Call me crazy, but I’d have to say that one of my favorite publications of all time is a National Center for Education Statistics report mapping state standards onto NAEP, allowing comparison of where state proficiency benchmarks align with NAEP scores. I’ve likely provided links to that report more than a few times in my blog. Well, they’ve done it again. The new MAPPING STATE PROFICIENCY STANDARDS report is out, and you can find it here:

http://nces.ed.gov/nationsreportcard/pubs/studies/2010456.asp

A frequent, but much debated conclusion from teacher quality research is that teachers’ own academic ability, measured by test scores or even more bluntly the “competitiveness” of the colleges teachers attended as undergraduates, is associated with student outcomes. This occurs even when we use such crude classifications as the Barrons Guide rating system. In recent, exceptionally methodologically strong piece, Boyd, Lankford, Loeb and Wyckoff found:

“Furthermore, almost half of the teachers in the most effective quintile (based on student outcomes) graduated from a college ranked competitive or higher by Barron’s, compared to only ten percent of the teachers in the least effective quintile.”(p. 23)

http://www.teacherpolicyresearch.org/portals/1/pdfs/Matching_of_Public_School_Teachers_to_Jobs.pdf

Okay, for some this may hurt, and smacks of elitism. But nonetheless, it is a strong and relatively consistent finding which we should likely give some attention.

This finding is highly relevant to Arne Duncan’s talk yesterday at Teachers College, Columbia University where he took aim at the role of university based preparation programs in Education. Notably, Duncan referred to related work by these very authors, but did not mention this finding. In an effort to be egalitarian, Duncan promoted the virtues of institutions like Teachers College but also those like Emporia State in Kansas.

As I noted in my previous post, one thing we know is that the majority of teachers come through relatively non-competitive undergraduate colleges (on a 6 point rating system, from non-competitive, less competitive, competitive, very competitive, highly competitive, most competitive). On average, public school teachers come from the less competitive and competitive categories (about 2/3 of all teachers in these two categories alone) far more so than the highly and most competitive (about 6.5%). So too do most college students generally. That’s just the way the higher education system is distributed.

Teachers also come in large numbers – 42% – from 1994 carnegie classification – Comprehensive I – colleges and increasingly from less selective liberal arts colleges (carnegie 1994 Liberal Arts II colleges). And these Carnegie classifications from 1994 are somewhat associated with Barrons ratings. In short, the system of teacher education, nationwide, is not set up to produce large numbers of teachers who have the attributes that authors above find to be associated with higher student outcomes. A teacher preparation or administrator preparation program is only as good as its students.

Other authors have argued that the dominance of non-selective colleges in preparing teachers and higher costs of pursing teaching through more selective colleges creates a disincentive for academically strong high school students to pursue teaching. Add this to relatively low salaries, and the problem is exacerbated. Perhaps its the PIPELINE and system as a whole and not so much the individual institutions and prep programs that need reforming. Perhaps we need some incentives to encourage academically talented students to pursue teaching and some incentives to encourage the “highly and most competitive colleges” to get in the game of teacher preparation. At the same time, we may need to make some tough policy decisions about academically weak undergraduate and graduate institutions which have increased their role over time.

An interesting twist related to teacher academic preparation is that principals with stronger academic backgrounds seem more likely to recruit and retain teachers with stronger academic backgrounds (http://eaq.sagepub.com/cgi/content/abstract/41/3/449) So, we’ve got to find some way to get stronger principals into schools where they are needed most, and make sure there is a supply of stronger teachers produced through a better pipeline, from which those principals can build strong teams.

One problem here is that rather than becoming more concentrated in strong academic institutions over time, educational administration programs have become more distributed across more diverse… and quite honestly academically weaker institutions. For example, between 1993 and 2003, comprehensive colleges went from producing about 3% of education leadership doctorates to about 25% (http://eaq.sagepub.com/cgi/content/abstract/43/3/279).

