(PDF) “Exploring social barriers in Chile: The influence of Parental Educational Capital and Gender on Salary in People with University Degree”

1 | P a g e

LONDON SCHOOL OF ECONOMICS AND POLITICAL

SCIENCE

DEPARTMENT OF SOCIOLOGY

SO499 - DISSERTATION

“Exploring social barriers in Chile:

The influence of Parental Educational Capital and Gender

on Salary in People with University Degree”

Msc Sociology (Research)

Candidate Number: 61888

Date: 30th August 2012

Word Count: 9,922

2 | P a g e

Content

I. Introduction............................................................................................3

II. Literature Review....................................................................................6

III. Methods................................................................................................15

IV. Results..................................................................................................23

V. Discussion..............................................................................................47

VI. Conclusions ...........................................................................................51

VII. Appendix A.............................................................................................52

VIII. Appendix B...........................................................................................54

IX. References .............................................................................................58

3 | P a g e

I. Introduction

According to Torche (2010a and 2010b) and Nuñez and Tartakowsky (2007, 2011), in

societies with high socioeconomic inequality, such as the Chilean one, benefits and

earnings are distributed unevenly from the beginning of people’s lives, allowing some

to make use of the meritocratic principles, while others cannot. In this context it is not

surprising that both the individuals receiving these benefits or disadvantages, and the

society as a whole, forget that "personal stories unfold in a context that provides

different opportunities to its members and that when they reach adulthood, the

cumulative result of this process appears as essentially personal attribute" (Torche,

2010a, pp.28).

Following that statement I would like to argue, that in the privileged groups such as

professionals with a University degree, we can expect unequal performance in their

professions expressed by the salary that they received as the most notable indicator

which represents that social exclusion and classist system that there is in Chile.

Indeed, these aspects tell us about inequalities or barriers that come primarily from

origin and gender, but we want to explore in this thesis other sociological variables

available in our data set.

To support that, I would like to do an analysis on a special and paradoxically branded

group, considered a more successful group of people in Chile. I am referring to

practitioners who are finish theirs University`s studies and could be considered the

“elite” in knowledge, cultural, and social positional terms. Based on this elite group I

would like to show that independently of academic attainment, theoretically, there are

4 | P a g e

some (many social and cultural) factors, that I would like to consider as barriers, which

do not allow the equity in their professional life. We will discuss that standpoint

referring to the relevant sociological theories.

Two points support the idea that will be carried out as a main dissertation topic. On

the one hand, Torche (2005 and 2010a) and Nuñez and Miranda (2009) focus their

main analysis mixing on poor, vulnerable, middle and upper class people as a whole. In

fact, they have tried to understand social mobility and inherited mobility as entire

social phenomena considering all the social strata. Therefore, these researchers have

not focuses on an important piece of the population that represented around 70% of

professional people whose family of origin came from a low or intermediate

Educational Background. I am referring to people that constitute a new middle class, a

middle class more educated and highly qualified than others in the past decades

(Bazoret,2010; Espinoza,2011). Furthermore, they have not revealed intensively the

impact of gender in that group of people that conjoin perversely with the family´s

social origin. That fact seems to be associated with a highly structured view about

mobility and inequality in Chile.

On the other hand, research on higher education has tended to stress the role of

educational attainment has had in the last 30 years in the Chilean society. There are

more people in the Higher Educational system and more Universities than ever in the

history of the country. Can this social inequality context and their barriers be the

starting point in undermining the power of the Higher Education in the society?

5 | P a g e

I will do a statistical analysis based on “Principal Occupational Salary” monthly earning

considered as the main response variable and using as explanatory variables the Social

Origin operationalized as Parental Educational Capital and gender. However, during

this thesis I would like to explore some other explanatory variables associated with the

response variable “Principal Occupational Salary” available in the CASEN data sets

analysed, such as changes in job between 1996 to 2001, the city where someone was

born, how they got a job, ethnicity and age. As will be shown these variables are

reflexive of some theoretical issues that will be developed through process of

conceptualization and operationalization in this thesis. That last has been doing in the

times and conditions that a short thesis allows.

I would like to use the Caracterización Socioeconomica (CASEN) Survey´s data sets

from the Ministry of Social Development of Chile. In order to respond to my statement

I will use three similar but different surveys. First, the Panel Casen data set (wave 1 and

2), and the cross sectional surveys CASEN 2006.

Research Questions

The research questions are as follows:

Do people with university degrees whose parental educational backgrounds are low

and intermediate obtain a similar monthly salary than peers from high parental

educational backgrounds in Chile?

1) Are these gaps in salary evident in males and females in the cases analysed?

2) Which other sociological variables could be associated with that gap in the available

data sets?

6 | P a g e

II. Literature Review

The Chilean Higher Educational System

During the last forty years Chile has reinvented itself as a country in many regards. This has for

instance been highlighted by Undurraga (2011), who has pointed out major transformations in

the political, economic, social and cultural spheres, which have triggered profound changes in

the direction and destination of the country.

One of these areas has been the educational system and specially the Higher Educational

System. According to Urzúa (2012) and Torche (2005), as in rest of the world, the interest in

getting a University degree in the last 30 years has expanded at rates that at the beginning of

the 90’s were not expected. On the one hand, in Chile this growth could be explained by

changes in the Higher Education System , which resulted from a fast increase in the enrolment

of young people at private university that started in the 80’s.

On the other hand, that system has been supported by different political centre–left-wing

coalitions (Concertación 1990-2010) and it has also been continually supported by the new

right-wing coalition government (Alianza 2010-2014) both based on the strong belief that a

system like this could open more significant opportunities and generate social mobility. For

that reason, it seems to be that all the governments have been introducing certain reforms

that focused on access more than on the quality of the system and changing the free-market

fundamentalism of the entire educational system. Therefore, the idea of better education is a

strong cultural belief.

7 | P a g e

Graph 1

110,133

287,670

634,733

-

100,000

200,000

300,000

400,000

500,000

600,000

700,000

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Evolution of people enrolled at Tertiary Education in Chile by

Sort of Institution

(1983-2011)

Technical Training Centres

Professional Learning Institutes

University

Source : Information Service for Higher Education (2012). MInistry of Education Chile.

From a very positive position, supporting this increasing educational openness, some

researchers have encouraged the idea that this process has been incarnated by "Subject type",

young people who have parents who were unable to access tertiary education achieve this

(Castillo,2010). They have been called the "first generation in higher education". According to

Castillo, "Seven out of ten young people studying a degree come from families in which no

member had had the opportunity to do it” (Armanet, 2005 in Castillo,2010). This reinforces the

idea of a growing trend in our educational system.

However, according to Caballing (2012), during the years 2006 and 2011 thousands of students

filled the streets to demand better public education, more social justice and equal

opportunities. They rejected the free-market fundamentalism in the entire educational system

that has generated segregation, stratification and inequalities. Does the Chilean society have

to more to move to a fairer and more just system? Are there more actual opportunities in the

social and labour scenario for people independent of their social background?

Educational and Income mobility: Social Inequality similarities in Chile and UK

8 | P a g e

As academics responding to the latest questions, Núñez and Tartakowsky (2007) pointed out

that despite the range of educational changes, openness of higher education and more young

people doing a degree career, Chile’s social inequality-levels have remained virtually

unchanged over this period. Based on an analysis of intergenerational income mobility, Núñez

and Miranda (2009) deepen the analysis by Torche (2005) by providing evidence of increased

intergenerational mobility in education (years of schooling) in the last decades. However,

these authors have not found any significant evidence of increases in intergenerational income

mobility. For Núñez and Miranda (2009), this indicates that increases in educational mobility

that have apparently existed in Chile did not translate into intergenerational income mobility

to the same extent, at least for the reported period of cohorts between 1960 and 1983. This

seems to indicate that there are factors which are limiting the transformation of increased

educational mobility to increased income mobility.

According to Núñez and Miranda (2009), these could include differences in quality of primary,

secondary and tertiary education to which different socioeconomic segments of the

population have access to, as well as the role of social networks and the existence of

discrimination in employment based on social origin.

Overall, according to Torche (2010), there has been an increase in socioeconomic inequality at

the secondary and postsecondary in cohorts that experienced the economic crises, such as the

Asiatic Crisis in 1998 and the Mexican Tequilazo in 2000. However, there is a decline in

inequality in elementary and secondary level transitions due to their universalization. In

addition, this author pointed out the reduced advantage of males, and in some cases a

growing advantage of females at the same levels of education.

9 | P a g e

Similar to the analyses of Chile, sociologic and economic research from the UK recognizes the

importance of education in determining the positions in the labour market, including

subsequent rewards; however, it also recognizes that other factors play an important role

(Jackson, Onwuegbuzie and Turner, 2007). In addition, Jackson et al. (2005) state that "the

association between education and class position is found to be declining, this is likely to be

because education is of reduced value to employers in making their personnel decisions in one

or other, or both, of its two main roles: i.e. as serving to certify relevant competencies or to

signal desirable attributes of employees that are not in themselves directly observable."(p.26)

As shown, there is evidence supporting the idea that more education is a necessity but not a

sufficient condition for increased social mobility (Jackson et al,2007; Goldthorpe and Jackson,

2008 and Boliver,2011). This is because there are barriers to a full realization of education as

the catalyst for social mobility, i.e. there are obstacles that prevent social subjects to enter

meritocratic positions (for a discussion of education-based meritocracy see Young,1958;

Sanders,1995, Themelis,2008). It can be shown that in highly industrialized societies,

characterised by a segmented capitalist economy such as the UK, education is increasingly

losing strength as a trigger of mobility and distinction in the labour market.

