Sunday, March 1, 2015

Charter Schools: The Large Group Orientation

This post is a follow-up from a previous through experiment applying education as an adaptive group theory to charter schools.

Part 1 - Framing the situation
Part 2 - Large group orientation
Part 3 - Small group orientation
Part 4 - Equivalent orientations
Part 5 - Evaluating other approaches


Education-as-an-adaptive group theory specifies two competing orientations within education: a large group orientation and a individualistic (small group) orientation.  Tension between orientations of similar strength produces complex behaviour. The large group orientation is embodied by institutionalized education.  The small group orientation is, in this test, embodied by charter schools or charter school systems.

Operationalization of these orientations is not straightforward.  Fortunately, I've ported ontological assumptions from multi-level selection theory, a productive and increasingly mature study area.  This means that adaptive group emergence is ontologically* based on fitness covariance: a group level only exists when fitness is not adequately explained at a lower (individual) level.  This yields a bottom-up social theory.  Ed-as-an-adaptive group focusses on the small-group-large-group interface.

However, in practice many social constructs are not present long enough to produce valid fitness measures.  This means social researches can:
  1. find comparative groups operating across long time spans, or,
  2. deal with behavioural expressions highly correlated (in an exclusive way) to adaptive group function, or, 
  3. deal with propensities for participation in group classes which are probabilistically correlated to adaptive group function.  
Option 3 is several steps removed from hard fitness measures.  Error propagation is disconcertingly significant.  I tend to apply option 2, limiting my study to adaptive groups with a strong to moderately strong moral mission.  I've found this filters out a lot of noise.  Plus, the religious like dynamics associated with strong moral missions correlate (qualitatively right now) extremely well with the presence of an adaptive group. It also enables study of social structures which, while not stable (i.e. many generations), are moderately so (say on the order of decades).  Institutionalized education in North America fits this time range (~180y**) as do many instantiations of specific charter school practices (~10y+). 

*Note: Technical debates exist on whether group levels are ontologically significant in the greater than equal to case (>), or only in the greater than case (>).  Multi-level selection theory (MLS2) takes the greater equal case(via strong altruism).  Selfish gene theories take the greater than case (via weak altruism).  Although Okasha has specified boundary conditions for this dicothemy, in practice, parsimony's significance is at least the great as computational precision.  This is because measurement fuzziness challenges precision's exclusivity.  While it's easy to get bogged down in technicalities, the main take-away is that fitness differences are the ultimate arbiter of what constitutes an adaptive group.

**Note: 180y or about 5-9 generations is probably pushing the limit for fitness measures.  10y is not enough time to even produce 1 generation.  Therefore hard fitness measures are not possible in this time range.  However group commitment on 10y spans do imply a fair bit of stability toward a specific group rather than just a series of similar classes of ephemeral groups.


Following the outlined methodology, an actor's large-group orientation is characterized by their engagement in group directed actions that have individual fitness costs (probabilistically).  In an adaptive group, on average, large-group provided fitness bumps offset individual commitment costs.  Give-and-take means that within a group some people get taken for a ride (losing fitness), many people come out even or a bit ahead, and some people get a big bump***.  Once too many people get taken for a ride, the group loses its adaptive function (via lowering the mean fitness bump provided by the group, so mean costs exceed mean benefits).  Orientations may remain, but are counter-adaptive.  Hence the evolution of sophisticated norm variance detection heuristics and the emergence of otherwise illogical sacred values (for spotting commitment levels and binding groups via costly commitment displays, ritual, and embodiment of moral big brothers).

Freeloaders within an adaptive group are positioned to reap significant windfalls: superficial compliance costs little and you really only suffer a fitness loss if you get caught.  Those who know how to "play" the system are positioned to gain even more.  They can adjust group operation to benefit themselves or other freeloaders.  However, this comes at greater risk of detection and punishment.  It's also why adaptive groups, like religions and cultures and highly resistant to planned reforms****.

***Note:  Obviously these measures are most meaningful on long time scales which span numerous generations.  Give and take occurs both on short terms and long terms.  Fitness measures are always probabilistic.

****Note:  Why we lie is a decent starter on the freeloader topic.  It posits that self-rationalization of norm variance lowers the risk of detection.  If you can convince yourself that your freeloading is actually helpful (or fairly neutral) to the group, then you're more likely to exhibit group compliant signals and less likely to express a cheater's tell.  The net result is cognitive dissonance surrounding group commitment: most people don't want to consciously abuse the group (due to evolutionary selection - group success is good for creating selective pressure for groupishness) but they also don't want to avoid doing what's in their own self-interest (from game theory, the best specific spot is for you to be a cheater and everyone else to be altruists).


image from

While education-as-an-adaptive-group theory, like its parent multi-level selection theory, posits two competing fitness orientations; large-group (competition between groups) & small group (competition within a large-group), as with multi-level selection theory, nesting among groups is permissible.  Thus theory operationalization suggests many types of orientational tensions are possible (even if many are unlikely).  Here's a sample:
  • large-group orientation vs. small-group orientation
  • small-group orientation vs. individual orientation
  • medium-group orientation vs. small-group orientation 
  • etc.  
Covairance, rather than informed speculation determines whether or not an additional (usually higher) group level is required (or useful) to explain fitness contributions. The complicated social dynamics people engage in mean that many nesting options are ephemeral. The ephemeralness of many nesting options means that associated fitness enhancements are extremely probabilistic (and only start to become meaningful with large n's).

