Shared Flashcard Set


institutional arrangements and stratification
institutional arrangements and stratification

Additional Sociology Flashcards




Breen and Goldthorpe 1999

direct effect of origin on test scores, secondary effect on evaluation of meaning of education

suggest a revision to rational choice which acknowledges:

1. cost befinets,

2. perceived abilities and

3. available resources


for working class people have guarantee that they maintain working class position by jobs and education to maximize utility also comes with risk which may lower their social standing




Breen and Goldthrope – risk aversion, want to stay at or above current class. for upperclass, don’t want to risk failing below their SES status and to do this must attend school.  For lower-class, there is a guarantee to get job at their level, don’t want to end up in the lowest class, and the safer strategy is to just get a job, less uncertainity, less of a risk.

Idea is if you go to college and fail the key exam you are in a worse place than if you never go. The perception of risk is more important than the reality



Breen and Goldthrope develop an economic model considering risk aversion – not wanting to end up in a position lower than one’s parents, resource cost, and belief about one’s ability to succeed.  They argue that for working class families, the decision to leave school makes rational sense given their assumptions and available resources.


Breen and Johnson 2000

critique the binomial logit model of educational transitions popularized by Mare (1980, 1981)
and recommend a multinomial model of educational careers. The authors use longitudinal data
from Swedish statistical registries to empirically test the strengths and weaknesses of their
proposed “MT Model” in comparison to the binomial model. Results indicate that the specific
pathways taken through the education system influence the probability of making subsequent
education transitions. Furthermore, these findings are robust to unmeasured heterogeneity. The
authors recommend use of the multinomial transition model in future stratification studies.


technical recommendation is straightforward: analysts should replace the use of a binary logit model with a multinomial logit model. This means that instead of modeling two outcomes (e.g., school continuation or drop out), analysts should model three or more outcomes (e.g., academic track, vocational track, or drop out). This relatively simple technical change has important implications for the conceptualization of educational careers and the
precision with which the relationship between social background and educational attainment is

Breen and Jonsson argue that individuals do not necessarily move through the educational system in a unilinear sequential mode as assumed by the logit model. Many European systems, for example, have qualitatively different parallel branches of education (e.g., academic and vocational) that may or may not converge at a later stage. Furthermore, social background effects may be stronger/weaker in shaping transition rates to specific branches. For example, “classorigin differences between those who choose the vocational path and those leave school will generally be rather less than class-origin differences between those who follow the academic
path and those who leave school” (pg. 759). In the binomial model, the academic and vocation
paths are often collapsed which means that the social-origin coefficient is a weighted average of
the true* coefficients.


Alternatively, one track could be thrown out of the binomial analysis, but this ignores potentially valuable information.


Breen and Jonsson are also interested in understanding how educational careers shape potential career outcomes in two ways that the Mare Model is unable to estimate.

1. They want to examine path dependency or the extent to which prior educational choices influence subsequent educational transitions.


2. investigate the “possibility of variations in the
effects of class-origins in determining the probability of choosing among a set of educational options” (pg. 755).


The Swedish registry provides a high-quality dataset with measures of parents’ occupation, social class, gender, educational transitions, and grade point average for over 350,000 individuals born between 1954 and 1967. Sweden’s educational system includes three transitions between approximately age 16 and the start of tertiary education. At each transition point, students are able to select an academic path, vocational path, or leave school and the system is relatively “open” meaning there is technically the possibility of switching paths at each transition.


Finally, Sweden has implemented public policies with the goal of reducing educational inequality by
equalizing resources rather than through educational expansion, which makes it an interesting case study.


The authors analyze the Swedish data using the Mare Model and the MT Model and compare the
results. In these data, the Mare Model tends to deflate class-origin effects at the lower transition
points and inflate them at the transition to higher education. Note that this is not a feature of the
models and these results will vary depending on the data analyzed.


The authors also use average log odds ratios to determine the extent of path dependency. In the
Swedish case, they find that there is a “marked persistence effect” at the second transition point
such that the rate of switching paths is negligible. At the third transition, those who followed an academic-academic path or a vocational-academic path have the highest odds of continuing their schooling. Given this evidence of path dependency, Breen and Jonsson investigate its relationship to class-origin using the MT model. They find that less common educational pathways are characterized by higher class-origin inequalities in comparison to more common educational pathways.


