Career patterns for IT engineering graduates

Page 1

http://eid.sagepub.com/Democracy

Economic and Industrial

http://eid.sagepub.com/content/early/2013/07/22/0143831X13492832The online version of this article can be found at:

 DOI: 10.1177/0143831X13492832

published online 22 July 2013Economic and Industrial DemocracyLine Holth, Abdullah Almasri and Lena Gonäs

Career patterns for IT engineering graduates  

Published by:

http://www.sagepublications.com

On behalf of: 

Department of Economic History, Uppsala University, Sweden

can be found at:Economic and Industrial DemocracyAdditional services and information for    

  http://eid.sagepub.com/cgi/alertsEmail Alerts:

 

http://eid.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

What is This? 

- Jul 22, 2013OnlineFirst Version of Record >>

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 2

Economic and Industrial Democracy0(0) 1 –17

© The Author(s) 2013Reprints and permissions:

sagepub.co.uk/journalsPermissions.navDOI: 10.1177/0143831X13492832

eid.sagepub.com

Career patterns for IT engineering graduates

Line HolthKarlstad University, Sweden

Abdullah AlmasriKarlstad University, Sweden

Lena GonäsKarlstad University, Sweden; Karolinska Institutet, Sweden

AbstractWomen constitute a clear minority in the field of information and communications technology (ICT) in higher education as well as in the job market. At the same time, this field is expected to have a shortage of qualified people in the future. Do women and men engineering graduates have the same career opportunities? This article problematizes the relationship between higher education in engineering and opportunities on the job market. The results show that men reach higher positions to a greater extent than women, and that women remain in low-qualification jobs to a greater extent than men.

KeywordsGender segregation, higher education, information and communications technology (ICT), labour market positions, multilevel analysis

Introduction

The development in Sweden regarding gender distribution in higher education and the differences in women’s and men’s job opportunities in relation to their qualifications is similar to that of many other countries, even though there are some important differ-ences (Berggren, 2007, 2008; Goldin and Katz, 2008; SOU, 2004). One difference is the narrow gender gap in the employment rate among the population with tertiary

Corresponding author:Line Holth, Department of Working Life Science, Karlstad Business School, Karlstad University, Universitetsgatan 2, Karlstad, 651 88, Sweden. Email: [email protected]

492832 EID0010.1177/0143831X13492832Economic and Industrial DemocracyHolth et al.2013

Article

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 3

2 Economic and Industrial Democracy 0(0)

education, which is much smaller in the Nordic countries than for example in the US, continental Europe, or in relation to the averages for the OECD countries (OECD, 2009). A similarity is that the majority of students entering undergraduate studies and receiving an undergraduate degree are women (HSV, 2008). Regarding new entrants to higher education, there has been a female dominance in Sweden for three decades now. Men were only in a clear majority in the field of technology. The situation is similar in the other Nordic countries (Kivinen et al., 2007). Gender stereotyped educational choices have been shown to increase horizontal segregation as demonstrated by the concentration of women and men in certain sectors and occupations (SOU, 2004). Women constitute around 20% of the total number of people in the ICT field (Computer Sweden, 2012) which is a heavily horizontally segregated field with women and men separated into different sectors, occupations and tasks (SOU, 2004; Tillväxtanalys, 2012).

The male dominance in technology and engineering has been documented in sev-eral studies (see Cockburn, 1983; Faulkner, 2000; Hacker, 1989; Lagesen, 2005; Mellström, 1995). Computer science has a strong association with masculinity and has been dominated by the male hacker and nerd culture since the 1960s (Mellström, 2009). Women have difficulty in identifying with these role models and therefore fail to see engineering as a suitable career choice. This has led to the exclusion of many women (Lagesen, 2005).