So why does that matter? How does this finding relate in any way to the fact that principals with stronger academic backgrounds (measured crudely by Barrons ratings of undergrad colleges) are more likely to hire teachers with stronger backgrounds, and those teachers are shown to make a difference? Isn’t it likely that there exists no relationship between graduate preparation and undergraduate preparation, and that we should be unconcerned that comprehensive colleges are the ones producing the doctorates? Well, again the dots connect logically. As it turns out, we show in the same article above that about 22 to 25% of doctoral recipients from Top 20 ranked (US News, of all things) education schools and 13% to 15% of doctoral recipients at all Research Universities attended highly or most selective undergraduate colleges, compared to only 5% to 10% of doctoral recipients in comprehensive colleges.

Yes, these are relatively harsh and elitist realities. And yes, I am implying that having a strong academic background is likely an important attribute for someone who wishes to lead an educational institution. That seems to make sense.

In his speech yesterday, Arne Duncan invoked the usual comparison to medical training:

http://www.ed.gov/news/speeches/2009/10/10222009.html

The point, of course, was to emphasize the importance of clinical training. But, let us not forget that the medical model relies on two critical prerequisites to clinical training – 1) highly selective entrance criteria and 2) successful completion of rigorous undergraduate + 2 years of rigorous content upload of basic sciences and other relevant curriculum. Without academically strong candidates to begin with, the model fails. Without rigorous up-front information uploading, and students who can handle it, the model fails. The medical model is equally reliant on all of its parts, not just the clinical training.

Just connecting some dots here. Cheers.

Secretary Duncan is again on the stump today, at Teachers College (where I attended) where he is expected to make the case that education schools are “cash cows” of the university, generating large sums of tuition revenue which are then diverted to other parts of the university.

http://www.ed.gov/news/speeches/2009/10/10222009.html

This proposition is hardly new, and appears to come from the pages of past TC president Arthur Levine in his report on ed schools a few years back.

http://elan.wallacefoundation.org/SiteCollectionDocuments/WF/ELAN/2007%20Second%20Half/EducatingSchoolLeaders.pdf

(this is the one on preparing school leaders. there was also one on teacher education)

At the time, my colleagues and I were intrigued by a number of the assertions being made and engaged in a series of research projects trying to untangle the “realities”, but actually did not explore specifically the cash cow notion. But, our research from that time does have a few facts to offer with respect to the cash cow argument, as well as important general context issues.

First of all, who is producing the teachers, and administrators? One implication of the current rhetoric is that major universities like state flagship universities and major private universities which offer a diverse array of programs, undergraduate and graduate majors are also producing large shares of all teachers. As it turns out, the major research universities actually produced about 13% to 15% of teachers who were working in public schools in 2003-04 and 2007-08. 42% of public school teachers received their undergraduate training at regional comprehensive colleges, many of which were the former “normal schools” or “teachers colleges.” In many of these schools, education majors are the dominant major, perhaps helping to sustain the institutions, but with a minority of other program areas to draw on education tuition dollars – except by the role that liberal arts and science departments play in providing undergraduate credit hours to teachers in their content areas. But, this is revenue received for credits delivered – not a redistribution of profit margin, per se.

The role of education schools in major research universities is potentially more interesting, but again, ed schools in research universities produce a relatively small share of all teachers and that share appears to be declining. The same is true of graduate degrees in educational administration. In the early 1990s, regional comprehensive colleges produced about 3% of doctorates in educational administration, and now produce about 25% or more (as of 2003). That is, graduate degrees in educational administration are being increasingly produced by institutions whose primary goal is to produce educators and education related professionals. So, from these perspectives, it’s getting harder to see how ed schools or programs are substantially subsidizing other schools or programs within universities, when increasingly, the production of educators and educational leaders is being concentrated in schools focused on education.

I’m unsure whether there’s other evidence to contradict this pattern. I’ve not studied it for a few years. There is some evidence that small cash strapped formerly undergrad only liberal arts colleges have expanded delivery of online certificate programs including administrative masters degrees, but they are hardly a major producer yet. In fact, they’ve expanded production in areas such as MBAs even more so than teacher and administrator education.

Now, on to the basic premise laid out by Levine and echoed now by Duncan, that ed school tuition dollars subsidize the rest of the university. This could be the case if tuition was constant for a credit hour across all students in all units in the university and if the average cost of providing a credit hour to undergrads was lower for education students than for other students in the university. One might imagine this to be the case, if we assume that education faculty are simply less well paid, for example, than engineering, business or economics faculty and that ed school classes are large. Actually, the bigger driver of cost per credit hour produced is the class size piece.

