Public Education
File: c:/ddc/Angel/BestIntentions/PublicEducation.html
Date: Mon Aug 13 21:11:08 2007
Wed May 21 19:10:34 2008
Sat Jan 09 20:03:07 2010
(C) OntoOO/ Dennis de Champeaux
Public education has deep roots, which we are not going to pursue here. Racial and gender discrimination has been rampant, topics we are also ignoring. Whether or not religion can be part of a publicly funded education is yet another nasty topic we will avoid. Who pays for public education: local property taxes, local taxes, state taxes, etc. is also a topic we will bypass, but we must observe that - obviously but with profound consequences - there are no producer-consumer financial transactions, as for example with private universities.
There are two different motivations for public education:
- Society's need to obtain as cheaply as possible qualified
individuals that can take the place of those who die
- Parent's desire to develop the potential of their children
The legislation introduced to establish public education used the former motivation, while the public believes - by and large - that public education is there to develop their children. The disconnect between the two is clearly intrinsic. Since the consumers, by definition, do not pay for public education there are no market forces to obtain the best value for the society's money.
A recent brochure to assist voters listed about ten parties. They all discussed public education and they all emphasized the need to improve it. However, they all are vague about how and about the required funding, although they may suggest that more is to be "invested" in education - which sounds right for the electorate.
The public has by and large not been involved in deciding what is to
be produced by the educational system. Rapid changes in the society are
likely the cause for the non-involvement. In particular:
- Science expanded rapidly in the 20th century, which was trickling
down in the high school curricula and from there further down
- An always-increasing segment of the population had to go to college;
from 2% in 1900, to 60% in 2000. This required dumbing down the
quality to accommodate the masses
Still the net effect was that children had to learn things that the parents were not familiar with and thus the parents were happy to delegate the curriculum decisions to the school boards. These in turn delegated the curriculum to the teaching staff who had to hurry up adjusting to the new and newer school books.
The battles in the US about teaching evolution and/or creationism - twice in a century - is something we better ignore. Still we mention in passing that a recent Newsweek issue reported that 25% (in the US) does not know that the Earth circles the Sun once a year.
The quality of public education is worrisome according to [Deeptha Thattai]:
A federal report published in 1983 indicated very low academic
achievement in public schools.
While public education is not a right guaranteed by the US
Constitution and is considered to be the responsibility of the states,
the situation had not improved so that the Bush administration jumped
in with the "No child left behind" education plan in 2001. Schools
must yearly test the proficiency of their students using standardized
test. Soon after "failing school" entered our vocabulary - after
hundred year of running public schools.
What is going on?
A special case is identical twins. They have identical DNA and their birth order is immaterial. Some of them were separated at birth and raised in different circumstances. They are the perfect case to settle the raging battle whether a person is ultimately shaped by the genes or by nurture and education. It turns out that they are very much alike, including IQ, when they are tested many decades later.
In short, a person's IQ has a positive correlation (around 0.6 on a scale of [-1, +1]) with the IQ of the parents. Hence while there are exceptions (smart parents having dumb children and the other way around) as a rule of thumb we have that smart parents have smart children and the other way around.
Thus we know that:
- There is a wide range (a bell-curve) of cognitive skills, among
which IQ
- A person's IQ potential is at a substantial rate statistically
inherited
- A person's IQ at any age is at a substantial rate statistically
determined by that person's IQ potential and the amount and quality of
the education received
There is no exclusive winner in the nature-nurture battle:
- IQ potential must be developed through education to be fully
developed
- Education will not allow a person to reach for an unbounded IQ
These findings have been known since around 1970 in academic circles but the society still has not yet fully accepted the consequences.
The "dictatorship" of a public entity specifying what was to be taught created a revolt based on the insight that children are each unique and thus need room for individual development. Standardized testing was certainly a no-no because it did not recognize the preciousness of each individual. This idealistic stance did not increase the average quality of what the educational machine produced.
The interests of the society swung the pendulum in the other direction when international comparisons described a nation at risk. Finally, after heavy protests from the educational establishment, the nation decided to introduce standardized testing. A fine argument is that without measurements one cannot determine whether interventions are actually beneficial or not.
However there is a "little" problem when a school is declared to be
failing when, for example, the average results of testing children in
fifth grade is below a certain level. The test results of the
children are determined (by and large) by:
(1) the amount and quality of the education they received
(2) the IQ potential of the children
If inheritance of IQ potential did not exist - which was believed for a long time in the 20th century - then low grades implying bad teachers would make sense.