The bulk of IT employment is in the private sector. But the proportion of women in the public sector is far greater than in the private sector. In 2010, 30% of public sector employees in IT occupations were women, while the corresponding figure for the private sector was 18% (Tillväxtanalys, 2012: 60). There are also clear patterns of segregation in IT job categorization, which shows that woman and men work in different areas of specialization (Wahl et al., 2001). Only one-fifth of women work as a programmer, systems designer or computer technician (Tillväxtanalys, 2012). Statistics Sweden’s (SCB, 2010) analyses show that the occupational distribution among different educational groups varies considerably. In this study, the IT gradu-ates were specialized in electrical engineering, physics engineering and computer science, areas with a relatively low distribution across different occupational groups compared with other degree programmes in technology (SCB, 2010: 21). This means that these graduates tend to find jobs in IT engineering occupations – that is, the jobs they qualified for. A recent report from Tillväxtanalys (2012), however, shows that women IT graduates move into IT occupations to a lesser extent than men, and to a lesser extent than in previous years, which means that the proportion of women in the field is further reduced. As we focus on a specific group of IT graduates in this study, we do not make any claims on the positions of other graduate groups in the ICT field.

The proportion of highly educated people in the labour force is growing faster than the changes in the qualification structures in the labour market (Åberg, 2002). This state of affairs has resulted in a debate on over-education as a labour market problem since academically educated individuals have qualifications beyond those required for their current jobs. Women run this risk to a greater extent than men (Le Grand et al., 2005; Löfström, 2009a). A Swedish government report on the gender

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 4

Holth et al. 3

segregated labour market pointed to the future risk of a workforce of highly qualified women without the chance of getting positions that correspond to their level of educa-tion (SOU, 2004: 19). The Employment Agency report from June 2012 states that there is a shortage of highly qualified computer specialists. In particular, there is a demand for software and systems developers and IT architects. This indicates that in relation to labour market needs, too few women and men have chosen IT degree pro-grammes (Arbetsförmedlingen, 2012). The same pattern of computer specialist short-age and small proportion of women prevails in the whole of Europe (European Commission, 2009).

Löfström (2009b) specifies four reasons why women are at greater risk of being overqualified than men. The first is that more women than men are enrolled in higher education. A second reason is that the length of academic programmes leads to a rise in labour market entry age, which can mean that women who want to start a family before they are professionally established accept jobs below their qualification level. In addition, family formation means geographical immobility. The fourth reason is employers’ attitudes. If employers perceive women as a more ‘risky investment’ than men, it will affect the jobs offered to women. Swedish as well as international studies show that women and men have different career pathways into the labour market (Bettio and Verashchagina, 2009; Bihagen and Ohls, 2006) and that the horizontal segregation is strongly connected to the processes shaping the vertical segregation (SOU, 2004). The quantitative egalitarian position for women measured in employ-ment rates has to be placed in relation to more qualitative measures. Today, highly qualified women have almost the same employment rate as men, not only in Sweden (90% for men and 89% for women), but also in many other OECD countries (OECD, 2009). But the uneven distribution of women and men is evident not only in relation to the horizontal division but also in relation to hierarchical positions (Hebson and Rubery, 2004; Holter et al., 2009).The vertical dimension of the labour division shows that women in the ICT field are underrepresented in leading positions; for example, the proportion of women in a managing director position is only 7.6% (Computer Sweden, 2012). The gender segregation on the labour market reflects a societal gender order that is related to the structure of power between the sexes. The gender difference in access to power also explains the difference between women’s and men’s opportunities in organizations (Wahl et al., 2001).

The different careers and prevailing selection processes over time have implica-tions for the economic rewards of higher education for the individual (Johansson and Katz, 2007; Löfström, 2009a). The gender order in the labour market can be seen as an obstacle to the full use of women’s resources on societal, organizational and indi-vidual levels. It has consequences for women’s economic rewards from gainful employment. The salaries of women in the IT sector are, on average, 6% lower than men’s, which is the same as the labour market in general. The lowest difference in salary between women and men is to be found at the IT management level where women earn 99% of men’s salary (Tillväxtanalys, 2012). An unequal and discrimina-tory lower income decreases a woman’s consumption level and the capacity to sup-port herself and her family. And it also means a loss of power and influence (Löfström, 2009b; Smith and Bettio, 2008). Some studies show that university educated women

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 5

4 Economic and Industrial Democracy 0(0)

have a better chance of career advancement (Crompton, 2006; Crompton et al., 2007), while other studies point in the opposite direction (Goldin and Katz, 2008), where educated women choose not to follow but to challenge the traditional career logic rules (Halrynjo, 2010).