To sum up, one important point can be noted based on the UK and European experience:

“There is little sign of economic growth, rising levels of consumption or educational expansion

per se having any connection with greater educational equality” (Goldthorpe, 2010, pp.231).

Therefore, from a theoretical point of view, the assumed role of education as a source of

achievement in the meritocratic ideal has been questioned in the last years (Goldthorpe and

Jackson, 2008). The crisis is not only embodied by Chile, but seems to be global.

Applying the concepts and analyses of Torche (2005, 2010a), the meritocratic ideal would lose

its foundation as an effective agent of change. Contrary to meritocratic assumptions, what we

10 | P a g e

witness in Chile is an example of inherited meritocracy, or in the words of Núñez and Miranda

(2009), a reproduction of Chile’s unequal social structure. Therefore, the interpretation of the

social conditions and aspects of the actual life, such as social origin, and being part of a

vulnerable family, are vital to reveal insights into some apparent consequences that the social

system produces in its subjects. These characteristics have had an important impact on

people’s life, and in most of the cases they are conditioning people’s experience, life and

death. Some of the concepts that could be pinpointed here refer to theories about social

reproduction and aspects that leave unchanged the social condition of people, such as gender

and racial-ethnicity issues that could be associated in this scenario.

The Role of Social “Class” Origin: Bourdieu’s Social and cultural Capital

The theory of cultural reproduction (Bourdieu , 1992, 1993; Bourdieu and Passeron, 1990)

explains the connections between social class of origin and social class of destination in terms

of the impact of cultural capital on educational attainment. According to Bourdieu, cultural

capital consists of familiarity with the dominant culture in a society. He argues that the

possession of cultural capital varies with social class, and it is likely that parents’ transferred to

their children their understanding, information (knowledge) and beliefs about the society, the

educational system and their expected result in real life.

“This consists mainly of linguistic and cultural competence and that relationship of familiarity

with culture which can only be produced by family upbringing when it transmits the dominant

culture” (Bourdieu, 1977 in Zimdars, Sullivan & Heath, 2009 pp.650)

At the same time, according to Portes (1998), Bourdieu emphasizes the intangible character of

social and cultural capital relative to other forms. Whereas economic capital is in people’s bank

11 | P a g e

accounts and social and cultural capital is inside their heads, social capital is inherent in the

structure of their relationships and the cultural inheritance of sophistication and distinction.

Ultimately, to possess social and cultural capital, people must be related to others, and it is

those others, not himself, who are the actual source of his or her advantage. It can be noticed

that in the concept of social capital, the motivation of others to make resources available on

concessionary terms is not uniform. At the broadest level, one may distinguish between

cultural-talented versus instrumental motivations to do so, depending on the level of

sophistication and distinction which social influence is added to in that relation.

Therefore, the possession of social and cultural capital varies with social class and thus, it is

very difficult for people from a low parental cultural capital to succeed in the educational

system. Again and from a socioestructuralist standpoint, a conditioning establishment is

revealed. One system, where the symbols and codes (Bernstein, 1996) are reproducing

inequalities in a system that some authors call, sharing the Bourdieusian concept,

`resemblance within a difference' (Bourdieu and Wacquant, 1992, pp. 106, in Naidoo, 2004,

p.459), consists of cognitive and structural mechanisms that mediate socio-political and

economic forces while simultaneously reproducing fundamental principles of social

stratification.

From a different standpoint but connected with this resemblance system, Lucas (2001) shows

that among the numbers of students entering higher education, students from the most

advantaged social backgrounds look for new strategies to try to maintain their advantage.

According to this theory of Effectively Maintained Inequality (EMI), ‘the socioeconomically

advantaged seek out whatever qualitative differences there are at level and use their

advantages to secure quantitatively similar but qualitatively better education’ (Lucas, 2001,

pp.1652). Therefore, as Torche (2005) points out, the “quantitative inequality” may be

12 | P a g e

replaced by “qualitative inequality,” that is, the advantaged classes will be able to obtain

educational credentials that provide them with enhanced opportunities for further attainment.

In Bourdieu’s terms, they are looking for more and more distinction.

However, according to Zimdars et al (2009), Bourdieu’s analyses of cultural reproduction focus

just on social class, and ignores issues related to gender and ethnicity. Nevertheless, it may be

that differences in particular cultural resources and parental educational attainment can be

explained in conjunction with gender and ethnic differentials in educational outcome. In other

words, those aspects could be genuinely connected to each other rather than just the social

class differentials as a sole and main factor. This is an open theoretical assumption that will be

explored through this dissertation.

The Role of Gender: More Discrimination in the social system.

According to Chang (2000), the long-standing belief has been that occupations are the core

aspect of the class stratification system and social reproduction aspects. It means that for

some occupations, despite women entering into the formal economy in ever-increasing

numbers, the occupational structure becomes the main locus of gender stratification as well.

“That is, occupational segregation affects the gender gap in earnings, the likelihood of career

mobility, and the possibility of having some degree of autonomy over one’s work”

(Chang,2000, p.1658). However, those gaps are better shown in salary differences between

males and females.

At that point, from a sociological and theoretical standpoint, salary differences between males

and females caused by discrimination can result from several processes. The first is where

women are differentially allocated to occupations and establishments that pay lower wages.

13 | P a g e

According to Petersen and Morgan (1995) and Castilla (2010), that primary mode can be

conceptualized as "allocative discrimination". The second is where occupations held by women

are paid at a lower salary than those held primarily by men, although skill requirements and

other salary-relevant factors are the same; this is called "valuative discrimination." The third is

where women receive a lower salary than men within a given occupation within a given

establishment, combining the allocative and valuative discrimination that involves the

segregation of men and women into different occupations, establishments, or both, and may

occur without within-job wage discrimination. Thus, it may be the case that where men and

women share the same jobs they receives the same pay but that in most cases they simply do

not share the same jobs.

Some studies carried out by Hogan, R. Perrucci., C. , and Behringer. A. (2005) and Hogan, R.

(2001 and 2011), considering the USA context, have asserted that females, particularly

professional married women, do not face the same disadvantage in family wealth or in

earnings that low-income self-employed women face, but that self-employment, in general,

seems to predict lower earnings for women, while in less developed countries such as Chile the

evidence does not share the same diagnosis. According to Bravo, Sanhueza & Urzúa (2008)

“Women are effectively discriminated against in the labour market, with the largest gender

gaps observed among the less educated groups” (pp.19)

Ethnic assumptions: more and more discrimination

According to Hogan, R. (2001 and 2011), the “competition versus exploitation” distinction

between labour market and class theories is not sufficiently nuanced to capture ethnicity and

gender barriers that are either independent of class or that interact with class and labour

14 | P a g e

market, as in the racial, ethnic and gender barriers to and effects of self-employment,

professional credentials, and managerial and supervisorial authority.

According to Padilla (1985), ethnic identity develops from the cultural and structural

similarities of two or more groups and often in response to common experiences of social

inequality. He describes how such conditions as poverty and racial discrimination have

necessitated the assertion of a broader ethnic consciousness and behaviour, often more

successful in social action than individual, cultural or national associations. Nevertheless,

those issues seem to be no related with the life of some ethnic group in Chile. This country is

less ethnic differentiated in South America; we will like to know if that is real.

Chile: A brief history of its culture and geography

However, to put forward those theoretical concepts it is necessary to contextualize some

aspects of the Chilean culture. Chile, as has been pointed out, has been a kind of “laboratory”

of neoliberal policies (Harvey D., 2005 and 2010) that have been systematically carried out by

different governments from 1990. It is in fact a small country of around 16 million people

where 5% percent is said to belong to ethnic groups. However this could be rapidly increased

by the massive migration of Peruvians, Colombians, and Ecuadorians in the last ten years

(Cano and Soffia 2009). Geographically, it is a long country in physical and spatial terms,

where the political, economical and administrative power is controlled from the capital. For

those reasons, topics such as the region where people live or were born, or whether they live

in the capital, are concepts that can differentiate certain life conditions of the Chilean people.

Finally, as in other small countries, Chile has been so influenced by the economic crisis,

especially in 1998 and 2000. Both years can help to explain differences and contradiction in the

analysis.

15 | P a g e

III. Methods

The purpose of this research was to carry out a quantitative study to analyse if there

are differences in the main occupational salary given as a result of Parental

Educational Capital and Gender to professionals with a higher educational degree in

Chile. At the same time, this research explores other sociological variables that can

additionally explain these differences. The analysis of the information collected was

carried out with a t-test of differences of means to non-paired, paired samples and

multiple linear regression analysis.

The chapter is divided as follows. First, a brief description of data and sources utilised

in the current research. Second, a description of the number of cases and the

definition of the cases selected to carry out this dissertation. Third, the analytical

techniques are explained. Fourth, the main concepts are conceptualized and

operationalized, as well as considering measurement issues, and defining response and

explanatory variables. Finally, the chapter states some pitfalls, as well as, strengths,

weaknesses and limitations of this study.