If it would be true that the IQ bell-curves of the in-flowing children would be identical for all schools, then it still would make sense to conclude that bad teachers cause low grades.
Since we all know that schools recruit children from different areas, which can be "good" and from those that are "not so good" (as reflected by property prices), it is a fallacy to assume that the IQ bell-curves of the inflowing children are the same.
Thus judging a school on average test results without taking into account the IQ bell-curve of the inflowing students is a serious mistake.
To test the quality of the teaching staff we need to determine its productivity by measuring for a class the cognitive skills at the beginning and at the end of a school year. A really failing school is one where there is not enough difference between the scores at the beginning and the end of a school year. [To be fair: certain districts have truly dysfunctional children and keeping them alive is already quite an achievement. For the sake of the argument we ignore in the remainder these luckily, (still?) exceptional cases.]
The key observation remains that during the preceding century we
have not bothered measuring the productivity of teachers, schools,
districts, the nation's public schools. At the same time, we allocate
a sizable amount of the nation's GDP to public education (5% of GDP in
the Netherlands). There are two problems that come to mind as a
consequence of not tracking in detail the cognitive skills of
children:
- How can we improve education if we do not measure systematically what
is going?
- What is the long-term trend of the cognitive skills in the society?
The first problem is tricky, because the education machine has always been able to defend itself that the society never provided clear, unambiguous, required deliverables. The decentralization of public education has been a key "advantage" that precluded the development of criteria to held the education machine accountable to. Only fairly recently awareness is growing that something has to change.
The difficulty of answering the second question is where - in our opinion - a calamity of major proportion has been hiding. We turn to this topic next.
We don't know the real answer, because we have never asked the education system to perform these tests every year. We propose that the distribution has likely shifted to the left; i.e. on average we have not become smarter, instead we have become duller -- a pleasant word used in the Bell-Curve [Herrnstein&Murray].
Let's set the stage first:
- We consider the 20th century
- We consider a population region for which we can ignore in- and
out-migrations
- A person's intelligence is a complex function over time, thus
consider, say, only 10 year olds
- Ten year olds have experienced in the 20th century varying
educational regimes, varying family sizes, varying economic
circumstances, varying nutritional philosophies, varying disciplinary
philosophies, varying amounts of basic knowledge to be learned,
varying exposure to radio, television, movie, etc. media, etc.
Hence it is no surprise that even with our severe age and region restrictions, we still have as input a population cohort that has gone through widely varying experiences. Longitudinal studies about IQ distributions on such a segment are spotty at best. Consequently, any finding in any direction can be readily attacked easily.
The Flynn effect suggests that we have become smarter in the 20th
century [Flynn]:
- "James R. Flynn discovered that IQ scores increased from one
generation to the next for all of the countries for which data
existed."
For example [Flynn effect]:
- "Dutch conscripts gained 21 points in only 30 years, or 7 points per
decade, between 1952 and 1982"
- "Two large samples of Spanish children were assessed with a 30-year
gap. Comparison of the IQ distributions indicated that ... the mean IQ
had increased by 9.7 points"
The experts have discounted these results based on the different
exposures that the test subjects had experienced (as outlined above).
The effects measured are unrealistic as argued in [Flynn effect]:
- "... however Arthur Jensen warns that extrapolating beyond the data
leads to absurd results such as an IQ of -1000 for Aristotle (even
assuming he would have scored 200 in his day)."
A fair conclusion is that the IQ tests - even after 100 years of refinement - are still not stable enough to be used for longitudinal studies. Thus we appear to be in an impossible situation with our claim that we have become duller.
There is a way out of this impasse. But first let us agree that there
is not an independent definition of intelligence beyond what these
tests measure. Hence after any careful deliberation one can always
use a sledgehammer argument that any conclusion is nonsensical
because we don't know what intelligence is in the first place. For
example one may claim that:
- No IQ test helps to assess the ability to survive in the Amazon rain
forest.
Getting these objections out of the way, we can agree that IQ tests
can be used to measure differences in "intelligence" between different
people in a certain narrow time period of, say, one year length. This
has been used to show whether or not children inherit IQ from their
parents and whether identical twins reared in different contexts have
comparable IQs after 40 years, as discussed above. To reiterate:
- Assuming that a person is educated, a person's IQ is statistically
(with a correlation of 0.6) determined by the IQ of the parents.