Particularly for men, class-origin effects vary according to path.

Like Cameron ad Heckman (1998), Breen and Jonsson are also concerned about the potential
problem of unmeasured heterogeneity.


Unfortunately, the MT Model is susceptible the same
critique, although to a lesser extent.


1 Closely related is the problem of selection bias, which is actually more problematic in the MT Model than the Mare Model because of non-independence
of transitions. “An apparent path dependence may be spurious because it is induced by the common effect of the unmeasured variable on earlier choices (the paths) and current choices.”

To address these selection bias concerns, Breen and Jonsson use Vermun’s latent class approach and “focus on the robustness of the effects of path dependence at the transition to tertiary education” (pg. 768). I am not familiar with this procedure, but the authors state that the findings are robust and did not change any prior conclusions.

Breen and Jonsson conclude, “the MT model improves our understanding of underlying processes of educational inequality and provides better guidance for which political measures are appropriate to reduce such inequality” (pg. 771). I agree that the MT model allows for more complex and nuanced studies of educational inequality. In this regard, it should be a ‘tool’ in every analyst’s ‘toolbox.’ However, I am not convinced that is appropriate for all or even many studies of intergenerational process related to educational attainment. Breen and Jonsson note that their criticisms of the Mare Model do not apply to educational systems that permit only binary choices or when these choices are unimportant (pg. 755).


However, they also state “such differentiation is present in (virtually) all educational systems, including that of the United States” (pg. 759). While the MT model seems very appropriate in the Swedish context (and several others), it is less clear how it would apply to the U.S. context. According to research on tracking and school organization more broadly, the U.S. educational system is more
differentiated within schools than across schools.


Even if alternative stratifying measures were
available (e.g., prestige is recommended by the authors), it still seems implausible at the k-12
level in the U.S. It is possible that this model could be used to study horizontal stratification in
postsecondary education. Furthermore, cross-national studies that prioritize standardization of
measures and models will be faced with difficult decisions regarding how to classify each
nation’s educational system into several categories.


Additionally, the authors gloss over the limitations associated with using a multinomial outcome.
In my personal experience, I have found that researchers are significantly less familiar with a
multinomial outcome, which makes interpretation more difficult. Therefore, papers using
multinomial logit models must be written in a more careful and explanatory manner.


Additionally, some specialty statistical analyses and certain combinations of procedures with a multinomial outcome are not possible in standard statistical software packages. Finally, the MT
model simply asks a lot of the data so a relatively large sample size is necessary to avoid small
cell size issues or problems of ‘perfect prediction.’
Therefore, I believe that the MT Model can serve as an important advancement to the Mare Model under certain conditions. However, I do not believe that it will revolutionize the field in the way that the Mare Model did in the early 1980s. It is important for analysts to understand its potential given certain research questions and datasets, but like any model, it has limitations that
were not very clearly represented in this paper.


Burt 1997

Burt suggests that not only does the density of social networks matter, but structural weaknesses or holes in the network can also present opportunities for social entrepreneurs, i.e. chances for aware individuals to network like-minded people for some greater ambitiion or


networks with structural holes present opportunities for entrepreneurial behavior

Cameron and Heckman 1999

·         Potentially the death of the “Mare model” of educational transitions. 


criticisms to the ET approach (Mare model).

1. they suggest that logistic coecients are not substantively meaningful.

2. they suggest that the Mare model does not capture how individuals make decisions, specifically, the Mare approach suggests that individuals are rationally able to assess their own abilities and make decisions
accordingly based on perceived abilities and costs and benfiets.



3. The declining effect of social background is a product of the functional form assumed.


4. they suggest that selection creates unobserved heterogeneity not accounted for in the logit model.


5. suggest the formation of a model without distributional assumptions.


argues logit models may not be most meaningful way of ascertaining whether family background has an effect on schooling.
problem is that logit models invoke a distributional assumption paramaterization of logit models and unobserved heterogeneity make them imperfectly suited model school transitions instead they use a less strict random effects type model


A main conclusion of this is that educational selectivity caused by omitted variables obscures family background effects.

argues declining relation between family and educational transitions is an artifact of the logit operationalization.