Data and sources

This research was carried out using available data sets on the proper social realty of

Chile. For that reason, it was decided to use an available secondary data set used by

the government and social research institutions in that country. The main data source

was the CASEN surveys panel data set 1996 to 2001 (waves 1 and 2) and the Cross

16 | P a g e

sectional CASEN 2006 (Ministry of Social Development, 2012; Social Observatory

Alberto Hurtado University, 2007).

I used these three data sets to better understand and infer my statement. On the one

hand, the panel data set allows the consideration of a longitudinal view about

practitioners in a period of time from 1996 to 2001. On the other hand, the cross-

sectional survey of CASEN, as will explain further, allows us to undertake a deeper

analysis to better understand those professionals. Additionally, the CASEN 2006 allows

for a more explanatory model to better understand and included some different or

unused variables in the panel CASEN 1996-2001.

Overall, the data sets allow us to have a general view based on reliable and

probabilistic information collected during a period of time from 1996 to 2006 in Chile.

Sample framework

The group of people that is at the core of this research are professionals having

finished their careers at the moment of response to both surveys. However,

information about Parent's Education and their jobs was just asked to Households

(jefes de hogar) and his/her Couples (Cónyuge o Pareja). Additionally, for analytical

purposes and to have cases with the most similar labour conditions, the cases were

selected considering males and females that at that time had been working at least 30

hours per week in the labour market, and whose ages there were between 26 and 65.

Likewise, for similar purposes, the sample did not consider professional males and

17 | P a g e

females working at the FFAA (Army, Police etc.) and personal or domestic service. On

the other hand, for statistical purposes, it has to take into account to the omission of

extreme cases, called outliers, in the “Main Occupational Salary” variable.

Therefore, not all the people involved , especially, in the CASEN surveys in panel 1996-

2001 and in the Cross sectionals surveys CASEN 2006, passed those requirements on

the key issues to be carried out in this research. As a result the dataset was filtered.

CASEN PANEL

1996-2001

CASEN

2006

N

n

153

3267

Kind of Analytical Techniques

The main analysis that I will carry out is T test and multiple regression analysis. In the

case of the multiple linear regression analysis, the main response variable was “Main

Occupational salary” and the explanatory variables were “Parental Educational

Capital”, “Gender”, “Head or Manager job Father”, “To become a Postgraduate”,

among other things.

18 | P a g e

Five separate regressions models are built to explore how available sociological

variables explain predicted outcomes on the “Main Occupational Salary” and to

establish which are statistically significant and have, therefore, more influence

depending on the model created.

The software STATA 12.0 was utilized as statistical tool, using a mix of descriptive and

inferential analysis.

Variables and Measurement

1

As noted, the main response variable here is “Main Occupational Salary”, and this

variable has a continued measure level.

Response Variable

Brief Definition

Question

Level of

measurement

Year

Collected

“Main

Occupational

Salary”2

Income earned by

employed in their main

occupation, either by

way of dependent

work in the case of

employees or

independent work

concept in the case of

employers or

employers and self-

employee (Ministry of

Social Development,

2012; Social

Observatory Alberto

Hurtado University,

2007)

Income module

question: “Last month

How much your salary in

his/her main

occupation/job was?”

Continues

Variable

1996

2001

2006

1

Summary Statistics see Appendix A.

2

The panel the income was corrected by current money at 2001 to be equal and comparable

measurement with 1996. That was doing by “IPC” an inflationary index from Chile.

19 | P a g e

The explanatory ones referred to consider more sociological aspects than economic

ones.

Regarding our main definition about some sociological variables that can be used from

the data sets available, an important point has to be clarified and explained to

understand better our research analysis.

We are not able to consider embodied aspects of social class and socio-cultural

distinction as some cultural sociologists have pointed out (i.e. “habitus”).

Unfortunately, for concrete reasons and available information, this is outside of the

scope of the current study. Indeed, according to Zimdars et al (2009), measures of

cultural , educational and social capital are limited, and cannot properly capture each

dimension of those concepts. In order to operationalize a concept, the researcher is

forced to define clearly and reliably the concept in question.

Explanatory

Variables

Brief

Definition

Questions

in surveys

Procedure

Categories

created

Year

Live in the

Capital’s

Country

Interviewee’s

current living

place.

District or Commune’

where are you living?

Recoding the variable

into a dummy

0==Living in Other

district

1==Living in Capital

1996

2001

2006

To get a

Postgraduate

degree

Interviewee

receive a

postgraduate

title recently

Which level of

education you got it?

Recoding the variable

into a dummy

0== other else

1==postgraduate

1996

2001

2006

Changes in

Occupation

between 1996

and 2001

Interviewee’s

changes in their

main occupation

from 1996 to

2001

Which is his/her

main occupation

1996?

Which is his/her

main occupation

2001?

Recoding both

variables into a one

dummy

0== Did not changes

1==Did changes

1996

2001

Kind of

Occupation

Interviewee’

main occupation

2006

Which is his/her

main occupation

2006?

Recoding both

categories into

several dummy

variables which base

line was to be Head

and Manager

Head or Manager==0

Self-employed

person==1

Worker or Employee

at Public Sector==1

Worker or Employee

at Private==1

2006

How do

generally got

a Job?:

Other way

Modes that

people generally

get a job in Chile

Independently of

your own effort: Who

helps you to get

his/her Cuurent Job

Recoding the variable

into a dummy

0==Relatives+friends+pr

ivate institutions

1==Other ways to get a

job

2006

Born in a

different

district to

where they

live

Changes in the

original district

where

interviewee born

Are you living in the

same district where

your mother lived?

Recoding the variable

into a dummy

Yes==1

No==0

2006

Ethnicity

Belong to some

original ethnic in

Chile

Do you belong to the

follow Ethnics?

Recoding the variable

into a dummy

Yes==1

No==0

2006

Strengths, weaknesses and pitfalls

As has been pointed out through this chapter, some of the strengths of the current

research have been based on the use of reliable and valid information on the statistical

characteristics of professionals in Chile. Indeed, this strategy to carry out the research

allowed the saving of time and money.

21 | P a g e

However, that point has been a weakness as well. Due to constraints the

conceptualization and theoretical assumption have been based on secondary sources.

In fact, the current data sets sometimes generated a change of research focus and a

lack of information in related topics. In particular, this occurred in the analysis of the

Panel data set 1996-2001, where the level of attrition reached 48% between 1996 and

2001 and some variables were just considered in one wave, but not in another. With

regard to this, it has been convenient for this dissertation to construct and use the

CASMIN (Comparative Study of Social Mobility in Industrial Nations) and the

Goldthorpe, Erikson–Goldthorpe class scheme based on employment status and

occupation of parents ( Erickson and Goldthorpe, 1993; Goldthorpe J, Withllewellyn, C.

and Payne, 1987). However, the level of attrition was so high in consideration of the

parental occupation in the panel data set that it was not possible used it.

On the other hand, it can be commented that based on statistical techniques, all the

data were analyzed applying econometrics principles, such as tests of hesteroscadicity

and graphs of residuals, as well as testing and assessing linearity and graphical tools to

see the distribution of the data available (see appendix C).

Limitations

It is necessary to keep in mind that the use of the Panel Data Set (1996-2001) was just

referential information and just represents a sample with some attrition problems.

However, the information and cases provided by the same one has high quality in

terms of the reliability of the information collected to each case considered. For that

reason, an analysis of the Cross-sectional CASEN was carried out to follow up the

22 | P a g e

consistency and trends with the 1996-2001 sample analysed in this dissertation. The

cross-sectional 2006 data set was a relievable and truthful data set and was very useful

to show the main statistical result in this dissertation.

23 | P a g e

IV. Results

In this chapter, the main differences in occupational salary during the last years (1996 -

2006) will be provided. The analysis will draw on different models, whose main aim is

to reveal which variables could be associated with our explanatory variables “Main

Occupational Salary”. Certainly, our focus will be around some vital variables that

seem to be the core of our research, that is, “Parental Educational Capital” and

“Gender”. Using those variables as initial explanatory variables, different and available

variables will be analysed using descriptive and inference statistical analysis.

As can be noticed, there are some significant differences in “Principal Occupational

Salary”, contrasting mainly with Parental Educational Capital, gender and others sort of

variables. The evidence and previous research reveals that these gaps have remained

stable during the last years with some differentiations depending on social strata,

principally those which are supported by research into the social mobility and

inequality that underly the land of flux and opportunity land that is Chile. However, as

this dissertation claims, these gaps are very strong and deep-rooted in the Chilean

middle and upper middle class. What is more, it already co-exists in the lives of those

who have had educational attainment as well as having received a degree at

University. I will reveal this fact based on information coming from the longitudinal

Panel CASEN 1996 to 2001 and CASEN 2006 data sets.

24 | P a g e

In order to present in a simple way some differences based on means analysis, a t

statistical test (t-test) has been used primarily to draw a general and open view. Then

testing based on multiple linear regressions analysis has been added and more

relevant models will then be shown.

What do panel data show in 1996 and 2001?

Apparently, there was no change in “Main Occupational Salary” if you had a university

certificate between 1996 and 2001. In fact, doing a T test analysis it is possible to show

that there are no differences between “principal occupation’s salaries” (by 153 people)

in 1996 compared to 2001.

Table 1

Comparison of Principal Occupational Salary 1996 and 2001 (n=153)

(t=0.086, DF=152, p<0.932)

The T statistic test to paired samples shows that there is not a significant statistical

difference in the mean of “Principal occupational salary” for both years. That means

that there is not enough statistical evidence to reject the null hypothesis (Ho) = mean

(MOS 96 – MOS 01)=0 at the 0.05 level.