Segregation processes at organization level

The linking of individual and career takes place at the organization level. The matching process indicates labour market demand for the qualifications acquired by individuals in the educational system (Bills, 2004). Recruitment strategies are important factors in shaping the gender structure in an organization and central to creating new patterns in the division of labour between the sexes (Acker, 2011; Gonäs, 2005). Network recruitment is a common form of recruiting. A Swedish study in the region of Norrköping at the beginning of the 21st century showed that this form of recruiting in the IT sector was manifested in informal networks and that women do not have access to these to the same extent as men (Knocke et al., 2003 : 32).

Career opportunities and terms of employment for women with Master of Science degrees in Engineering have been described by other researchers (see Berner, 2003; Hertzberg, 1989; Kvande and Rasmussen, 1993; Wahl, 2003). They have shown that women have lower salaries, different job tasks and worse career prospects than men. According to Wahl (2003), this situation is due to gender segregation on struc-tural grounds, such as the distribution of women and men in the organization, degree of segregation in relation to positions and tasks, and the hierarchical level entailing differences in influence and power. Kvande and Rasmussen (1993) show that women with Master of Science degrees in Engineering are interested in making a career and to shoulder management responsibility, but that they are not attracted to strict hierarchies and would prefer to manage network types of organizations, in which they would have better promotion prospects. A Finnish study (Hertzberg, 1989) shows that women with Master of Science degrees in the engineering profes-sions thought that they had the same opportunities as men for upward advancement in a hierarchy, but that they held specialist positions, which meant that they advanced horizontally to positions without opportunities for further advancement. In a report from the Office of the Equal Opportunities Ombudsman (JÄMO, 2002) on the IT field in the period 1999–2000, women technology graduates are overrepresented in administrative positions in IT organizations. In a study of 31 IT consultants and managing directors in Swedish IT consultancy businesses, Peterson (2007) points to a horizontal division of labour when it comes to jobs, positions and tasks between women and men. Men dominate in technological core areas such as programming and systems architecture, which are prestigious areas in the organization and offer the best career prospects. Women tend to hold jobs and perform tasks with lower technological status, for instance, developing user-friendly applications and inter-faces, or have ‘soft’ administrative roles, which are undervalued and lead away from the technological core area and career advancement. International studies indicate similar career patterns. Guerrier et al. (2009) show in their study of 28

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 6

Holth et al. 5

managing directors in IT-related employment in Great Britain that technically quali-fied women tend to move away from the core area and end up in internal roles with a focus on the organization and the team. Men hold prestigious and outreaching positions more often, dealing with customers to a greater extent. These isolated qualitative studies do not constitute a basis for generalizing about the labour mar-ket, but they reflect empirical patterns and exemplify horizontal segregation in IT consultancy organizations. A large number of graduate engineers find employment as IT consultants (Computer Sweden, 2012).