·         Address the consistent finding in sociological literature that the effects of family background and family resources diminish at children progress through school.

·         These findings are based on the Mare model, aka the educational transitions model.

o   Treats educational attainment as a series of transitions, recognizing that the determinants of continuing in school can change in importance at different levels of the school system.

o   Basically, a logistic regression of making a given transition on background variables, conditional on having completed the prior transition.

o   The major advantages of this model over the traditional logistic regression was that it recognized the qualitative differences among various transitions throughout the educational career, and that the coefficients were not confounded with changes in the marginal distribution of educational attainment (since educational attainment increased substantially throughout the 20th century, this was important.)

·         However, the logistic model has serious flaws that bring into question many of the findings of the educational transitions literature.

·         Cameron and Heckman’s main critiques:

o   Logit coefficients are hard to interpret (odd’s ratios)

o   Model is only loosely behaviorally motivated, and assumes myopia on the part of individuals (i.e. that they are focused on the next transition rather than their long-run attainment). 

o    *The pattern of declining coefficients across transitions is an artifact of the functional form.

§  Unobserved ability can lead to bias

§  Over transitions, the distribution of unobserved ability shifts right, causing the bias to increase

§  This biases coefficients (family background effects) toward zero, because those from the most disadvantaged backgrounds are likely to have the highest levels of unobserved ability/motivation at higher transitions.

§  Even if the true effects don’t change, the coefficients will.

·         Cameron and Heckman develop a technique to correct for omitted variable bias (latent class model), and find that the effects of most background variables are actually higher at the transition to college than at lower transitions.

o   *Main conclusion is that educational selectivity caused by omitted variables obscures family background effects.

·         Their proposed alternative is an ordered discrete-choice model, which assumes that individuals observe their endowments at birth and choose the level of schooling that maximizes their net returns.

o   Relies on weaker statistical assumptions

o   Assumes rational forward-thinking behavior.

o   Does not address the qualitative differences among transitions.

·         Find that long-run factors give rise to the relationship between family income and educational attainment. 

o   Controlling for ability reduces family income effect

o   It is long-term factors such as permanent income that affect academic ability that matters for educational attainment, not short term factors like credit constraints. 

grodksy 2001

suggests that the degree of wage inequality actually increases for jobs at the upper end of the occupational spectrum.


In a decomposition, it seemed some but not all of these differences were explained by dierences in human capital. An interesting variable that I
don't think entered their decomposition was wealth.


People on the job market may be more or less likely to bargain for wages


looks at how location in occupation structure affects black and white wage inequalities

recent trends in more equality in occupation and less equality in earnings

individual level variables do not explain all race gap but education does reduce it

majority of race gap accounted for by differences in human capital

look at combined occupation as a mediator


Grosky 2001 suggests that the degree of wage inequality actually increases for jobs at the upper end of the occupational spectrum. In a decomposition, it seemed some but not all of these dierences were explained by dierences in human capital.

Domhoff 2002

Tries to explain the paradox of why corporations have so much control and
power in a democratic society like the United States
Argues upper class and corporate upper class are the same and maintain it
through cooperation and tight bounds
This training begins in childhood and elite education leading to social bonding

Fernandez 2001

tests technological bias a cause for rising racial inequality, demand for lower skilled jobs

looks at retooling of food processing plant, effect on workers truly exogenous

retooling about the people change in literacy requirements


after turnover decline in wages below the median, this is in large part caused by exit from the company


wage gap increased dramatically at this new plant

Gerber 2002

goal in transition from communism to market in Russia was to make a more flexible market

newly privatized jobs, higher job transitions, low upward mobility


workers in higher privatized branches more job transitions low upward mobility

firm and branch effects increase over time

in privatized branches with poor perform more job loss
greater privatization in a region, greater turn over
better regional performance -lower job loss

finds at firm level private property effects modest at best branch privatization does have effects to push reform

DiPrete & Nonnemaker 1997

the average of the inward and outward flows for detailed labor market categories divided by the category size


definition of structural turbulence

find those with least resources most affected
for men clear push, pull effects

Gerber 2010

However, in developing countries and/or those countries which are in the process of institutional transitions, social networks might operate differently.

