Friday, July 07, 2006

Bit more on Tickle IQ scores

Infant mortality correlates more stongly with national average IQ as estimated in Lynn and Vanhanen's IQ and the Wealth of Nations than any other variable I'm aware of except for life expectancy (.84 and .85, respectively).

Does that relationship hold domestically? Unfortunately, the data on IQ by state are sparse. Tickle's results may approximate better than anything else out there--they correlate with GSP per capita at .53, white NAEP science scores at .63, Project Talent scores at .70, and my 'education index' (state's percentage of bachelor's degree and beyond minus state's percentage of less than high school diploma or equivalent) at a very high .86 (all statistically significant). They also correlate inversely with white infant mortality rates by state at .61 (unfortunately there are no recent data for white infant mortality in DC, the Tickle Test paragon).

The Tickle range for the fifty states plus DC is only five points, all presented as whole numbers. That the correlations remain so robust seems to evince the general accuracy of the results. I can only speculate on how they would be enhanced with numbers expanding out a couple of decimal places, but when I followed a friend's suggestion and contacted Tickle, a customer service rep told me it was proprietary information and then, sensing my frustration, advised me to see if I could find out more at Mensa. But he didn't expound and I found nothing there.

(Human biodiversity)


FuturePundit said...

I've been exploring

What I did: Search a state just with the state name entered. That'll give you every school the state. Then sort by math or reading score. Then change to showing 100 items at a time. Then page forward to see how many schools of the total have at least 90% average proficiency.

Well, South Dakota is much smarter than Wyoming or Nevada by this measure.

Nevada has only 4 schools out of 494 with over 90% math proficiency. 1%. Wyoming appears to be considerably dumber than South Dakota.

South Dakota has 542 schools with 3 unrated for math (and why are some unrated). 64 are 90+% math proficient. That's 11% that are in the top tenth versus 1% for Nevada.

Problem is that school sizes differ. Is the average lower scoring school bigger than the higher scoring ones?

I wonder if there is a way to get an aggregated state level percentage math proficient. Ditto English. Also, sorted by grade level. Match proficiency in 8th or 9th grade probably means the most. After those grades high school drop-outs disappear from the measures and make comparisons less useful.

crush41 said...

Wow, thanks for the website.

Poking around in my home state (16% are at 90% or above, btw), it appears that smaller schools overwhelmingly perform better than larger ones. For example, for schools with 0-199 students, about 70% of Kansas schools are above 90% in math proficiency. For schools with 1000-2999, the best performing school only enjoys 78% math proficiency (you can filter results by size).

Still, by looking at schools within a specific size parameter and comparing it across states would probably yield pretty reliable data for comparison. Problem is, if you filter by size, only the first 100 schools are accessible!

I wonder what qualifies as "proficient". In Kansas, for example, schools with high proportions of economically disadvantaged students score phenomenally well in some cases (100% for 85% economically disadvantaged?!)

crush41 said...

The proficiency rates are based on state standards. Not surprisingly, the rates trend upward (to meet NCLB standards no doubt), and probably not because the population is getting smarter or educator's are becoming more effective (see the graphics showing that on state reading tests, 8th grade Kansans have improved from 2002 to 2004, but on the NAEP reading test, 8th grade Kansans have declined in ability).

Unfortunately, when it comes to comparing schools, the NAEP is the best we have.

Another interesting piece: Kansas failed to make adequate yearly progress because "students with disabilities" in the state didn't increase enough in reading ability!