A few years back Chris Morphew and I did an analysis of data from the National Survey of Postsecondary Faculty, estimating wage models and models of “cost per credit hour” by field in which those credit hours were delivered. We accounted for relative salary of similar rank faculty, share of salary to teaching and class sizes of average undergrad load of teaching faculty. We actually found that ed school credit hour costs were about average, comparable to business for that matter. While b-school salaries were higher, ed school class sizes were smaller, on average across the full range of undergrad courses. Next, we linked our credit hour cost estimates to course taking data on students in different majors to come up with estimates of the relative cost of producing and ed major versus an econ major, etc. based on the full mix of courses students take across units in a university and the relative price of those units. Again, cost of producing an ed major was relatively average – not low.

Now, there are factors we could not and did not consider with our limited data – including shares of undergrad credits delivered by teaching assistants  and whether this rate is significantly higher, or lower for ed schools. We also were unable to generate estimates of other “overhead” costs such as equipment that might be necessary in engineering or sciences, but this would hardly seem to compromise comparisons between ed schools and other areas such as social sciences that would seemingly have comparable non-faculty expenses.

That said, I’d be curious as to what other evidence is now out there to support, or refute this assertion that Duncan is now making. Here’s my preliminary reading list for anyone interested:
Morphew, C., Baker, B.D. (2007) On the Utility of National Data for Estimating Generalizable Price and Cost Indices in Higher Education. Journal of Education Finance 33 (1) 20-49

Baker, B.D, Orr, M.T., Young, M.D. (2007) Academic Drift, Institutional Production and Professional Distribution of Graduate Degrees in Educational Administration. Educational Administration Quarterly 43 (3)  279-318

Baker, B.D., Wolf-Wendel, L.E., Twombly, S.B. (2007) Exploring the Faculty Pipeline in Educational Administration: Evidence from the Survey of Earned Doctorates 1990 to 2000. Educational Administration Quarterly 43 (2) 189-220

Wolf-Wendel, L, Baker, B.D., Twombly, S., Tollefson, N., & Mahlios, M.  (2006) Who’s Teaching the Teachers? Evidence from the National Survey of Postsecondary Faculty and Survey of Earned Doctorates.  American Journal of Education 112 (2) 273-300

Here’s a link to a new report on the effect of the economic downturn on state tax revenues. Particularly interesting is the table ranking overall budget impact across states on Page 20 (Table 12).

http://www.rockinst.org/pdf/government_finance/state_revenue_report/2009-10-15-SRR_77.pdf

Here’s a link to a solid, recent dissertation out of the University of Michigan regarding New Jersey’s Abbott School districts and the effects of Abbott litigation on overall outcomes and outcome gaps.

http://deepblue.lib.umich.edu/bitstream/2027.42/61592/1/aresch_1.pdf

Here are some highlights quoted from the Intro of the dissertation:

The first essay is an empirical analysis of the effects of the Abbott school finance reform on educational expenditures in New Jersey. This reform dramatically increased the funding available to poor, urban schools with the goal of improving achievement in those districts. My analysis suggests that districts directed the added resources largely to instructional personnel. They hired additional teachers and support staff.

The second essay asks the obvious next question: Did this increase in funding and spending improve the achievement of students in the affected school districts? I focus primarily on the statewide 11th grade assessment that is the only test that spans the policy change. I find that the policy improves test scores for minority students in the affected districts by one-fifth to one-quarter of a standard deviation.

Resch’s third essay (following the common 3 essay structure of econ dissertations) is on higher education.

Resch’s analysis is not entirely uncritical of Abbott reforms. Resch expresses concerns that emphasis in later rounds of Abbott on early grades reforms may have diverted resources to early grades education at the expense of upper grades (where early gains had been shown – see above). Resch also questions whether the approach to financing Abbott’s was sustainable, but notes that SFRA introduces significant structural changes to school funding in NJ (I could go on about this part, but not now). Resch is most critical of the state’s own lack of efforts to generate and maintain data useful for evaluating the effects of reforms – either prior Abbott reforms or SFRA.