Gerber's findings of the use of social networks in job acquisition in Russia are consistent with theories that suggest that in periods of greater job scarcity, uncertainty, and market transition, use of social networks becomes more common and important for outcomes.


concludes effects of social networks grew following the market transition


in areas with higher unemployment, greater reliance on networks


emphasizing job scarcity (crisis), uncertainty (transition), and our own explanation


(privatization) for why the labor market has gotten more personal in Russia.

Gerber and Hout 1995

The MMI proposition has 4 parts: (Raftery and Hout)


1. increases in education reect increasing population
demand for education and a social upgrading of education for those of the highest social origins


2. an increase in the supply of education will lead to
an increase in the proportion of individuals acquiring that education across all social origins but the relative disparity in making an educational transition by social origin (often measured in the form of an odds ratio) will remain unchanged


3. an exception to 2 will occur when all individuals of a particular social origin complete the education levelwhen saturation occurs, the odds ratio of transitioning by social origin will decline


4. (I don't think this was mentioned in the original 1993 denition of the MMI but is mentioned in Gerber and Hout's 2005 article) equalizing forces at one level of education may be offset by increasing
stratication at a higher level.


educational stratication in Soviet era Russia continued


gender stratication in education dramatically decreased but stratication by parents and geography continued

blossfeld and shavit characterize rise in education relative to cohort size, decreasing association between social origin and years of schooling while association between years of schooling but a relatively constant association between
level of schooling and social origin.


Meanwhile social origin seems most important
for early schooling but less so for later transitions
under maximally maintained inequality theory
1. growth in secondary and higher ed leads to gradually maintained odds ratios for education by social origin and is due to change in demand
2. other increases in education lead to increases by all social origins but preserve odds ratios of transitioning
3. if upper class education totally saturated, odds ratios of transitioning may go down
4. equalization can be reserved, rise in one education may reduce conditional probabilities of future ed

in russia some experimentation in school system but all was standardized and centralized starting in 1934 through the 1970s

entering vocational school vs. higher secondary school precluded university entrance

Soviet era mass expansion of education which outstripped population growths


government tried to make reforms to counteract continued trends in those with more educated parents and better off backgrounds being more likely to enter higher ed


use approach following after mare that looks at logit predictions of transitions not just accumulation of years

findings that explosion in higher ed non VUZ training, urban origins and parents of means were advantageous to educational completion.


female disadvantage in education erased


increases in education increased MMI but for women same trends actually led to a decline in stratfication


increased university stratification cancelled out egalitarian effects at lower level


"A national survey of educational stratification in Russia reveals substantial inequality of educational attainments throughout the Soviet period. Parents' education, main earner's occupation, and geographical origin contributed to these inequalities. Gender prefer- ences for men were removed, and for some transitions reversed. Although secondary education grew rapidly, higher education failed to keep pace. This disparity led to a university-level enrollment squeeze, and the resulting bottleneck hurt disadvantaged classes more than advantaged ones. In turn the effect of social origins on entering university increased after 1965. The upshot was no net change in the origin-based differences in the likelihood of attaining a VUZ degree across three postwar cohorts"


By reforming only parts of the system at a time (starting at the lowest levels), educational policymakers put pressure on the levels above. The planners could not anticipate that opening access at one level would increase stratification at the next level, but that outcome is now clear.

Granovetter 1973

strength of a tie
1. duration
2. intensity
3. intimacy
4. reciprocity

information can be transferred to more people via weak ties

Grusky 2005

An occupation is a category of "functionally similar jobs"

detailed occupations frequently represent the the subjective aspirations individuals, they are recognized widely in society, they promote their own subcultures and lifestytles, and :indivudal identities and self-defintions are strongly affected  by occupational affiliations., almost to the point of bearing out a Durkheimian 'essentialist' view that such ties provide a master identity.


training by senior officers "introduces further homogeneity in the attitudes, behaviors and worldviews of prospective incuments and the social interations amoung occupation member reinforces occupation-specific attitudes, values, and lifetyles.


disaggregated structuralization highlighs how "the instutionalization of an ocupational classification scheme is so deeply entrench in society that it "trains us to regard between-category disparities as appropriate and legitmate"