Following women and men in the same profession at workplace level provides the opportunity to study the ‘sliding gender segregation’ in the workplace. This concept captures the combined processes that reproduce gender division. Even though women and men have the same education and job tasks in the workplace at the beginning of their careers, women gradually lag behind with time (Holt et al., 2006). There is reason to believe that the sliding workplace segregation is also formed in the intersection between work and family (Halrynjo, 2010). Interviews with employers on recruiting strategies in the IT sector display conditional opportunities for promotion in the form of attendance and accessibility and demands for high performance irrespective of indi-vidual circumstances in other spheres of life (Gonäs and Rosenberg, 2012). Parenthood generally affects women’s career prospects to a higher degree than men’s (see Crompton, 2006). Acker (2011) uses the concepts of the encumbered and unencumbered employee, seeing the woman as the encumbered and the man as the unencumbered in relation to family responsibilities. Holth and Mellström (2011), however, found in their compara-tive interview studies of computer and mechanical engineers in 1995 and 2009 respec-tively that a redirection had taken place towards changed and more equal gender relations. This is evident in men’s increasing focus on fatherhood and parental respon-sibility, and in the lesser importance ascribed to career-making, which indicates that changes take place in the intersection between traditional structures and individual action.

Aim and research questions

On the basis of quantitative data, this article aims to provide insights into the transition from education to occupation for the two groups of new Bachelor and Master of Science graduates in IT engineering and into their employment patterns and career movements over time. The aim is also to increase knowledge of the long-term effects of efforts to make the IT programme meet labour market needs in the field. The reason for choosing IT engineering graduates was that we wanted to study how the male dominance in engi-neering education is reflected in the employment patterns of women and men. Our research questions are: To what extent do newly graduated IT engineers work in a job or a position that corresponds to their graduate level and how does this relation change over time? What is the impact of educational level, age, parenthood, industrial sector and gender structure of the organization for the individual’s employment opportunities? This approach involves studying individual career paths in terms of occupation and position.

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 7

6 Economic and Industrial Democracy 0(0)

Method and material

Our study is a comprehensive investigation of all the individuals who graduated with a Bachelor or a Master of Science Engineering degree in the technological areas of electri-cal engineering, physics engineering and computer science in Sweden during the period 2001–2007 (including operating technicians).

We have used multilevel analysis to determine the significance of certain explanatory variables for reaching a job position that requires the level of qualifications that these two educational groups have. Multilevel modelling is suitable when data are hierarchi-cally structured, i.e. when they consist of units (e.g. individuals or measurements) that can be grouped at different levels in a hierarchy (Gelman and Hill, 2007; Hox, 2002). This may mean that individual results are more correlated within a certain group than individual results between different groups. By applying multilevel analysis, we can account for and quantify such group effects. The data consist of repeated surveys in which each individual has one or more measurements.

Data sources

The study is based on uniquely designed data material assembled via Statistics Sweden (SCB) and its MONA system.1 The material derives from three sources: the LISA data-base (Longitudinal Integration Database for Health Insurance and Labour Market Studies) provided by SCB, the record of graduates at the Swedish Agency for Higher Education2 and the Register-Based Labour Market Statistics (RAMS).3 The LISA data-base makes it possible to follow an individual’s transition between employment posi-tions. It contains demographic background variables, education, occupation, income and sick leave. Employed individuals are linked to data for the industry and companies – more specifically, the workplace. The material covers the period 2001–2007. This data-base allows us to trace the transition from education to the labour market for each student cohort. The two groups we are studying will be referred to as MSEs and BSEs in IT. All in all, the dataset consisted of 4116 individuals, of whom 3662 are included in our analy-sis. Exclusions (n = 454) were due to incomplete data.

Statistical analysis

The statistical analysis is a multilevel multinominial logistic regression analysis where we have calculated the odds ratio for the correlation between position and the explana-tory variables (fixed effects) gender, time, education, age, parenthood, industry sector and workplace gender distribution (qualindex) (see Appendix1).

The dependent variable (position) is a categorical variable with more than two out-comes. Gender, education, age and parenthood have two categories and the rest have been grouped into three categories. Odds ratio in this analysis is interpreted in relation to the reference categories for the respective variable. The reference category for gender is ‘woman’, i.e. the category ‘man’ has a higher chance if the odds ratio is higher than 1 and lower if the odds ratio is lower than 1. The MSE category (55E) is the reference category for education. With regard to the explanatory variables with three categories, the odds

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 8

Holth et al. 7

ratio is interpreted for two of these in relation to the reference category. For the depend-ent variable position we have here chosen to present the results from two equations (models) or transitions. The first refers to jobs with no education requirements versus short-cycle education (levels 1–2) and the second to jobs with no education requirements versus long-cycle education (levels 1–3).