In any case, if you happen to be trying to decide whether to spend an hour or two with a sound analysis of actual data treated with reasonable rigor, or whether to go see the Cartel Movie filled with misguided and intellectually sloppy assumptions, misleading if not outright fabricated numbers, I would recommend the econ dissertation. Yeah… it’s not as sexy – doesn’t have the slick production – and won’t be in a theater near you. But you can just click and download above, and actually get some decent real numbers and analyses. Cheers.

If you do read my previous posts regarding the Cartel, please see all 3, along with the original materials that prompted me to post (http://www.njecea.org/cartel/?page_id=10)

Post 1: https://schoolfinance101.wordpress.com/2009/05/30/idiot-of-week-award-the-cartel-check-this-out/

Post 2: https://schoolfinance101.wordpress.com/2009/06/02/i-just-cant-let-go-of-this-one/

Post 3: https://schoolfinance101.wordpress.com/2009/06/17/vacuous_bowdon/

Note also that if you disagree with my framing of the film’s objectives, please see the “crisis” page here, where the movie is framed largely as I have framed it.

I had relatively modest expectations for the Gannett series in New Jersey on state taxes to begin with. Sadly, this series managed to fall way short of these expectations by trying too hard to construct the point that New Jersey’s taxes are simply way out of line and that, for example, NJ would much better off if it behaved like all of those  smart, business friendly states out there like Delaware, Maryland, Missouri, Georgia and Washington. These states are indeed strange bedfellows.

http://www.app.com/article/20091002/NEWS/310020010/NJ+Tax+Crush++How+five+states+keep+their+tax+burden+down

First, lets get some numbers squared away. I like the fact that this article listed in the right margin, the average state and local per capita taxes, the average state per capita expenditure, and the average per capita income. Note that this second number should have been state and local spending. More importantly, the article failed to do the last calculation and related ranking – state and local taxes as a percent of personal income (perhaps I missed it). That is, how does our effort compare, given our income?  http://www.taxpolicycenter.org/taxfacts/displayafact.cfm?Docid=531

Typically, when looking at taxes as a share of income, NJ is not among the top few states. Above average, yes, but not the top. More specifically on the question of public education spending… state and local… NJ does rank second in the percent of Gross State Product spent on K-12 public school districts. By the way, Vermont is way ahead of us. But, as I have previously argued, to a large degree though not entirely, you get what you pay for. And the five states listed in the article as model states seem to get what they pay for.

On this previous post, I explain how “small business friendly” states tend to have particularly weak public school systems, if we assume small business friendliness to be only a function of low taxes and not high quality services. https://schoolfinance101.wordpress.com/2008/12/17/state-rankings-small-businesses-school-quality-and-economic-productivity/

As far as I’m concerned, having good schools is probably a critical element of business friendliness at least if you have any dreams of recruiting and retaining well educated employees who want their children to also be well educated.

So, what about those miraculous low tax states identified in this article? Within their regions, all but Georgia (which has less competition for bragging rights, and is relatively average) have relatively poorly funded public education systems and to a large extent, the outcomes to go with it.

Delaware has the lowest percent of school aged children even in the public school system. Only 77.6% of 6 to 16 year olds in Del. attend the public schools according to American Community Survey data from 2005 to 2007.  Delaware’s education spending, however, is relatively average or slightly better than average among states (after adjustments for competitive wages, size, location and relative poverty) and its outcomes are relatively average too. Yes, NJ spends more on schools than Delaware, and NJ serves a large share of its children in those schools, and NJ children generally outperform DE children on a variety of assessments.

http://www.nces.ed.gov/nationsreportcard/statecomparisons/Default.aspx?usrSelections=1%2cMAT%2c4%2c0%2cwithin%2c0%2c0

Quite surprisingly, Washington operates one of the least well funded state school finance systems in the nation. And Washington provides negligible additional support for higher poverty and urban schools (not always one and the same). And, Washington also does less well on outcome measures.