Occupational schemes are so widely accpected that we focus almost entriely on disparitie within occupation, which are "closely scruntized and are sometimes take as evidence of discrimination (esp when correlated with race, gender, ethnicisy)



different article same year (Weeden and Grusky) Functional niches in the division of laor that typically become deeply institionalized in the labor market

Hodson 1982

provides a critical view of dual market theory by suggesting that too often these characteristics of dual market theory are conated (periphery firms, bad jobs, part-time work) versus (core firms, good jobs, full-time
work) and that dual market theory often makes these types of assumptions that might be interesting areas of study, not taken for granted. Their conclusions are that dual market theory is valuable in creating a discourse challenging neoclassical economics, but it is ultimately limiting. As they put it (far more eloquently than I could try), ....the assumptions of duality in structure and a
one-to-one parallelism between economic sectors and labor markets impose an overly restrictive conception of economic and labor market segmentation


There are four basic elements in the dual model:


1. organizational structure of capital

2. the organization of labor within capital structures


3. a set of outcomes for workers which result from their participation in the labor market

4. and a social division of labor in terms of racial, ethnic, and gender groups.

two sectors of economy primary/periphery periphery firms, there is free moves in and out of the labor market

core firms, firm specic training barriers to moving between markets

discipline in primary sector is bueracratic in secondary it is harsh

Kalleberg 2000

31% of americans in non-standard employment
Kalleberg finds many bad jobs

uncertainty, employers don't invest fringe beneftis
unionization can't help temps

Kalleberg 2006

temporary work driven by employer needs


differences across countries in extent part-time jobs are good or bad


Kallberg suggests another classication
system for jobs, part-time and full-time.


Part-time jobs can be thought of as bad but this is really country and context specific as it depends on both the characteristics of the job and other systems of social support.



Lin 1999

resources in labor market can be classied into


1. personal
2. social


strength in weak ties might lie in access to positions vertically above ones own

Lucas 2001

Based on United States data. Effectively maintained inequality suggests that advantaged classes will more easily obtain advantaged positions in educational institutions and once in those positions are more readily able to maintain them.

Eectively maintained inequality is less about a numerical inequality but about a qualitative inequality and seems to suggest that advantaged classes are at the top of horizontal stratication.


Lucas proposes an important amendment to the proposition of maximally maintained inequality, what he terms, effectively maintained inequality. Lucas suggests that depending on available educational quantity and quality, parents
of higher social origins will be able to secure the educational advantages available regardless and that parents and maintain these advantages.


An important problem with the MMI theory at least in my view (and it seems Lucas agrees) is that it focuses on the quantity of education but does not
speak to qualitative dierences in the education received. Lucas thus proposes to amend MMI theory with a theory called eectively maintained inequality.Lucas suggests that inequality depends on variation in the quantity and quality of education
available to a society. In circumstances where there is greater variation in the quantity of education, higher social origin will increase the chances of a
child getting that education; in circumstances where there is greater variation in the quality of education, higher social origin will increase the chances of a
child getting the education judged to be better.


cites two contradicting literatures: tracking students vs. looking at educational transitions
declining logit coeffcients have been explained via


1. life course perspective as children age, they can make more of their own choices


2. mmi structural factors contribute to declines or maintenence of inequality


criticism of educational transition literature is that it cannot account for qualitative dierences in school experiences


for tracking to be important
1. tracking must be predictive of outcomes
2. social origin must be associated with starting tracking position


tracking matters as middle class parents may be more active in keeping kids in particular track
and social origins also start children in a particular track

Effectively maintained inequality posits that socioeconomically advantaged actors secure for themselves and their children some degree of advantage wherever advantages are commonly possible. On the one hand, if quantitative dierences
are common, the socioeconomically advantaged will obtain quantitative
advantage; on the other hand, if qualitative differences are common they will be observed.