In education, three group levels seem to emerge (although no measurements have yet been taken to verify that these levels are indeed what the fitness co-variance actually produces).  These levels are:
  1. the large group embodied by institutionalized education
  2. various small groups embodied by identifiable sub-populations which have collective adaptive function
  3. an individual person 
Small group orientation introduces the most source of framing error.  How many sub-sizings and sub-nestings are possible for "small groups"?  As with regression analysis, over-fitting is as much a problem as under-fitting.  Without hard data, limiting things to specific charter schools and charter school systems "feels right" (hence current avoidance of rigid small-group definitions).  This means when we deal with charter schools we get a couple of different possible orientational tensions.  
  1. institutional education vs. a charter school
  2. institutional education vs. a charter school system
  3. a charter school system vs. a charter school
  4. charter school vs. individual
  5. charter school system vs. individual
The effect size of each orientational tension obviously requires investigation (fitness based studies in education have a lot of work cut out for them).  From a methodological perspective, narrowing down which orientational tensions are and are not at play is analogous to standard factor analysis studies.  The distinction is that education-as-an-adaptive group theory specifies how these options emerge via first principles rather than from (expertly informed) dogma.


The chosen test study was framed around individual charter schools.  Because of this, analysis from the charter school system perspective is out of scope.  Thus three relevant orientational tensions emerge:
  1. institutional education vs. a charter school
  2. a charter school system vs. a charter school
  3. charter school vs. individual
These orientations seem to match common sense: a charter school can either become like a regular school, pool resources with others, go it alone, or get taken-in by a self-serving leader.  The distinction is that these options emerge from first principles.

Because this post is about charter schools' large group orientations, only the first two tensions will be discussed.  


image from

In terms of charter schools, corollaries from existing research (Atran, Wilson, List & Pettit) suggest a groupish orientation involves:
  1. sacrificing some autonomy, 
  2. rationalizing behaviour and intentions according to the group's "moral big brother", 
  3. normalized behaviour,
  4. costly commitment display, 
  5. rituals,
  6. freeloader detection & punishment.
List & Pettit's work on group agents (2014) shows that groups don't need to micro-manage individual orientations.  Group agent autonomy simply requires acquiescence to the group's collective judgment orientation.  Science of religion literature uses the term "moral big brother" instead or collective judgment.  In layman's terms it means respecting the group's central (moral) mission or purpose.  List & Pettit show that biasing effects from the group to the individual are sufficient to eschew the need for acquiescence to group morals on every issue.

In addition, Haidt's work on ex post facto rationalization (and the cognitive science literature on group influence) illustrate the degree to which rationalization to a group is pretty much inevitable.  Further work is needed to tease out the degree of group rationalization which occurs within adaptive groups compared to non-adaptive groups.  Obviously the level of rationalization is proportional to the degree of an individual's commitment to that group.  As an aside, identity fusion strikes me as a productive approach to this sub-topic.


Institutional Education Dominance
Education-as-an-adaptive group suggests charter school and average actors within them will do a number of unique things when oriented toward a large-group (embodied by institutionalized i.e. public education).

  • The most obvious thing is the rationalization of moral judgment.  Moral rationalization is based on institutionalized education's mores.  The charter school accedes to institutionalized  education's moral judgment. As per List & Pettit, education's moral judgement is based on a rationalized aggregate of the laity's morals. The charter school is free to pursue individual judgements that may or may not conflict with institutionalized education's morality.  When judgements in these areas differ, again following List and Pettit's work, a process of rationalization should occur.  Cognitive dissonance theory may be informative in these scenarios.  Ex post facto judgment rationalization is highly likely.  This means actors' emotional "tails" will wag their rational "dogs".  In effect, there will be an inherit emotional attachment to education's social equity tendencies (or whatever education's moral mission turns out to be).  This attachment will supercede rational attempts to re-order zero-sum resources for the charter school's target population.  In other words, if high-needs students on the fringe of the charter school's target demography are present, instructional practices will be moderated to serve these fringe needs at the expense of target demographic specialization and benefit.  These acts will probably be (ex post facto) rationalized as ethical practice, mandated expectation, reputationally important, of minor cost but high return, etc.
    • Falsifiability test - use Haidt's disgust research tests to look for the orientational direction of ex post facto rationalization.  Compare this with behaviour which favours the general population or the charter school's target population.
    From the 2013 CREDO study of Charter School effects