To investigate the impact of hierarchical structure compared to a one-level structure, we calculated the so-called ‘intraclass correlation’ (ICC) (Hox, 2002). ICC indicates the extent to which the variation refers to the second level, i.e. measuring over time. The ICC for the whole material amounts to 42%. The ICC for data referring to men is 43%–38%. All these ICCs are sufficiently high to confirm the importance of multilevel analysis in this study.4

Results

The rate of women graduates from university IT programmes is low: 18% of MSE gradu-ates and 20% of BSE graduates are women. Table 1 shows that the rate of employment obtained in the year of graduation is considerably higher for MSEs than BSEs in our material.

We have chosen to include only those who were employed at some point during the survey period and for whom we have information about occupation, the workplace industry sector and the gender distribution among those who are university educated in the organization. This means that a total of 3662 individuals are included in the following results. Table 2 shows the distribution of the explanatory variables for the individuals who graduated at some point in the period 2001–2007. The results are presented for women, men and the total. This is followed by Table 3 and the presentations of the results from the different models.

A slightly higher percentage of women than of men have BSE qualifications. There is also a difference in age. A higher percentage of women were 35 years and older when first employed after graduating. Likewise, the family situation differs: 17.2% of the women compared to 13.0% of the men had one or more children when embarking on their careers. Position level 1 refers to occupations that do not require higher education, level 2 includes occupations for which short-cycle higher education is required and level 3 are occupations that require long-cycle higher education. There is a clear difference between women and men regarding distribution at the level of occupational position. Women had to a greater extent obtained employment at the lowest level, at which no academic qualifications are required, as well as at the intermediate level with some degree of academic training required. Of the men 60.7% had obtained employment in an

Table 1. Percentage of women and men MSE and BSE graduates in IT engineering who had employment in the year of graduation.

Employment Women Men Total

MSE in IT 88.4 86.9 87.2

BSE in IT 72.5 72.0 72.1

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 9

8 Economic and Industrial Democracy 0(0)T

able

2.

Cha

ract

eris

tics

of t

he s

tudy

pop

ulat

ion

(N =

366

2) b

y ed

ucat

ion,

age

, fam

ily/c

hild

ren,

pos

ition

, ind

ustr

y se

ctor

and

wor

kpla

ce g

ende

r di

stri

butio

n fo

r w

omen

, men

and

tot

al; 9

5% c

onfid

ence

inte

rval

s ar

e sh

own

in p

aren

thes

es.

Var

iabl

eW

omen

Men

Tot

al

N

o.%

No.

%N

o.%

Educ

atio

nM

SE in

IT40

155

.7 (

52.1

,59.

3)17

0257

.9 (

56.1

,59.

7)21

0357

.4 (

55.8

,59.

0)

BSE

in IT

319

44.3

(40

.7,4

7.9)

1240

42.1

(40

.3,4

3.9)

1559

42.6

(41

.0,4

4.2)

Age

20–3

461

885

.8 (

83.2

,88.

3)27

0591

.9 (

90.9

,92.

9)33

2390

.7 (

89.8

,91.

6)

35+

102

14.2

(11

.7,1

6.7)

237

8.1

(7.1

,9.1

)33

99.

3 (8

.4,1

0.2)

Fam

ily/c

hild

ren

No

child

ren

596

82.8

(80

.0,8

5.6)

2560

87.0

(85

.8,8

8.2)

3156

86.2

(85

.1,8

7.3)

C

hild

ren

124

17.2

(14

.4,2

0.0)

382

13.0

(11

.8,1

4.2)

506

13.8

(12

.7,1

4.9)

Posi

tion

1 (lo

wes

t)18

726

.0 (

22.8

,29.

2)61

320

.8 (

19.3

,22.