This proposition comes out of the tracking literature
solves issue put by cameron and heckman by including time varying covariates solve heterogeneity issue
Lucas argue that students make decisions both sequentially and myopically opts for an approach to include time varying and time invariant covariate 


I am skeptical
life course perspective decline of parental effects, mmi shocks to the system not always possible to adjudicate



The efforts of Lucas are to introduce the ordered
logit model. Lucas suggests the ordered logit allows for transitions following Cameron and Heckman's advice of use of an ordered logit model when the
error term is homoskedastic.  Lucas uses an ordered logit, because he's interested in a
tracking model


Lucas proposes a more complex model that accounts for an exhaustive list of covariates to measure both the likelihood of transitioning to a
higher level of education and for remaining in a particular educational track or trajectory



Mare 1981

developed new system for measuring education in terms of transitions not just in terms of years of schooling


educational transitions (ET) approach in which the
steps of moving between grade levels or types of education are modeled directly.


shows formally that even when schooling is increasing differences in educational attainment by class over cohorts may actually be invariant


a. Explain the shortcomings of the years of education approach, as Mare described them. Then describe the alternative ET approach introduced by Mare and explain the reasons that he gave for preferring it.

The original years of education used a linear regression model. I'll highlight two pitfalls of this approach: Using a linear model assumes that the effects of social background are the same across transitions be it from year to year level of education or from types of educational institution. Conceptually, there is no apriori reason to believe this is the case. A key finding of the Mare model and a number of studies which adopted this approach is that the effect of social background declined with time. Another limitation of the linear model is that it does not allow one to study trends in educational progression nor does it speak to the probabilities of moving across educational transitions.



argues that trends in education should
be measured by educational transitions not years. Mare advocates for predicting
conditional probabilities of making an educational transitions by using a
series of logistic models. This practice seems to have been an institutionalized
practice in the literature through the 1990s. However, using a logistic model
requires an assumption about the distribution of education for a given population
and an assumption that there is a hierarchy to educational transitions
such that individuals can only move between two categories.

Mateo 2009
unrealistic that firm specic human capital really matters
but they evaluate case where all capital required is general
Morgan and Sorenson 1999

find for public schools (though not Catholic schools) parental social closure associated with negative effects on hs graduation,


Coleman expected such effects to be positive
school with closed social functioning referred to by authors as a norm enforcing school, catholic schools effective example of this, downside is schools can
be oppressive argues social closure doesn't explain Catholic school performance


While the findings have always supported the potential effect of intergenerational closure in Catholic schools, their analyses did not lead them to conclude that intergenerational closure has the same effect in public school; in fact, closure might have its “cost”, and would even have negative association with students’ academic achievement (Morgan and Sorensen 1999a). They came to the conclusion that the reason why intergenerational closure might have positive effect on students’ achievement is that the parental network is coupled with the religious norm that could potentially encourage learning.


Morgan and Todd 2009



Better methods and different data find there is not a positive association with social networks and success in public schools

In Morgan and Todd (2009), they claimed to add the following “new” things to the literature. First of all, they used the data from Education Longitudinal Study in 2002 and 2004 (ELS), as opposed to the data from National Education Longitudinal Study of 1988 (NELS 88) that the previous studies used. The ELS data is not only more recent but they also include network measures that can better capture Coleman’s idea. That is to say, while NELS 88 only include the quantity of students’ and their parents’ networks, ELS provide information on the characteristics of those networks (e.g. the sex and the academic achievement of the students’ friends). They employed those measures in response to previous critiques (Hallinan and Kubitschek 1999). Secondly, to engage the discussion on the “cost” of intergenerational closure, they ran separate models for students from Catholic schools and from public schools and in order to see the difference in the effects on the outcomes. This strategy is similar to what Morgan and Sorensen (1999a) did as they included interaction terms of “Catholic schools” with other independent variables in the models. Finally, though Morgan and Todd (2009) used “parents know parents” as the measure for intergenerational closure, while the previous studies used the number of students’ friends the students’ parents know for the measure. Morgan and Todd’s (2009) measure is potentially more in line with Coleman’s (1988) idea of intergenerational closure being a network among parents.

Another important advancement in Morgan and Todd (2009) is their model specification. Previous studies have used a “causal” language while it is only association/correlation they were referring to (and hence I hereby use them interchangeably to certain degree). Morgan and Sorensen (1999a) only measure intergenerational closure on the school level by using the school-specific means. They argued that closure on the school level is more in line with Coleman’s (1988) concept of closure as the measure on the school level emphasizes its contextual nature in its effect. In order to draw a more holistic map of the causal relationships Morgan and Todd (2009) attempted to, they allow the slope to be random and observe the effect of closure across levels. It is by doing so that they found the difference in the role of intergenerational closure in Catholic schools and in public schools. While including network characteristics and family background measures do reduce the magnitude of parents know parents in Catholic schools, the coefficient is still substantial in the full model with all of the controls. On the contrary, the coefficients of parents know parents are greatly reduced with all other controls for students in public schools[1].