  • Instructional practice will tend to be normalized.  This doesn't mean innovation won't happen, only that it will be biased to consider general demographics rather than the specific demographics possible within a charter school.  The standard deviation of innovative practice should increase (as individual actors distribute across orientational continuum).  Actors and the school will feel "pressure" to engage in practice which maintains an entry point for the general population.  i.e. a student transferring into the school may need to get-up-to speed on innovative practice, but practice isn't so innovative that its only reachable to a specific demography, it's history and it's learning style. 
    • Falsifiability test - compare direction & severity of rationalization to degree of normalized practice.
    • Falsifiability test - compare direction & severity of rationalization to the CREDO graph above.  Large group oriented charter schools should have low variance in"catch-up" time and a small slope (they're the same as the public system!).  Small-group oriented charter schools which should have high variance in "catch-up time" and a steep slope (learning dip adapting to new norms for some & learning spike from great fit to new norms for others).  
    • Falsifiability test - Students deviating from the charter school's target demographic should suffer more in a small-group oriented charter school than in a large-group oriented charter school or large-group oriented public school.

  • Actors within the school and the school itself will continue to do things that are costly for their target population.  This may be maintaining a Carnegie unit, resourcing special needs students, maintaining education's bureaucratic structures, maintaining extra-curricular activities that don't serve the target population, etc.  In education I've found distinctions between ritual and costly commitment displays are hard to tease out.
    • Falsifiability test - research costly commitment displays in education and see how levels match up with school orientation (as measured by the first falsifiability test)

  • Freeloading is directed to the small-group.  Actors are committed to a large group orientation but still play lip service to the small-group orientation.  In a charter school this would be embodied by teachers committed to larger social equity concerns but framing this behaviour as if it was in the best interest of the school's filtered target population.  It could involve connecting with lower performing schools for optics rather than for student benefits.  School branding PR would be very important, because in reality, the charter school is minimally different from any branded public school attenuating to institutionalized education. 
    • Falsifiability test - actors state orientation to the large-group but measure small-group orientation (via Haidt's disgust styled tests) and commit small-group orientation acts.

In general a large-group orientation for a charter school is characterized by putting the needs of (public) institutionalized education before the needs of the school's target demographic.  Charter schools in this orientation may certainly have unique and innovative practices, however, these practices are framed as being for the greater good of all students.  This framing happens by rationalizing that public education benefits from a certain degree of niche optimization.  When push comes to shove, the needs of the general population or outsiders entering the charter school are given priority over exclusionary practice optimized for the school's target population.  Academic effects would be due to population self-selection which could be mimicked in the public school setting by gerrymandering boundaries or facilitating socio-economic streaming via scheduling options.


Most of the previous discussion ports over to charter schools which orient toward a charter-school-system.  

  • Autonomy should be sacrificed toward the charter school system.  Because charter school systems are smaller than public education, an individual charter school should have more influence on the large group's judgment.

  • A system of charter schools need not be an adaptive group; it could simply operate as quid-pro-quo collaborators (the theory specifies that institutionalized education is an adaptive group). Should it function adaptively, the charter school system should have an identifiable morality.

  • Costly commitment displays should be directed to the charter school system.  Rituals should bind the charter schools together (i.e. something like cricket tournaments in an North American charter school system - talk about a painful ritual!  School uniforms could also meet this standard.)

  • Freeloader detection and punishment is likely to be more important and harsher in a small system.  Group adaptiveness is more severely threatened by a single freeloader in group of 10 than in a group of 100.  Small adaptive charter school systems also haven't had the time to stabilize like institutionalized education.  Nor are they likely to have developed the same universalizing mission less they compete directly with the established beast.


So what would a large-group orientation look like in a charter school?  Chances are you'd mainly see superficial differences with mainstream practice.  You'd probably also see some schools starting to break with public education's mores and traditions.  Practices will start to diverge.  Divergence will accelerate as major norms, like bums-in-seats, percentage grades, externally selected curriculum, and social normalization via the hidden curriculum, etc. are eschewed.  This further filters the population and may draw in other people poorly served via public education but who fit well with the new practices and population.

Divergence may result in a go-it-alone orientation,  connections with other like-minded schools and systems, or reproduction for greater strength and anti-fragility.  In all these events, charter schools either risk confrontation with institutionalized education or they risk norm variation pressure and/or freeloader punishment.

Education-as-an-adaptive-group theory establishes multiple scenarios for Charter school tensions via first principle reasoning.  While common sense reproduces many lines of thought, the value of a theoretical base shouldn't be underestimated.  Testable hypothesis are produced and central issues are illuminated.  Tension is a key component of this theory.  As such, it highlights searches for cognitive dissonance, orientational vacillation and moral rationalization.

No comments:

Post a Comment