25 | P a g e

However, if we do a deeper analysis crossing that paired mean for both years, it is

possible to show that there is just one significant difference depending on the “Parent

Educational Capital”. Table 2 shows the main significant difference that appears for

people whose Parent Educational Capital is Low.

Table 2

Comparison of Principal Occupational Salary 1996 and 2001 by Low Parent Educational Capital (n=55)

(t=-4.630, DF=54, p<0.001)

The table reveals a significant difference between “Principal Occupational Salary”

between 1996 and 2001. This analysis allows us to reject the null hypothesis assuming

the alternative one that there is a significant difference between both years when the

mean of “Principal Occupational Salary” crosses “Low Parent Educational Capital”. On

the contrary, for people whose father’s educational background is intermediate and

high there are no significant differences. For that reason, it is not possible to reject the

null hypothesis for those groups.

26 | P a g e

Regression analysis based on “Monthly Principal Occupational Salary” panel survey

year 1996

3

Another way to examine that trend is by adding regression models which include

variables that seem to be associated with the logarithm

4

of “Monthly Principal

Occupational Salary” on 1996. Table 3 shows the models that explore the partial

association between the response variable and a set of explanatory variables in 1996.

These explanatory variables have been selected considering the theoretic and

pragmatic criteria that the literature and the dataset review have provided.

Table 3: Multiple linear regression Monthly Principal Occupational Income of Professionals in

1996 on selected explanatory variables

MODEL 1

b

Se

Parental Educational Capital Intermediate

0.466**

(0.166)

Parental Educational Capital High

0.660*

(0.314)

Female

-0.647***

(0.155)

Age 1996

-0.00821

(0.00969)

Live in the Capital’s Country 1996

0.306

(0.156)

To receive a Postgraduate degree 1996

0.413

(0.309)

Intercept

954395.8***

(274682.0)

N

147

R-sq

0.215

adj. R-sq

0.181

Rmse

0.922

Prob > F

0.001

= *p<0.05; **p<0.01; ***p<0.001

Source: Own elaboration based on Casen Panel 1996-2001

3

Please refer to all further confidence intervals in appendix B

4

Here the specialized literature (Tarling, 2009; Bravo and Vasquez, 2008; Agresti and Finlay, 2009)

recommend use of the natural logarithm of the salary because the empirical log of income is closer to a

normal distribution, thus solving statistical and linear problems with applied econometrics techniques.

27 | P a g e

It can be seen that the model is highly significant, as is indicated by the F test

(p<0.001). Additionally, the R2 indicates that close to 22% of the variance of “Principal

Occupational Salary” is explained by the model added.

Additionally, the model indicates a significant partial association between the

dependent Dummy variables “Parental Educational Capital of Family: Intermediate”

(t=2.82; p<0.01) and “High” (t=2.10; 0,05) and the dummy variable Female (t=-

4.16;p<0.001).

The interpretation of the estimated statistical significant coefficients is as follows:

The coefficient of dummy “Parental Educational Capital: Intermediate” is 0.47. This

shows that, controlling for the other explanatory variables, the predicted “Main

Occupational Income” is 47% higher for people whose Parental Educational Capital is

intermediate, compared to those whose Parental Educational Capital is Low. Similarly,

the coefficient of dummy “Parental Educational Capital: High” is 0.660. This reveals

that, controlling for the other explanatory variables, the expected “Monthly Principal

Occupational Income” here is 66% higher for people whose Parental Educational

Capital is High than for those whose Parental Educational Capital is Low.

On the other hand, the coefficient of dummy variable “Female” is -0.65, suggesting

that after controlling for other explanatory variables, the predicted monthly payment

to females could be 65% lower than to Males.

The other explanatory variables whose coefficients do not result in significant statistics

associated with the monthly salary seem to be a bit historically unexpected. The most

noticeable cases are the dummy variables “To become a Post Graduate” and Age. As

28 | P a g e

the statistical model suggests, maybe in 1996 the mode to “get ahead” for

professionals was not highly associated with continuing studies on a master’s course.

Additionally, the variable Age that could be associated with an increment in the

expected monthly “principal occupational salary” per year, does not present a

statistically significant prediction about the salary in the cases analysed in 1996. It is

important to take into account that an interaction and logarithmic transformation was

carried out on the variables without positive results being statistically associated with

the explanatory variables.

However, the question about the statistically significant differences expected in the

monthly occupational salary between the main categories of Educational Capital of

Parents and females/males in 1996 are highly connected with the theoretical evidence

and the previous studies considered for this thesis. Indeed, the analysis implemented

is coincidently connected with the literature about Chile supported by Torche (2005,

2010a), Nuñez and Miranda (2009), Espinoza (2011), (Bravo et al, 2008), who support

the idea about the important levels of differentiation and inequality in Chile. However,

the focus in this research has been on narrowing the analysis on Professionals with a

higher degree working in the labour market.

Are these conditions in 1996 fair or understandable to someone that reached a certain

level of educational attainment? This statement will be explored through the analysis

of the main results.

29 | P a g e

The next analysis looks to carry out a different analysis considering the same number

of professionals workers in the labour market with a higher degree (n=147), but five

years later. The question here is what happened to the same people five years later?

However, it is important to take into account that it is not a statistical longitudinal

study, given the limitations of time and knowledge gain for the researcher developing

this thesis. This Issue was discussed in the limitations of the current dissertation.

30 | P a g e

Regression analysis based on “Main Occupational Salary” panel survey Year 2001

In order to explore the state of the explanatory variable in the same sample (n=147), a

second regression was fitted. In the current analysis, it was carried out considering an

explanatory variable available and with less attrition problem. The most reliable

variable with the research variables were ‘Changes in the occupation’, considering the

actual changes of the cases between 1996 and 2001. The model was fitted presented

as follows.

Table 4: Multiple linear regression Principal Occupational Salary 2001 on selected independent

variables

MODEL 2

B

Se

Parental Educational Capital Intermediated

0.131

(0.105)

Parental Educational Capital High

0.355~

(0.201)

Female

-0.378***

(0.0999)

Age 1996

-0.00226

(0.00640)

Live in the Capital’s Country 1996

0.259*

(0.0992)

To receive a Postgraduate degree 1996

0.362

(0.197)

Changes in Occupation between 1996 and

2001

-0.343**

(0.111)

Intercept

13.04***

(0.315)

N

147

R-sq

0.259

adj. R-sq

0.222

Rmse

0.586

Prob > F

0.0000

=~ p < 0.10; *p<0.05; **p<0.01; ***p<0.001

Source: Own elaboration based on Casen Panel 1996-2001

The predicted model on the monthly average of “Principal Occupational Salary 2001” is

highly significant, as is indicated by the F test (p<0.001). Additionally, the R2 indicates

that close to 26% of the variance of dependent variables is explained by the model

fitted.

31 | P a g e

The model indicates a significant partial linear association between the monthly log of

“Main Occupational Salary 2001” and the dummy variable Female (t=-3.79; p<0.001).

In addition, the model suggests a significant partial association with the dummy

variable “Changes Occupation between 1996 and 2001” (t=-3.08).

The interpretation of the estimated statistical significant coefficient is as follows:

The coefficient of dummy variable “Female” is -0.378, suggesting that even after

accounting for the control variables the predicted monthly pay for women decreased

by 38% compared to that for men.

Interestingly, the coefficient of dummy variable “Changes Occupation between 1996

and 2001” is -0.343. That indicates that controlling for all the exploratory variables, the

expected salary for people who changed their work between 1996 and 2001

decreased by 34% more than for people who remained in their jobs.

One interesting issue from the 2001 selected model indicates that coming from an

Educational Capital of Family intermediate or High, compared with coming from

Educational Capital of Family Low, does not significantly affect the expected (log of)

Monthly Main occupational salary average. It can be underlined that the null

32 | P a g e

hypothesis of association between those dummy variables and the explanatory one

cannot be rejected at a significant level of 5%. That fact could indicate a continuity of

the differentiation in the expected main occupational salary after five years, from

people coming from a more educated family. However, this fact must be taken

cautiously because the research was not carried out using a proper longitudinal

analysis.

However, from a other statistical standpoint it can be argued that the null hypothesis

can be rejected at a significant level of 10%. Therefore, assuming that level of

significance, the predicted (log of) main occupational salary coming from a family with

a High Educational Capital (t=1.77; p<0.079) increased by 36% compared to those

coming from a Low Educated Family. That fact could indicate a continuity of the

differentiation in the expected main occupational salary after five years for people

coming from a more educated family. Unfortunately, in order to retain enough cases

to fit a multiple linear regression, it was not possible to analyse the sample doing a

cross tabulation by the same type of Work because the bases of cases involved

decreased so much.

From another contextual and historical standpoint, the statistical association indicated

in the model with the dummy variable “Changes Occupation between 1996 and 2001”

reveals an historical Issue. During the years 1998-2000 Chile and others countries from

Latin-America suffered the effect of the “Asiatic Crisis” (Chang and Velasco, 2000).

That can explain the statistical significance that this variable had in the predicted main

occupational salary for professionals at that time.

