« Fighting the Competition, One Legislature at a Time | Main | Biofuel Update: They Still Suck »

Selection Bias

I thought it was kind of interesting that upon reading this McKinsey & Co study (currently the top one in the list) on education, Kevin Drum and a number of other left 'o center blogs pulled out this one chart to highlight.  It shows starting teacher pay  (i.e. out of college) as a percent of the economy's average)


The author's of the study argue that the countries higher on this list also have better student results.  Now, I will confess that this is a pretty interesting finding in the study -- that starting teacher pay is more important than teacher pay in later years, because the key is to attract talented people right out of college away from other professions.  Interesting. 

But here is the quite fascinating selection bias by the lefty blogs:  I have read the whole report, and this is absolutely the only chart in the whole study that in any way, shape, or form might be interpreted as a call for higher government education spending.  Even more interesting is what these bloggers left out.  This is the other half of the starting teacher pay analysis Drum et. al. chose note to include, and makes clear that even this chart is not a call for more total spending:

South Korea and Singapore employ fewer teachers than other systems; in effect, this ensures that they can spend more money on each teacher at an equivalent funding level.  Both countries recognize that while class size has relatively little impact on the quality of student outcomes (see above), teacher quality does.  South Korea's student-to-teacher ration is 30:1, compared to an OECD average of 17.1, enabling it in effect to double teacher salaries while maintaining the same overall funding level as other OECD countries....

Singapore has pursued a similar strategy but has also front-loaded compensation.  THis combination allows it to spend less on primary education than almost any other OECD and yet still be able to attract strong candidates into the teaching profession.  In addition, because Singapore and South Korea need fewer teachers,  they are also in a position to be more selective about who becomes a teacher.  This, in turn, increases the status of teaching, making the profession even more attractive.

Whoops!  Don't want our friends at the NEA to see that!  Most of the study turns on McKinsey's finding that teacher quality drives student results, way ahead of any other factor, from class size to socioeconomic background:


Well, now the NEA might be getting really nervous.  Something like this might cause parents to do something rash, like demand that low-performing teachers get fired.  Gasp.

Anyway, to get back to the cherry-picking and selection bias issue, the study is pretty clear that it thinks that "more spending" is a failed strategy for improving public education

If school choice is off the table, then I would be very supportive of a program to increase starting teacher pay, funded by larger class sizes and substantial reductions in useless administrators and assistant principals.  Anyway, it is kind of an interesting study, though you may find the pdf file format really irritating to try to read.  Lots of funny formatting. 

Posted on April 3, 2008 at 09:22 AM | Permalink


OECD average of 17 kids per class? Wow! How much is the American average then, 10-12? When I was in school (in Hungary) usually we've been 26-33 kids in a class, with 30 being the target number thy tried to maintain and it wasn't a problem at all. Most kids don't really ask questions anyway.

Posted by: Miklos Hollender | Apr 9, 2008 7:21:24 AM

Again, no one actually can dispute the real statistics. Facts are stubborn things.

1. The average test scores of education majors is amongst the lowest of any other major.

2. When annualized to a full 12 month year teachers starting pay and average/median pay compares favorably with other professionals. If you look at total compensation after including the generous benefits and retiremnt packages teachers do very well.

3. The teaching profession is protected by tenure which makes it virtually impossible to fire teachers that have attained tenure. Tenure protection is something other professionals do not have.

4. If the average teacher is working so hard, the results clearly do not indicate it. Drop out rates and the overall quality of education given the cost, leave much to be desired.

5. Teachers and their unions, supported by the Democratic Party, have opposed every initiative to improve education, save spending more money on teachers. They will only support initiatives like reduced class size, which studies have shown do not improve education, because it will increase the union's constituency. The study above shows that improved teacher quality is perhaps the biggest contributor to improved educational results, yet the teacher's union opposes any effort designed to do this like merit pay, clinging stubbornly to the blue-collar like union wage scales that simply reward time of service and not quality.

Posted by: mark | Apr 9, 2008 11:34:18 AM

The comments to this entry are closed.