Step two in our argument is the observation that statistically people
in the lower 50% of the society (regarding educational level) have had
in the 20th century more children than in the top 50%. For example
[Flynn2]:
- "... new research from the 2006 New Zealand census showed that those
women without a tertiary education had produced 2.57 babies each,
compared to 1.85 babies for those women with a higher education."
Combining these two statistical effects we reach our conclusion that we are dumbing down. However, we want to be the first to acknowledge that this conclusion pertains only to the potential of a person to develop IQ.
Thus our conclusion is compatible with the Flynn effect with the following line of reasoning: there was so much undeveloped potential available in the beginning of the 20th century that IQ was increasing because the society made free education available and increased the funding in the competition against other nations (and possibly due to improved nutrition).
This still begs the question how to validate our "dumbing down"
mechanism in an ocean of increasing IQ scores. While there are no IQ
tests on newborns, here some indirect, supporting observations:
- The Flynn effect has maxed out [Flynn effect]:
The Flynn effect may have ended in some developed nations starting in
the mid 1990s. Teasdale & Owen (2005): "report intelligence test
results from over 500,000 young Danish men, tested between 1959 and
2004, showing that performance peaked in the late 1990s, and has since
declined moderately to pre-1991 levels."
- A Dutch committee admitted that Dutch children are performing worse:
10% of the children were doing poorly in 2000; this increased to 15% in
2007 [Dijsselbloem].
IQ is inversely correlated with the propensity to commit crimes. The
following quotes needs careful interpretation. The numbers cannot be
taken as being absolute representatives of IQ declines. The society
responds slowly to actual phenomena. We do not want to go into
politics/ philosophies here regarding the 2nd half of the 20th century
beyond observing that the idea of "make ability" gave way to "lock them
up".
- "The incarceration rate per 100000 in the US has increased from 601
in 1995 to 738 in 2005" [Wikipedia-Incarceration]
- "After sharp increases in the 1980s and 1990s, the incarceration
rate has recently grown at a slower pace from 130 offenders per
100000 in 1980 to 500 in 2005" [offenders]
[The different figures in these quotes for 2005 are likely caused by the inclusion/ exclusion of inmates held in local jails.]
The following statistic on the number of people on probation can less
easily be explained away as the consequence of "lock them up":
- The number of people on probation (= court ordered community
supervision of convicted offenders) increased from slightly over one
million in 1980 to over four million in 2005 [probation]
Another set of data regarding who owns the nation's assets and who pays income taxes sheds light on what is currently (2007) in the US the middle class.
Traditionally there was a large middle class with smaller lower and
upper classes. This pattern appears to have been broken in the US
(and possibly elsewhere) around 2000 (and possibly earlier). Consider
the following statistics:
- the bottom 40% does not pay Federal income tax
- the bottom 50% owns less than 6% of the nation's assets (in spite of
a century of progressive taxation)
- the top 5% pays 50% of the Federal income tax
- the top 3% pays 50% of the California income tax
Based on these statistics we have a small upper class, a larger middle class, but which is still a minority, and a lower class that is larger than 50% of the population.
This data suggests that we have already entered (in the US) the period of declining average IQ.
Since in every region in the world there has been a period in the 20th century (or during the whole 20th century) where the least educated procreated more than the most educated, we must conclude that there is reasonable evidence that the potential of human IQ - as determined by genetics - has gone down in all regions of the world.
This trend is ironic when contrasted with a 19th century position
[Deeptha Thattai]:
- "The common-school reformers argued for the case on the belief that
common schooling could create good citizens, unite society and prevent
crime and poverty."
There is no direct causality between public education and the dumbing down of the gene-pool. Plausibly one can argue that public education being free made it a facilitator for having more children. A family recently welcoming their 17e child - running up a tab of at least $1,224,000 for public education alone - is an extreme example. This advantage counts more for the poor and the less educated than for the more affluent and better educated.
If so, we have here a "fine" example of an un-intended, negative side effect of a well-intended policy.
The press talking for decades about "the rich are getting richer and the poor are getting poorer" is correct regarding the consequences of this effect, but they are not willing to report about its cause.