33 | P a g e

In spite of there being two totally different statistical models (different R2, residuals,

coefficients, statistical level of acceptance of null hypothesis, etc.) and conceptual

terms, there is a noticeable figure here. On the one hand, the models show that after 5

years, coming from a family with a Higher Educational capital is found to have a

positive effect on the predicted main occupational salary compared to people coming

from a family with a Low educational Capital, holding constant all the explanatory

variables in the models. On the other hand, the gap between those people coming

from intermediate Educational Capital and people coming from families with a low

Educational Capital presents an apparent convergence through the years. However,

this result considers a general sample without considering differences by age, kind of

job, or occupation. The sample size and the attrition problems created difficulties for

that task.

At the same time, result was important for noticing that females five later keep

significant differences theoretically connected with the theoretical framework and the

research evidence shown earlier. However, it is possible to notice with the current

evidence that in our sample the predicted salary of females is reduced compared with

males, but that those predicted outcomes or differences after the five year result are

no lower than in 1996. In other words, in 2001 the female salary is lower but not lower

than in 1996.

34 | P a g e

What does a cross sectional view shows?

In this section I would like to show figures from the cross sectional survey CASEN 2006.

This analysis will be carried out using the same variables presented in the previous

section. However, as was demonstrated in the last section, some variables have been

presented in different categories, changing a little the meaning of the interpretation.

The idea is to deepen the analysis using a larger sample size and complementing some

explanatory aspects related to our response variable “Principal Occupational Salary”.

This perspective allows us to follow some general patterns and underline some other

variables that were not present in the panel data set because some of them were

omitted.

In sum, this different view can be more helpful to support our statement that has

been followed through this thesis.

On the other hand, graph 2 shows from a general viewpoint, that there are significant

statistical differences in the means of the main occupational salary by age groups.

The one way ANOVA reveals that there are significant differences between those

groups called Low, Intermediate and High “Parental Educational Capital”. Therefore,

there is significant evidence against the null hypothesis which indicates that the means

are not different for the population (p-value <0.001) in the ANOVA analysis and

Bartlett's test for equal variances.

35 | P a g e

Graph 2

991,931

1,139,557

1,232,710

1,375,156

719,074

892,110

884,446

991,222

582,731

643,078

702,970

811,107

0

200,000

400,000

600,000

800,000

1,000,000

1,200,000

1,400,000

1,600,000

26-36

37-46

47-56

57-65

Princip al Occupational Income (Chilean Pesos)

Principal Occupational Income by Parenta l Educational Capital and Age

Groups (2006)

Professionals with University Degree Finished (n=3243)

High

Intermediated

Low

Source: Own Elaboration based on Casen 2009. Ministry of Planification of Chile

Analysis of Variances: Pr ob>F="*p<0 .05 ** p<0.01 *** p<0.0 01"

***

*** *** *** ***

Considering these significant statistical differences and characteristics

5

two linear

regression models were fitted.

Regression analysis based on “Principal Occupational Salary” Cross sectional Survey

CASEN 2006

The table 5 shows the two models fitted. The first model was designed considering just

three dummy variables and one continuer. The second model included more

explanatory variables that seem to be statistically significant to predict the log of

“Main Occupational Salary”. Some of these variables have been suggested by the

5

A normality test it was carried out, and is presented in the appendix B.

36 | P a g e

theoretical or contextual assumptions guiding our research questions. Likewise, the

two models complement each other to present how much variance of the explanatory

variable is explained in conjunction with the explanatory variables considered by the

models. Therefore, R2 reflects the variation around the fitted regression line.

One of the main guiding hypotheses is that there are statistically significant differences

in the predicted log of “Main occupational salary” by Educational Capital of families

and gender in the Chilean professionals. Now with a data set with a wide sample of

cases we would like to fit a more reliable prediction. However, the analysis on the

Cross sectional CASEN 2006 does not have a view after 5 years as has been previously

indicated by the panel CASEN 1996-2001.

37 | P a g e

Table 5: Multiple linear regression6 Log Principal Occupational Salary of Professionals in 2006

on selected independent variables

Model 1

Model 2

B

Se

B

Se

Educational Capital Family

Intermediate

0.131***

(0.0272)

0.0556

(0.0291)

Educational Capital Family High

0.327***

(0.0297)

0.214***

(0.0302)

Female 2006

-0.414***

(0.0232)

0.0886*

(0.0394)

Age 2006

0.00898***

(0.00121)

-0.369***

(0.0224)

Father Occupation as Head or

Manager

0.0153***

(0.00405)

Live in the Capital’s Country

2006

0.291***

(0.0243)

To receive a Postgraduate degree

2006

0.122***

(0.0352)

Occupation: as Self-employed

person 2001

-0.394***

(0.0585)

Occupation: as Worker or

Employee at Public Sector 2001

-0.578***

(0.0499)

Occupation: as Worker or

Employee at Private Sector 2001

-0.578***

(0.0490)

How do generally got a Job?:

Other way

0.0977***

(0.0229)

Born in a different district to

where they live

0.0806***

(0.0238)

Was born between 1980-1976

0.282

(0.153)

Was born between 1975-1966

0.279*

(0.125)

Was born between 1967-1955

0.229*

(0.0899)

Was born between 1954-1946

0.113

(0.0668)

Ethnicity

-0.0901

(0.0596)

Intercept

12.86***

(0.0626)

12.72***

(0.266)

N

3243

R-sq

0.140

0.235

adj. R-sq

0.139

0.231

Rmse

0.654

0.617

Prob > F

0.001

= *p<0.05; **p<0.01; ***p<0.001

Source: Own elaboration based on Casen 2006

6

Consulting on linearly and heteroskedasticity assessment - please see in the appendix B for fitted

models.

38 | P a g e

Model 1 2006

Model 1 indicates statistically significant evidence that there is a linear association

between the logarithm of “Main Occupational Salary” of professionals and the

explanatory variables in the population. The R2=0.14 demonstrates that 14% of the

observed variation of “Main Occupational Salary” by professionals is explained by

variation in the variables included in the model.

The table 5 also show statistically significant association between the predicted log of

monthly “main occupation salary” by the dummy variables “Educational Capital of

family” Intermediate (t=4.38; p<0.001) and high (t=10.70; p<0.001), as well as with the

dummy variables of “Females” (t=-17.93; p<0.001) and the continue variable of age

(t=7.67; p<0.001), holding constant the correspondent explanatory variables. Each of

these variables, as in the previous analysis on the years 1996 and 2001, revealed a

statistically significant association with the response variables log of Main occupational

Salary.

Interestingly, the variable Age in this model’s result was positively associated with the

predicted Monthly “Main occupation Salary”. However, it can be seen that the effect in

the prediction of the salary is very small. The coefficient to this exploratory variable in

the model 1 is 0,009, suggesting that holding Educational Capital of Family

Intermediate and High, gender constant, and increasing age by one year, increases the

expected Monthly “Main Occupational Salary” by almost 1%.

It can be noticed that an interaction was fitted between the dummy variable female

and age but this was not shown to be statistically significant associated with the

39 | P a g e

response variable in this model. In addition, there was an interaction between the

dummy variables, Parental Educational Capital Low, Intermediate, High and females,

but this did not provide a statistically significant result fitting a regression model with

those interactions.

Model 2 2006

Model 2 indicates statistically significant evidence that there is a linear association

between the average “Main Occupational Salary” of professionals and the explanatory

variables in the population. The R2=0.28 demonstrates that 28% of the observed

variation of “Main Occupational Salary” by professionals is explained by variation in

the variables included in the model.

The current model contrasts with mode 1 and shows that the dummy variables

referred to as Parent Educational Capital Intermediate leave off being partially

associated with the explanatory variables. However, the explanatory dummy variables

referred to as Parent Educational Capital High (t= 7.07; p<0.001), the dummy variables

female (t=-16.48; p<0.001) and the continue Age (t=3.78; p<0.001),still have almost the

same statistical significant pattern that indicates a partial association with the

predicted value of the response variable.

A particular note takes the dummy variables about kinds of Occupations that are

compared with the baseline Head and Managers. In fact, every single dummy variable

of kinds of occupations results are partially linearly associated with the predicted value

of log of main occupational salary, suggesting that for their coefficient of being a

professional Self-employed (t=-6.74; p<0.001), Worker or Employee at Public Sector

(t=-11.60; p<0.001) or Private Sector(t=-11.78; p<0.001), the predicted salary could be

40 | P a g e

between 39% or 58% lower than being a Head or Manager, holding constant each

single dummy variable one by one, respectively.

Likewise, the model indicates that other explanatory variables can be partially

associated with Monthly Main Occupational Salary. Such is the case of the dummy

variable “Live in the Capital’s Country 2006” (t=11.96; p<0,001), that its coefficient is

0.290. This suggests that for a professional living in Santiago, Chile (Administrative

Capital) the expected average pay increases by 29% compared to some professionals

living outside the Capital, holding constant all the variables.

On the other hand, the dummy variable “Getting a Postgraduate degree 2006” (t=3.70)

is indicated in the model as having a statistical association with Monthly “Main

Occupational Salary”, with its coefficient being 0.122. Therefore, the expected salary of

professionals that have a Postgraduate degree is 12% higher than those who do not,

holding all the explanatory variables constant, respectively to each single dummy

variable.

The other interesting finding indicated by the model is the partial statistical association

of the dummy variable referred to as “Live in a different district to where you were

born” (t=3.39; p<0.001) with the predicted monthly salary, whose coefficient is 0.080.

That Suggests that the expected monthly salary increased by 8% in those professionals

who leave their birthplace compared to those who do not leave the district where they

they were born.