(I have some graphs here from an earlier post on LA and MS: https://schoolfinance101.wordpress.com/2009/02/25/public-schooling-in-louisiana-and-mississippi/)

Yes, within region, combining South Atlantic and Gulf Coast states, Georgia is relatively well funded (though again, not supporting poor and urban districts) compared to its peers and does outperform many states in its low performing region. But, Georgia performs much less well than NJ (okay… not a fair comparison… but that’s partly the point). GA also performs less well than Texas or Kentucky for that matter (in graphs in LA & MS post above).

Missouri, like Pennsylvania has made some efforts of late to straighten out their school funding mess into a more logical formula. But the fact remains that poor inner urban fringe and poor rural districts in Missouri have been starved of educational resources for decades and have the outcomes to match. On average, Missouri falls into that consistently below average category on inputs and outcomes. The bigger story in Missouri are the disparities – not so much with the major urban centers themselves (though they are part of the picture) but with those others noted above.

Maryland is another state with fewer than 80% of 6 to 16 year olds even in the public school system. Maryland’s spending is relatively average as are its outcomes. Which is, perhaps fine for Maryland.

By contrast, as I have pointed out here – https://schoolfinance101.wordpress.com/2009/06/17/vacuous_bowdon/

New Jersey, perhaps in part because of its high investment in public schooling – a major public expense, has strong performance measures either compared against other states or if treated as a separate country as I discuss in my previous post. Yes, much is cyclical here. NJ is a richer and more educated state that, as a result, values education and has the money to pay for it, compared to these states.

Do we really want to be like DE, MD, MO, GA or WA? I do like the seafood in WA better. I just moved back to NJ (no not the other way) from living a few yards away from the state line with MO and had the pleasure of both working for MO on their formula redesign and testifying against them on the remaining disparities. But I’m glad to be back.

Sadly, I think that one of the main things that would be learned from mimicking the taxing behavior of these states is that you can have lower taxes if you simply want to sacrifice the quality of your public services – primarily public education.

That said, I will continue to rant against certain major organizational inefficiencies in New Jersey public schooling and government services – that is, the multiple municipal madness issue, which I have written about and speak about (slides here)  https://schoolfinance101.wordpress.com/2009/08/14/small-districts-racial-isolation-and-new-jersey/

Indeed there is real progress to be made on reducing organizational inefficiencies in local governance. I’m not sure that I hold out much hope for resolution here.

======

PS: I hope at some point in the near future to post some detailed numbers on the characteristics of individuals moving into and out of NJ. Last I checked, while there were net exits of middle to lower income families, there was a net inflow of individuals with graduate level education. This would conflict with the unfounded rhetoric I heard spewed on the radio the other morning that NJ is losing all of its hard working small business types and replacing them will undocumented immigrants. Then again, the undocumented part may be a bit hard to track.

In this article, Jay Greene and Marcus Winters present a grossly oversimplified perspective of what we really know about the relationship between state school finance systems – special education aid formulas – and state special education classification rates.

http://www.ajc.com/opinion/funding-may-push-special-145257.html?printArticle=y

This supposed problem plays out at two levels. First, it is assumed that states which allocate funding based on local school district rates of classifying special education students will see greater overall growth in special education student populations than states that a) provide flat funding per fixed share of students in each district, b) cap the number of students classified for which funding will be provided or c) use some other better measures of local district resident rates of children with “real” disabilities – a  measure outside the influence of local district classification procedures. Second, some go so far as to assume that not only are state average disability rates different solely because of local responses to differences in state fiscal incentives, but that local rates of disability classification within states also vary largely because of differences in the extent to which local school officials play the special education fiscal incentives game. Green and Winters seem to be speaking primarily on the first point – state average differences and headcount incentives.

There is a small body of research, some of which is pretty solid, that supports the notion that there is a relationship between fiscal incentives and classification rates. That is not to say, however, that such incentives explain most or all of the differences, as implied by Greene and Winters in their absurd Maine to California anecdote. Second, studies that show that classification rates are partially responsive to fiscal incentives do not address whether the incentivised classification rate may actually be closer to the true rate of disabilities than the non-incentivised rate. Without a measure of true prevalence it is difficult to make the leap that the incentive is necessarily a bad one – one that distorts inappropriately the classification rates and services for children with disabilities.