3)80

021

.9 (

20.6

,23.

2)(S

syk3

)2

149

20.7

(17

.7,2

3.7)

544

18.5

(17

.1,2

0.0)

693

18.9

(17

.6,2

0.2)

3

384

53.3

(49

.7,5

6.9)

1785

60.7

(58

.9,6

2.5)

2169

59.2

(57

.6,6

0.8)

Indu

stry

sec

tor

Uns

peci

fied

20.

3 (–

0.1,

0.7)

20.

1 (0

.0,0

.2)

40.

1 (0

.0,0

.2)

A

grar

ian/

fore

stry

30.

4 (–

0.1,

0.9)

20.

1 (0

.0,0

.2)

50.

1 (0

.0,0

.2)

M

anuf

actu

ring

139

19.3

(16

.4,2

2.2)

554

18.8

(17

.4,2

0.2)

693

18.9

(17

.6,2

0.2)

En

ergy

, wat

er9

1.2

(0.4

,2.0

)29

1.0

(0.6

,1.4

)38

1.0

(0.7

,1.3

)

Con

stru

ctio

n10

1.4

(0.5

,2.3

)34

1.2

(0.8

,1.6

)44

1.2

(0.8

,1.6

)

Tra

de a

nd c

omm

erce

669.

2 (7

.1,1

1.3)

311

10.6

(9.

5,11

.7)

377

10.3

(9.

3,11

.3)

Fi

nanc

ial

322

44.7

(41

.1,4

8.3)

1506

51.2

(49

.4,5

3.0)

1828

49.9

(48

.3,5

1.5)

Ed

ucat

ion/

rese

arch

8011

.1 (

8.8,

13.4

)33

711

.5 (

10.3

,12.

7)41

711

.4 (

10.4

,12.

4)

Hea

lth/s

ocia

l car

e21

2.9

(1.7

,4.1

)33

1.1

(0.7

,1.5

)54

1.5

(1.1

,1.9

)

Pers

onal

and

cul

tura

l ser

vice

s20

2.8

(1.6

,4.0

)53

1.8

(1.3

,2.3

)73

2.0

(1.5

,2.5

)

Publ

ic a

genc

y48

6.7

(4.9

,8.5

)80

2.7

(2.1

,3.3

)12

83.

5 (2

.9,4

.1)

Wor

kpla

ce g

ende

r di

stri

butio

n (q

ualin

dex)

M

ale

dom

inat

ed31

443

.6 (

40.0

,47.

2)15

5152

.7 (

50.9

,54.

5)18

6550

.9 (

49.3

,52.

5)Ev

en d

istr

ibut

ion

152

21.1

(18

.1,2

4.1)

464

15.8

(14

.5,1

7.1)

616

16.8

(15

.6,1

8.0)

Fe

mal

e do

min

ated

254

35.3

(31

.8,3

8.8)

927

31.5

(29

.8,3

3.2)

1181

32.3

(30

.8,3

3.8)

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 10

Holth et al. 9

occupation requiring long academic specialist training; the corresponding figure for women was 53.3%. The fact that over a quarter of the women with BSE in IT or MSE in IT engineering obtained their first job in an occupation that does not require any aca-demic qualifications brings the debate on over-education to the fore. The same applies to a fifth of all the men (see Le Grand et al., 2005). The industrial affiliation of the work-places indicates certain differences between women and men. It is noteworthy that both women (44.7%) and men (51.2%) obtained jobs in the financial sector. The manufactur-ing industry has only attracted 18.9% of the two engineering graduate groups, a similar pattern for both women and men. This reflects the fact that these academic categories are in demand far beyond the manufacturing industry in a traditional sense. Even if they usu-ally move into a limited number of occupational areas, there is a demand for systems and programming graduates in many different labour market sectors (Arbetsförmedlingen, 2012; SCB, 2010).