In addition, the model indicates a statistically significant association with the response

variables in the dummy “How do generally got a Job?: Other way“(t=4.27; p<0.001).

The coefficient of this dummy variable is 0.097. In other words, the expected monthly

41 | P a g e

salary increases by 10% for those who got jobs based on “Other modes” compared to

those who get jobs based on contacts such as relatives and friends.

In addition to this is the fact that model 2 indicates and confirms an interesting natural

process in the labour market, which is how the salary is logarithmically distributed

during the course of life in Chile. The dummy variables indicate that those born during

the years 1975-1966 (t=2.24; p<0.05) and between 1967-1955 (t= 2.54; p<0.05) are

statistically associated with the explanatory variable. The statistically significant

coefficients in both variables indicate that those that were born in those years increase

their predicted salary by 27% and 23% respectively, compared to the rest of the people

in the sample, all other variables remaining constant. Indeed, these differences are the

product of the natural process that people in the labour market are more likely to

receive better salaries when they are in a more productive phase (31-51 years old).

After this year, the curve of salaries could be reduced. Hence, the fact that the other

dummy variables related with group of those years have not shown a statistically

significant association with the predicted salary compared to people that were not

born during those years.

Contrary to our expectations, some variables were not statistically significant in the

model. That is the case of the dummy variable ethnicity. That dummy variable

compared people who get jobs based on public information and by creating their own

jobs with the baseline of those who get jobs based on contacts with relatives and

friends. Unfortunately, for a research proposal that seems to be theoretically

associated with the prediction of main occupational salary, that model does not fit. In

42 | P a g e

contrast, the dummy variable compared the “Other way” of how people get a job with

getting jobs based on contacts with relatives and friends.

The dummy variables ethnicity is not associated statistically with the predicted

monthly main occupational income average in the model. Of course, those variables

were just used instrumentally to give an account that at a professional level there is

not an issue if people belong to some ethnic group. From my point of view,

additionally, there is just a small group of people that can achieve the educational

attainment necessary to be a professional in Chile. Unfortunately, that is not reflected

in the results that the model indicates.

43 | P a g e

Analysing more deeply the predicted Logs of Main Occupational Salary controlled by

Gender

Table 6 shows multiple linear regressions analysis predicting the log of Main

Occupational Salary controlled by gender on selected exploratory variables. In fact,

these variables are the same that model 3 presented in the table 6. The idea is to draw

out which explanatory variables are statistically associated with the predicted monthly

salary for males and females.

The model that refers to males as well as females indicates statistically significant

evidence that there is partial linear association between the logarithm of “Main

Occupational Salary” and the explanatory variables in the population referring to male

professionals in Chile. The R2 demonstrates that controlling for males and females a

19% and 13%, respectively, of the predicted variation of the Log of Main Occupational

Salary is explained by variation in the variables included in each of the fitted models.

According to the male and female regression models the dummy variables Parental

Educational Capital High (t=7.20; p<0,001), the dummy variable female (t=-16.2;

p<0,001), Age(t=7.20; p<0,001), Living at the Capital(t=7.20; p<0,001) and all the

dummy variables about occupation are statistically significant in predicting the log

monthly main occupational salary.

44 | P a g e

Table 6: Multiple linear regression Log Principal Occupational Salary of Professionals in 2006 Controlled by

Gender on selected independent variables

Males

Females

B

Se

B

Se

Educational Capital Family Intermediate

0.0587

(0.0435)

0.0578

(0.0390)

Educational Capital Family High

0.246***

(0.0439)

0.188***

(0.0417)

Father Occupation as Head or Manager

0.0957

(0.0560)

0.0833

(0.0548)

Age 2006

0.0147*

(0.00595)

0.0166**

(0.00552)

Live in the Capital’s Country 2006

0.284***

(0.0346)

0.300***

(0.0341)

To receive a Postgraduate degree 2006

0.205***

(0.0502)

0.0432

(0.0491)

Occupation: as Self-employed person 2001

-0.330***

(0.0743)

-0.497***

(0.0968)

Occupation: as Worker or Employee at

Public Sector 2001

-0.617***

(0.0641)

-0.588***

(0.0825)

Occupation: as Worker or Employee at

Private Sector 2001

-0.543***

(0.0617)

-0.643***

(0.0828)

How do generally got a Job?:

Other way

0.112***

(0.0334)

0.0859**

(0.0313)

Born in a different district that live

currently

0.119***

(0.0350)

0.0449

(0.0323)

Was born between 1980-1976

0.143

(0.221)

0.470*

(0.213)

Was born between 1975-1966

0.182

(0.178)

0.437*

(0.176)

Was born between 1967-1955

0.159

(0.126)

0.348**

(0.131)

Was born between 1954-1946

0.0334

(0.0891)

0.242*

(0.104)

Ethnicity

-0.181*

(0.0910)

-0.024

(0.1037)

Intercept

12.78***

(0.388)

12.24***

(0.369)

N

1587

1604

R-sq

0.185

0.128

adj. R-sq

0.177

0.120

Rmse

0.630

0.602

Prob > F

0.001

= *p<0.05; **p<0.01; ***p<0.001

Source: Own elaboration based on Casen 2006

One of the main differences is the partial association that a postgraduate has in the

predicted salary for males. In fact, the predicted log of main income for a male

postgraduate (t=4.09; p<0,001) increases by 20.5% more than for a females with a

degree, holding constant all the others variables . Likewise, males that have changed

in 2006 the district were they born (t=3.38; p>0, 01) increase their predicted main

occupational salary by 12% more than those who do not change until 2006 the district

where they were born, holding constant the rest of the explanatory variables.

45 | P a g e

Additionally, it can be seen that for males belonging to an ethnic group in Chile there is

a significant negative association with the predicted monthly salary. For males that

belong to an ethnic group their predicted pay is 18% lower than for men who do not

belong.

Meanwhile, females present other kinds of differences when the explanatory variables

and coefficient are compared to males. One of the most relevant is the concentration

of statistically significant coefficients predicting the monthly occupational salary

referring to dummy variables of groups of ages. For instance, females that were born

during the years 1967-1955 (t=2.62; p>0.001) receive a monthly salary 35% higher than

those not born in those years, holding constant all other explanatory variables. Similar

statistically significant and positive effects on the predicted salary occur from 1954 to

1945 (t=2.32; p>0.01), between 1980 and 1976 (t=2.18; p>0.05) and through 1975-

1966 (t=2.46; p>0.05), with an increment in the predicted monthly salary by 24%, 43%,

46%, respectively, compared with those who were not born in that period.

On the other hand, it is important to take into account that the effect in the predicted

main salary is different for males and females. In fact, the predicted salary for females

that have a postgraduate degree increases by 4% compared to females without one,

holding constant all variables. While the predicted salary for males that have a

postgraduate degree increases by 21% compared to those that do not have a

postgraduate degree title.

Overall, these results provide support for the view that Parental Educational Capital

plays an influential role in the differences in the predicted average of salary, in spite of

46 | P a g e

the differences between males and females that the different models reveal. Not

surprisingly, the results show that there are continuing and systematically differences

in the social system for professionals in Chile, which confirms recent research that has

evidenced a similar trend in other countries in Latin America and in some other

industrialized countries. However this difference has been revealed in a preferential,

or what we would call an “elite”, group.

47 | P a g e

V. Discussion

We now return to the questions proposed in the introduction. One of the main results

that can be drawn is the constant differences that have been found between the

predicted occupational salaries by Parental Educational Capital. Those facts are

coincident with the theoretical and research general approaches to the social

inequality system in Chile and Latin America (Torche; 2005, 2010a). However, it is not

related to evidence coming from industrialized and western countries that points out a

significant decline in the parental origin with regard to educational attainment and

labour market analyzing those effects by income or salary.

At that point, an interesting and debatable issue comes from the analysis of the period

1996 to 2001, not only in statistical terms, but also in conceptual and theoretical ones.

As it has been shown, assuming a 10% level of significance the predicted (log of) main

occupational salary in 2001, coming from a family with a High Educational Capital

increased by it 36% compared to those coming from a Low Educated Family. That fact

statistically indicates a continuity of the differentiation in the expected main

occupational salary after five years for people coming from a more educated family.

However, this fact must be taken cautiously, not only in assuming that level of

significance, but because the research has been not carried out using a proper

longitudinal analysis. More accurately, it has been a sort of differentiated analysis in

two times on two equal samples, but by carrying a cross-sectional analysis.

In spite of those facts, and from a theoretical standpoint, the statistical difference

between people coming from a high and a lower educational background reveals that

48 | P a g e

after 5 years the social systems seem to be working properly in classifying structural

differences in these groups of professional in Chile. This fact is in contrast to the

evidence provided by Torche (2007) that Chile is a more flux society in middle class

groups. How can we understand that claim if at a high educational level differences

from social origin are revealed?.

In that respect, two main ideas can be followed to respond to those differences. On

the one hand, it can be a product of the personal and individual selection of

particularly higher educational careers (such as lawyer, doctor, engineer etc.) when

these people were young. That issue, is moreover, one of the main weaknesses and

limitations of the current research. On the other hand, from a conceptual stand point

the definition of social class, origin and the entire concept created in the rest of the

research, has been operationalized and conceptualized in our case from just two

available variables, the Level of education of father and mother. That, of course, is due

to the availability of the data sets. However, independent of this fact our research has

pointed out some interesting similarities and differences that can open new questions

about this topic.