Further, removing the fiscal incentives entirely does not necessarily bring to a grinding halt the overall statewide growth in classification rates or the variations across districts – even if capitation or flat funding does create modest statistical differences in growth rates between states. Pennsylvania has provided flat, census based funding since the early 1990s, yet classification rates grew dramatically since that time and rates continue to vary widely across Pennsylvania districts, from about 5% to over 30%.

Further, as with most demographic characteristics, families of children with mental or physical disabilities are simply not uniformly distributed across neighborhoods, cities and towns within states or across states making it very difficult to say that the typical school district or state should have only X% of such children. Complicating the issue is that the uneven distribution of families of children with disabilities is endogenous to the quality of services available across communities within states and across states. So, for example, if a state provides generous funding based on actual needs of students and that funding leads to higher quality services for students, families of children with disabilities are more likely to consider relocating to those states. The same applies to more local moves where services vary across districts. And parents of children with disabilities may make these decisions based on more than the services provided by the school district alone. Large towns and small cities in otherwise rural areas tend to have elevated disability rates in part because of greater availability of social services and health-care services less available in surrounding areas.

So perhaps a state can export its children with disabilities to a neighboring state by adopting school finance policies that ensure low quality programming and limit district incentive to pursue diagnostic testing. And perhaps some of the differences we see across states – especially between neighboring states – are a function of these programming and service quality differences. This question is yet to be thoroughly addressed in the literature.

In any case, it is a huge unwarranted stretch to argue that state limitation of funding for special education necessarily leads to a more correct identification rate of children in need while holding constant (or even improving) the quality of programs and services and while not exporting children with disabilities.

The problem for state policymakers is to find the correct balance between sensitivity to the needs of individual children as identified by those charged with providing their educational services (local school districts, etc.), and measures of population differences across cities, towns and school districts within states that can serve as a guide in the distribution of resources while avoiding the wrong incentives.

I have written about this topic in the attached research article.

Baker.Ramsey.CBased.JEFSubmit.May28_09

I read this today:

http://www.ncpa.org/sub/dpd/index.php?Article_ID=18464

And was especially intrigued by the first bullet point: “Schools receive more than $10,000 per pupil per year, about one-third more than private schools spend per student.”

Having just completed my study of private school tax returns, this statement seemed a bit out of line and there was absolutely no support for it, not even in their main report: http://www.washingtonpolicy.org/Centers/education/policybrief/06_finne_schoolfunding.pdf

They do argue (but do not validate) in this report that the typical Washington private school spends about $6,000 per pupil.

So, I went back to my data set of private schools. Note that the main finding of my report was that private school spending varies widely and varies especially as a function of the affiliation of the schools. The lowest spending schools in my set of 1500 tax returns were those which are members of the major Christian Associations. My sample included 26 such schools in Washington state, which spent in 2007, on average about $6,656 per pupil. So, even the lowest spending group of private schools in Washington spend more than $6k per kid. The largest group filing their IRS 990 returns in Washington were private independent day schools. These schools spent, on average, $19,283 per kid per year.  Hey, that’s about twice what the posting said was the allocation for public schools. Sadly, only 2 catholic schools reported their IRS 990 in Washington, and those schools spent about $13k per kid per year, but are not necessarily representative of all Catholic schools in Washington.

Just checking my news alerts today. A lot going on:

Massachusetts formula review: http://www.wickedlocal.com/gloucester/news/education/x786214932/Lawmakers-seek-review-of-18-year-old-education-funding-formula

Kansas meeting among potential plaintiffs in response to aid cuts: http://www.dodgeglobe.com/education/x402527694/Schools-for-Fair-Funding-meeting-in-Dodge-City

New lawsuit brewing in Arizona: http://www.kold.com/Global/story.asp?S=11138528

Comments on these at a later point.

I picked this article up on twitter: http://www.courant.com/news/education/hc-education-commissioner-0819.artaug19,0,3631152.story

From these figures below, it looks to me like Connecticut has some other issues to deal with. This is  a perfect example of just how illogical state aid distribution and state school finance formulas can be.