modeling to medical students and faculty (the faculty refused to do their homework). One lecture set was on "how to read a graph." This is apparently still a mystery to many people.

In the Gallop poll data, the number of people who could be broadly called creationists were about 82% in 1982. This includes nearly anyone religious who thought that God either created Adam and Eve less than 10,000 years ago, or that God allowed/directed evolution to do the work. That year, only 9% accepted the fact of human evolution without any divine intervention. This left 9% undecided. The data are remarkably stable until 2000. Then, a trend lasting until today appeared. In 2001, about 3% of the "undecided" became acceptors of evolutionary biology regarding the origin of humanity. Was this a "post millennial" bounce for rationality? The number of creationists did not increase. From 2001 to 2006, the undecided fraction shifted another 1% to the evolution column. More significantly, from 2007 until today, there has been a decline in the percentage of creationists, and an increase in the number of Americans who agree that, "humans evolved, but that God had no part in the process." The number of "undecideds" dropped, but the deficit also went to the creationists, particularly those who believed that, "God created humans in their present form at one time within the last 10,000 years." This was clarified in the 2008 data which showed a restoration of the "undecided" percentage, and continued decline in creationists.

In the most recent Gallop poll, 78% of respondents could be said to be creationists, at the extremely broad criteria used earlier, and 7% are undecided. A mere 1% (within polling error) of "evolutionists" has shifted to "undecided." The only striking observation is that the "creationists"- most loosely defined- have not increased! In fact, there is a 4% deficit from the ~82% which had been essentially stable from 1982 until 2007, and additional weakness since 2008. On the other hand, the ~6% increase in pro-science thinking seems stable.

The question then arises, "Should theistic evolutionists be lumped with the young earth creationists?" I'll argue the negative in a new post.


EastwoodDC said...

Note that all the variation is within a 2 standard deviation error bound. The recent spike up isn't particularly meaningful (unless it continues).

Gary S. Hurd said...

Yeah. Almost all time series data should be looked at with a smoothing function as well.