Morgan and Todd (2009) then concluded that while the correlation between intergenerational closure and academic achievement in Catholic schools is consistent with Morgan and Sorensen (1999a), they did not find the negative association between them in public schools as Morgan and Sorensen (1999a) presented. They speculated that the difference is cause by 1) model specification, 2) changes in the effectiveness of some types of public schools and 3) revisions to the data collection instrument.

I mentioned the previous studies because they could account for some of the analytical decisions Morgan and Todd (2009) made in their ten-years-after study. In the ten-year debate, the critiques the two campaigns have on each other, along with the improvements the authors try to make, mostly center on the technical issues, including model choice and specification. Their purpose of doing so is to better operationalize the concepts Coleman (1988) presented.

[1] Morgan and Todd (2009) pointed out that the p-value significance might not provide much information because it is “entirely a consequence on the imbalance in the number of students across sectors—1,918 Catholic school students and 12,025 public school students--which yields an estimated standard error that is nearly three times larger for Catholic school students than for public school students.” They therefore depend on the substantive size of estimated coefficients for interpretation.

Mouw 2003

Mouw however finds that in the United States the causal effect of networks don't really play out rather who you are determines who you know.


does social capital affect labor market outcomes?

suggest much of effect of social capital is similar people becoming friends not causal effect

Mueller 1998

Education is a key institution


called education a crucial institution relevant to stratification.
Mueller important differences across countries in association of education and occupational attainment


differences in vocationally oriented countries workers can move across firms easily without much loss of human capital. US not one of these countries

neo-institutionalists expect that the association between education and occupational standing will converge as types of educational systems convege

industrialization hypotheses suggest that as social development converges

associations between education and occupational standing will converge


two types of education: qualification (high vocational training) and organizational which is less vocational


also differences in standardization within a country
across types of countries, slope differs based on dierent role of secondary education on occupational entry

unemployment is often lower with education but non-linear effects in many countries

standardization matters in terms of giving information

Piore 1970

role of employment on poverty can best be  understood by considering a dual labor market

primary sector offers good jobs with high wages

secondary sector bad jobs long hours low wages


suggests punctuality and regularity important skills to succeeding in primary labor market

suggests we should de-emphasize policies moving people from secondary to primary employment

Sakomoto 1991

finds evidence of institution specific capital


Sakomato's work seems to suggest evidence that there is an embededness to certain types of capital, and basically issues a challenge to neoclassical economists to demonstrate otherwise.

Sorensen and Kallberg 1981

qualitative differences across jobs deviate from predictions of neo-classical economics

employment relationships are social relationships


created in relation to production of goods and services


two types of autonomy in relation to jobs: control in doing the job and control in accessing job

goal is to identify matching process that leads to satisfaction of neo-classical labor market assumptions

employee access to job changes productivity for employers


only when wage employment is completely open can neoclassical assumptions be met

closed systems can be efficient but dif mechanisms
vacancy of jobs there is concern about characteristics of person since they are not easily replaced

majority of job changes now result of vacancy not wages

Gangl 2004

curious if welfare state reduces longterm detrimental consequences of unemployment

employment protection associated with higher unemployment but lower disadvantages associated from unemployment

compares US and west german case

Gangl 2004

low inequalities tend to have persistence of low inequality


looks at longitudinal data, to examine permanent income


there is both more income inequality and income mobility in the US than Europe

real income growth similar in US comparable to other places


20 to 25 percent of income inequality is by chance fluctuations and this variance is especially large in US

US exceptional in people at bottom of income distribution experiencing loss relative to other earners

at bottom of income distribution income is least stable
taking longitudinal view doesn't change view of inequality


genuine income instability is highest in the US


US unique in level of instability and negative income on bottom decile and positive income change on top

Supporting users have an ad free experience!