We cannot afford the achievement gap morally, socially or economically," state schools Superintendent Jack O'Connell said. "We know all students can learn no matter what ethnicity, economic status or native language. Now we must confront and change those things that might be in our education system that could be holding some students back.We agree that individuals differ in (potential) cognitive skills. Hence why assuming that clusters have identical bell-curves? Sure, one may try to convince the parents of children in a cluster that academics is really important, or that limiting family size helps to give more attention to children, etc., but that still may not bring equality of a cluster's bell-curve with a national one.
Clinging to the axiom that clusters should have identical bell-curves can lead to desperate arguments [Ventura]:
Linda Murray, an official with The Education Trust - West, agreed that factors other than economics contribute to the achievement gap. The trust is a nonprofit agency based in Oakland that focuses on education reform. 'It's a vicious cycle,' Murray said. 'From our research, we believe oftentimes that schools that serve a large population of Latino and African-American students expect less from their students. These schools give less rigorous assignments, so there is already a lower expectation of achievement. If they are not challenged by the coursework, then that reinforces low standards. Certainly there is something more going on than just poverty that's impacting these students.'Gender discrimination is being stamped out as an ongoing project. The key insight is that even when we agree that a few gender differences are real, the overlap is so large that one should treat a person not as a man or a woman but simply as a person. Herrnstein & Murray [Herrnstein&Murray] argued this stance very forcefully in the context of white-black discrimination. The same story applies to different clusters with different measured bell-curves. Denying these differences is ludicrous. Accepting them makes authoritative arguments against discrimination way more believable.
Public education started around 1900 and thereby contributed to the developed of cognitive skills that previously could be developed only by those having sufficient means. Hence all talents across all classes became available. Notice that this effect had no impact on the nature-dimension, the basic potential of the population.
At around the same time medical progress made in the 19th century, especially the insights by Pasteur, started to kick-in, leading to rapidly expanding populations. As argued above the top 50% procreated less than the bottom 50%. Due to the statistical ability to inherit cognitive dimensions, among which IQ, there has been a dysgenic (moving to the left) impact on the IQ bell-curve.
The first effect worked faster than the second effect. Thus if we would have had perfect IQ tests than we would have obtained for our 25 year olds an IQ curve that went up since 1900, reached a maximum somewhere and is slowly going down since then. Whether we have already past the maximum is an open question, although the increasing incarceration rate, the assets distribution, the income distribution and the Dutch report [Dijsselbloem] suggests we have.
Perceiving the essence of education as teaching these soft goals has been a forceful argument for a century against standardized testing of hard goals. The educational establishment not willing to compromise and not taking the initiative to some kind of self-assessments, performance tracking, process improvement and not willing to negotiate publicly what proper deliverables are, has pushed them now in a bad corner.
Reporting regional, state and national IQ-trends must be established just like the trends that are measured and reported in other segments of the economy.
[Dijsselbloem] http://www.nrc.nl/binnenland/article1094496.ece/Ministerie_erkent_daling_onderwijspeil
[Flynn] Flynn, J. R. (1994), "IQ gains over time". In R. J. Sternberg (Ed.), Encyclopedia of human intelligence (pp. 617-623), New York: Macmillan.
[Flynn2] http://en.wikipedia.org/wiki/James_R._Flynn
[Flynn effect] http://en.wikipedia.org/wiki/Flynn_effect
[Gardner] Gardner, H., "Frames of Mind: The Theory of Multiple Intelligence", ISBN 0-465-02510-2, 1983.
[Herrnstein&Murray] Herrnstein, R. & C. Murray, "Bell Curve: Intelligence and Class Structure in American Life", A Free Press Paperbacks Book, 1994
[offenders] http://www.ojp.usdoj.gov/bjs/glance/incrt.htm
[probation] http://www.ojp.usdoj.gov/bjs/glance/corr2.htm
[Ventura] http://www.venturacountystar.com/news/2007/aug/24/minorities-score-lower-on-state-test/
[Wikipedia-Incarceration] http://en.wikipedia.org/wiki/Incarceration
Here a simplified characterization of IQ-intelligence: it stands for the ability to deal with symbolic information, the ability to survive in New York City, the ability to do research that can culminate in a Nobel prize.
Different clusters having different IQ-bell-curves need not worry us. The people in a cluster had their genes over time optimized for dealing with the challenges of their geographic region, which can be substantially different from symbolic information processing. And more significantly: success in society (or happiness) depends on many other skills than raw IQ-intelligence.