Additionally, the current model 1996-2001, indicates a kind of convergence in the

predicted salary between those coming from an intermediate and those coming from a

low educational capital family. Again, independent of the conceptual and empirical

differences in how the research was carried out, that is in fact a finding that all the

research into social inequality and mobility in Latin America and Chile has concluded.

Maybe this important convergence from people with a low Parental Educational

Capital has been approaching their transitions to people that have an intermediate

49 | P a g e

Parental Educational Capital. From our point of view, it could be this gap and the

approaches to flux and faster change that those others studies support, that is the final

conclusion of the research carried out.

The overall research also raises evidence supported by different studies and research

about the income inequalities for gender, which is our second question. In fact, the

evidence in this research suggests that females are likely to be the group of people

that need more attention, especially if they come from a less educated background.

The data base from 2006 gives us the opportunity to account for the association of

other variables on the predicted main occupational income, and provides insight

regarding the importance of the main occupational income that the father has,

whether they live in the capital or have received a professionals’ postgraduate degree

that year. All these variables give us important and robust evidence to understand the

social context in which professionals from different origin are actually living. However,

the analysis of the predicted salary by gender is fruitful and reveals that it is men who

more likely to use some “strategies” to reach a kind of differentiation, specifically to

get a postgraduate degree. Theoretically, that fact can be supported by the theory of

effectively maintained inequality (Lucas, 2001) that points out the utilization of

strategies by more advantaged groups to maintain their position. However, specifically

how this qualitative difference is actually used by males in the labour market cannot

be totally revealed by the current research. Rather, it can just be suggested that there

is a special movement in the higher educational system that seems to be associated

more with males than with females. Is this a pragmatic strategy from those more

advantaged or a strategy from those less socio-cultural advantaged?

50 | P a g e

Overall, according to Bourdieu’s (1995, 1993, 1978) standpoint, we are witness of a

`resemblance within a difference' Chilean social system where entry to the higher

educational system seems to be very open, but with social inequalities that are

reproduced by generations. However, the structural mechanism that generates this

system is very difficult to reveal in this kind of research. Therefore, more complex

research models and research strategies using longitudinal statistical tools as well as a

deep use of qualitative techniques could help reveal those aspects. For the impact

that this study may have, it is necessary to focus research and public policy resources

on seeing the reality of the women in the labour market, especially if they come from a

disadvantaged social and cultural origin.

51 | P a g e

VI. Conclusions

Assuming the main points of the current dissertation, the conclusions can be divided as

follows.

First, the importance of new and more relevant applied research that looks for a

strong connection between public policy and academia. Concretely, the information

available about into the higher educational system has to generate useful research

that seeks to study the consequences of more and more people leaving the

educational system to find a job in the labour market. At the same time, the challenge

to academia is the outcomes that can be considered by the government and by public

policy in education. In short, the challenge will be how academia can translate properly

and easily the theoretical and applied findings.

Second, connected with the previous point, it is necessary to expand more and more

the research focus in actual strategies that people coming from low, intermediate and

high origins have used during their course of life. We need to study aspects of social

strategies that EMI has pointed out as interesting theoretical insights to develop

theoretically and empirically those sociological topics in Chile. For instance, research

about the educational and labour market and how people coming from middle and

lower class are currently obtaining position.

52 | P a g e

VII. Appendix A

Summary Statistics

Casen Panel 1996

Casen Panel 2001

53 | P a g e

Casen 2006

54 | P a g e

VIII. Appendix B

PANEL CASEN 1996-2001 Regressions Models and other test applied

Model 1

.

errores 3.9e+03 0.0 046 0.00 00 177.42 0.0000

Variable Obs Pr(Ske wness) Pr(Kurt osis) chi2( 2) Prob>chi2

joint

Skewness/K urtosis tests fo r Normality

. sktest errore s, noad

errores 4.857196 -.111 0446

variable kurtosis skew ness

. tabstat error es, stats(kurto sis skewness)

(22970 missing values generate d)

. predict error es, resid

Prob > chi2 = 0.00 00

chi2(1 ) = 32. 75

Variab les: fitted val ues of logyo96

Ho: Co nstant variance

Breusch-Pagan / Cook-Weisberg test for heteros kedasticity

010 20 30

sqrt(frequency)

-8 -6 -4 -2 0 2

Residuals

55 | P a g e

Model 2 Regressions Models and other test applied

errore s 208 0.0000 0.0000 139.07 0.0 000

Variabl e Obs Pr(Skewness) Pr(Kurtos is) chi2 (2) Prob> chi2

joint

Skew ness/Kurtosi s tests for Normality

. sktest er rores, noad

errore s 11.17724 2.72534

variabl e kurtosis skewness

. tabstat e rrores, stats (kurtosis sk ewness)

(26674 miss ing values ge nerated)

. predict e rrores, resid

Pr ob > chi2 = 0.4691

ch i2(1) = 0.52

Va riables: fitt ed values of logyo01

Ho : Constant va riance

Breusch-Pag an / Cook-Wei sberg test f or heteroske dasticity

-2 0246

sqrt(frequency)

-2 0 2 4 6

Residuals

56 | P a g e

CASEN 2006 Regressions Models Regressions Models and other test applied

Model 1

.

_cons 12.85778 .0614876 209.11 0.000 12.73722 12.97834

edad .0089835 .001171 7.67 0.000 .0066875 .0112796

_Isexo_2 -.4143171 .0231051 -17.93 0.000 -.4596192 -.3690151

_Icapfam3_2 .3267688 .0305467 10.70 0.000 .266876 .3866617

_Icapfam3_1 .1308008 .0298705 4.38 0.000 .0722337 .1893678

logyopraj Coef. Std. Err. t P>|t| [95% Conf. Interval]

Total 1611.4559 3242 .497056107 Root MS E = .65404

Adj R-squared = 0.1394

Residual 1385.097 3238 .427763124 R-squar ed = 0.1405

Model 226.358903 4 56.5897258 Prob > F = 0.0000

F( 4, 3238) = 132.29

Source SS df MS Number of obs = 3243

i.sexo _Isexo_1-2 (naturally coded; _Isexo_1 omitted)

Model 2

57 | P a g e

errores 4.0e+04 0. 000 0.00 0 1631.50 0.0000

Variable Obs Pr(Skewn ess) Pr(Kurtosi s) chi2(2) Prob>chi2

j oint

Skewness/Kurtosi s tests for Norma lity

. sktest errores , noad

errores 4.760454 .0554 801

variable kurtosis skewne ss

. tabstat errore s, stats(kurtosi s skewness)

(228581 missing values generated )

. predict errore s, resid

Prob > chi2 = 0.0000

chi2(1) = 39.5 2

Variabl es: fitted value s of logyopraj

Ho: Con stant variance

Breusch-Pagan / Cook-Weisberg te st for heterosked asticity

-20 020 40 60

sqrt(frequency)

-6 -4 -2 0 2 4

Residuals

58 | P a g e

IX. References

AGRESTI, A. AND FINLAY, B. 2009 Statistica Methods for the Social Scinces. Fourth

Edition. Pearson Prentice Hall.

BRAVO, D., SANHUEZA C. and URZÚA, S. 2008 Ability, Schooling Choices and Gender

Labor Market Discrimination: Evidence for Chile. Research Network Working papers ;

R-558) Inter-American Development Bank. Felipe Herrera Library

BRAVO D. AND VASQUEZ J. 2008. Apuntes de Microeconometría Aplicada. Magister en

Políticas Públicas. Facultad de Economía Universidad de Chile. Agosto, 2008

CABALIN CRISTIAN (2012) Neoliberal Education and Student Movements in Chile:

inequalities and malaise, Policy Futures in Education, 10(2), 219-228.

CASTILLO, J. & CABEZAS, G. (2010) Caracterización de jóvenes primera generación en

educación superior. Nuevas trayectorias hacia la equidad educativa. Revista Calidad en

la Educación. Nº 32, julio 2010 p.112 118.

CASTILLA, E. and SOFFIA, M. 2008. ‘‘Gender, Race, and Meritocracy in Organizational

Careers.’’ American Journal of Sociology 113(6): 1479–526.

CANO, V. 2009. Los estudios sobre migración internacional en Chile: apuntes y

comentarios para una agenda de investigación actualizada. Papeles de Población, Vol.

15, Núm. 61, julio-septiembre, 2009, pp. 129-167. Universidad Autónoma del Estado

de México – México

CHANG MARIKO LIN (2000) The Evolution of Sex Segregation Regimes American

Journal of Sociology , Vol. 105, No. 6 (May 2000), pp. 1658-1701

Chang R. and Velasco, A. 2000. "Liquidity Crises in Emerging Markets: Theory and

Policy," NBER Chapters, in: NBER Macroeconomics Annual 1999, Volume 14, pages 11-

78 National Bureau of Economic Research, Inc.

BAROZET, E (2010). Topología de las clases medias chilenas: Análisis de la mesocracia

en un país desigual. Ponencia presentada en el II Congreso Internacional Ciencia,

Tecnologías y Culturas: Diálogo entre las disciplinas del Conocimiento. Mirando al

futuro de América Latina y El Caribe, Universidad de Santiago de Chile, 29 de Octubre

al 1 de Noviembre de 2010

BOLIVER, VIKKI (2011). Expansion, differentiation and the persistence of social class

inequalities in British higher Education. Higher Education: The International Journal of

Higher Education Research 61(3): 229-242

BERNSTEIN B 1996 Pedagogy, symbolic control and identity: theory, research, critique

(London, Taylor and Francis)

59 | P a g e

BOURDIEU P. 1992 The rules of art: genesis and structure of the literary field

(Cambridge, Polity Press)

BOURDIEU P. 1993 The field of cultural production (Cambridge, Polity Press)

BOURDIEU P. and Passeron J.C. 1990 Reproduction in Education, Society and Culture.

SAGE Publications, 1 Oct 1990. Second Edition.

BOURDIEU P. WACQUANT LJD 1992 An invitation to reflexive sociology (Cambridge,

Polity Press)

BLALOCK, H. M. (JR.) 1968 ‘The measurement problem: A gap between the languages

of theory and research’, in H. M. Blalock (jr.) and A. B. Blalock (eds) Methodology in

Social Research, New York:McGraw Hill.

ERIKSON, R. AND GOLDTHORPE, J. H. 1993 The Constant Flux: A Study of Class Mobility

in Industrial Societies, Oxford, Clarendon Press.

ESPINOZA, VICENTE, (2011) Movilidad ocupacional en Chile 2001-2009. ¿Desigualdad

de ingreso con igualdad de Oportunidades”. Forthcoming Universidad de Santiago de

Chile.

GOLDTHORPE, J. WITH LLEWELLYN, C. AND PAYNE, C. (1987) Social mobility and class

structure in modern Britain. Oxford: Claredon Press 2nd edition.

GOLDTHORPE, J.H. (1996). Class Analysis and the Reorientation of Class Theory: The

Case of Persisting Differentials in Educational Attainment. The British Journal of

Sociology, Vol. 47, No. 3, Special Issue for Lockwood (Sep.,1996), pp. 481-505

GOLDTHORPE, J. AND BREEN (2001): Class Mobility and Merit. The experience of two

british Cohorts. European Sociological review Vol 17 N2 81-101.

GOLDTHORPE, J. AND JACKSON, MICHELLE.(2008)EDUCATION-BASED MERITOCRACY:

THE BARRIERS TO ITS REALISATION. In Social class: how does it work? Laureau and

Conley Editors. Sage publications.

GOLDTHORPE, J. (2010) Class analysis and the reorientation of class theory: the case of

persisting differentials in educational attainment. The British Journal of Sociology

Special Issue: The BJS: Shaping sociology over 60 years Volume 61, Issue Supplement

s1, pages 311–335, January 2010.

JACKSON, M. (2007), How far merit selection? Social stratification and the labour

market. The British Journal of Sociology, 58: 367–390. doi: 10.1111/j.1468-

4446.2007.00156.x

HARVEY, D (2005) A Brief History of Neoliberalism. Oxford University PressBerkeley:

University of California Press.

HARVEY, D (2010) The Enigma of Capital and the Crises of Capitalism. London: Profile

Books.

60 | P a g e

HOGAN, R. 2001. “Class, Race, and Gender Inequality,” Race, Gender & Class 8, 2:61-

93

HOGAN, R. 2011. Deconstructing Race, Class, and Gender Inequality in Personal

Earnings. Purdue University. Orphan Working Paper (still in need of a good home

somewhere in the discipline) Prepared in part for my Social Inequality Seminar at

Purdue. April 2011

HOGAN, RICHARD, PERRUCCI., C. , and BEHRINGER. A. 2005. "Enduring Inequality:

Gender and Employment Income in Late Career." Sociological Spectrum 25:53-77.

JACKSON, M, GOLDTHORPE J.H AND MILLS, COLIN (2005) Education, employers and

class mobility Research in Social Stratification and Mobility, Volume 23, 3–33. doi:

10.1016/S0276-5624(05)23001-9

JOHNSON R.B., ONWUEGBUZIE A.J., TURNER L.A. (2007). Toward a Definition of Mixed

Methods Research. Journal of Mixed Methods Research Volume 1 Number 2 April,

Sage Publications. 112-133

LUCAS, S. R. 2001 Effectively Maintained Inequality: Education Transitions, Track

Mobility, and Social Background Effects American Journal of Sociology , Vol. 106, No. 6

(May 2001), pp. 1642-1690

MINISTRY OF EDUCATION. (2005). Higher Education in Chile: Aiming for Quality.

Santiago, Chile: Division of Higher Education, March, 2005.

NÚÑEZ, J. and A. TARTAKOWSKY (2007) “Inequality of Outcomes vs. Inequality of

Opportunities in a Developing Country. An exploratory analysis for Chile”, Estudios de

Economía 34 (2), Diciembre, 185-202.

NÚÑEZ, J. and A. TARTAKOWSKY ( (2011): The relationship between income inequality

and inequality of opportunities in a high-inequality country: the case of Chile, Applied

Economics Letters, 18:4, 359-369

NÚÑEZ, J. and MIRANDA, LESLIE (2009) La movilidad intergeneracional del ingreso y la

educación en Chile in Aguilar et al. El Arte de Clasificar a los Chilenos. Enfoques sobre

los modelos de estratificación en Chile. Ediciones Universidad Diego Portales

MINISTRY OF SOCIAL DEVELOPMENT. 2012. OFFICIAL DATA SET CASEN 2006.

PADILLA, F. (1985) Latino Ethnic Consciousness: The Case of Mexican Americans and

Puerto Ricans in Chicago. Notre Dame: University of Notre Dame Press.

PETERSEN AND MORGAN (1995) Separate and Unequal: Occupation-Establishment Sex

Segregation and the Gender Wage Gap. American Journal of Sociology, Vol. 101, No. 2

(Sep., 1995), pp. 329-365

61 | P a g e

PORTES , A. (1998) Social Capital: Its Origins and Applications in Modern Sociology.

Annu. Rev. Sociol. 1998. 24:1–24

RAJANI NAIDOO (2004): Fields and institutional strategy: Bourdieu on the relationship

between higher education, inequality and society, British Journal of Sociology of

Education, 25:4, 457-471

SAUNDERS, P. (1995) 'Might Britain be a Meritocracy?', in Sociology, vol 29, no 1: 23 –

41

SOCIAL OBSERVATORY ALBERTO HURTADO UNIVERSITY (2007) La Encuesta Panel

CASEN: Metodología y Calidad de los Datos. On web page:

http://www.osuah.cl/documentacion_encuestapanelcasen/La_encuesta_Panel_CASEN

_-_Metodologia_y_Calidad_de_los_Datos.pdf visited 6th February 2012

TARLING, R. 2009 Statistical Modelling for Social Science. Principle and practice.

Routledge.

TORCHE FLORENCIA (2010a), Opinión pública y los determinantes de la pobreza y la

riqueza en Chile. In Movilidad social y el dilemma de más o menos Estado. Encuesta

Nacional Bicentenario Universidad Católica - Adimark : «Una Mirada Al Alma De Chile»

Enero. 2010

TORCHE FLORENCIA (2010b) Economic Crisis and Inequality of Educational Opportunity

in Latin America Sociology of Education April 2010 83: 85-110,

doi:10.1177/0038040710367935

TORCHE, F. and L. LOPEZ-CALVA (2011). ‚Stability and Vulnerability of the Latin

American Middle Class‛, forthcoming in K. NEWMAN ed. Dilemmas of the Middle Class

around the World, Oxford University Press, New York.

TORCHE FLORENCIA (2005) “Unequal but Fluid: Social Mobility in Chile in Comparative

Perspective”, American Sociological Review, 70 (3), pp. 422-450.

THEMELIS, SPYROS (2008): Meritocracy through education and social mobility in

post‐war Britain: a critical examination, British Journal of Sociology of Education, 29:5,

427-438

UNDURRAGA, JOSÉ TOMAS (2011) Rearticulación de grupos económicos y renovación

ideológica del empresariado en Chile 1980-2010: Antecedentes, preguntas e hipótesis

para un estudio de redes. Serie Working papers SO-UDP Nº - 2011. On web page:

http://www.icso.cl/wp-content/uploads/2011/03/Working-paper-elites-Tomas-

Undurraga1.pdf visited 6th February 2012

62 | P a g e

YOUNG, M. (1958) The rise of the Meritocracy. Penguin Harmondsworth. ERIKSON

ROBERT AND GOLDTHORPE JOHN H.(1996) Source: European Sociological Review, Vol.

8, No. 3, Special Edition on Social Stratification and Social Mobility (Dec., 1992), pp.

283-305

URZÚA, S. 2012 . La rentabilidad de la educación superior en Chile. Revisión de las

bases de 30 años de políticas públicas Estudios Públicos | N° 125.

http://www.cepchile.cl/

ZIMDARS, A., SULLIVAN, A., AND HEATH A. 2009 Elite Higher Education admissions in

the Arts and Sciences: Is cultural capital the key ?, Sociology 43 (4).

“Exploring social barriers in Chile: The influence of Parental Educational Capital and Gender on Salary in People with University Degree”

Figures

Recommended publications

Masculinity salience and homophobic reactions : affective and cognitive responses of sex-typed men t...

Femininity in The Strange Case of Dr. Jekyll and Mr. Hyde

France after Revolution. Urban Life, Gender, and the New Social Order

Kapitalismus und Gender. Eine Auseinandersetzung mit der kapitalismuskritischen Intersektionalitätsf...