Development over time

In the following, the results of the development over time are presented in terms of occu-pation and position careers. The results of the previous section indicate that many start out in occupations with low educational requirements. The next step is to investigate whether the first employment position determines the future career or if there is an upward mobility in terms of position levels over time and in what circumstances this takes place.

The results are presented in three sections: results of the entire data material, of data on women and of data on men. Table 3 presents the results in the form of coefficients for the various explanatory variables (fixed effects). Each variable has two coefficients since the material comprises two models. We start by presenting the results of the whole mate-rial followed by the gender specific analysis.

Table 3 shows that there is a 31% greater chance for men of changing the position to 3 with respect to (w.r.t.) 1 than there is for women. The chance of remaining in a low-qualification job was considerably higher for women than men. One explanation might be that women and men are employed in different types of positions, which in turn have different career prospects in IT organizations (see Guerrier et al., 2009; Peterson, 2007).

Time is a significant factor in both models. The chance increased 8% per year for such career moves.

Educational level is of significant importance to the chance of moving to the highest position with regard to lowest position. The odds ratio of getting to position 3 rather than position 1 is 0.26, indicating that the chance for a person with a BSE in IT is 74% (1–0.26) lower than for a person with an MSE in IT. Education does not play the same role for the transition from the lowest to the intermediate level (level 2 w.r.t. 1).

Having children had a positive influence on the chance of changing to intermediate w.r.t. the lowest positions and between the lowest and the highest. The chance of making one of these changes is 30% higher for people with children than for those without chil-dren. Generally speaking, we can say that there are significant correlations between par-enthood and the individual’s situation on the labour market. Being available and present

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 11

10 Economic and Industrial Democracy 0(0)

Table 3. Results from the multilevel regression analysis for transition between occupational position level 2 with regard to (w.r.t.) 1, and level 3 with regard to 1 and the explanatory variables (fixed effects) for the total material (T), women (W) and Men (M) in terms of odds ratio.

Estimated parameters, level 2 w.r.t. 1

Estimated parameters, level 3 w.r.t. 1

Odds ratio, level 2 w.r.t. 1

Odds ratio, level 3 w.r.t. 1

Explanatory variables (fixed effects)Gender T 0.005 (0.094) 0.268* (0.082) 1.00 1.31 W – – – – M – – – –Time T 0.081* (0.017) 0.078* (0.015) 1.08 1.08 W 0.140* (0.036) 0.089* (0.032) 1.15 1.09 M 0.070* (0.020) 0.079* (0.017) 1.07 1.08Education T –0.414* (0.078) –1.339* (0.067) 0.66 0.26 W –0.190 (0.172) –1.190* (0.155) 0.83 0.30 M –0.452* (0.087) –1.369* (0.075) 0.64 0.25Qualindex2 T Ref. Ref. Ref. Ref. W Ref. Ref. Ref. Ref. M Ref. Ref. Ref. Ref.Qualindex1 T 0.296* (0.085) 0.535* (0.073) 1.34 1.71 W 0.201 (0.167) 0.571* (0.147) 1.22 1.77 M 0.327* (0.099) 0.540* (0.084) 1.39 1.72Qualindex3 T –0.311* (0.102) –0.479* (0.087) 0.73 0.62 W –0.402* (0.186) –0.685* (0.165) 0.67 0.50 M –0.273** (0.122) –0.391* (0.102) 0.76 0.68Children T 0.266* (0.084) 0.274* (0.074) 1.30 1.31 W 0.252 (0.171) 0.342** (0.154) 1.29 1.41 M 0.263* (0.097) 0.255* (0.084) 1.30 1.29Sector1 T Ref. Ref. Ref. Ref. W Ref. Ref. Ref. Ref. M Ref. Ref. Ref. Ref.Sector2 T –0.355* (0.079) 0.607* (0.070) 0.70 1.83 W –0.486* (0.171) 0.260** (0.155) 0.62 1.30 M –0.319* (0.089) 0.699* (0.079) 0.73 2.01Sector3 T 0.112 (0.107) 1.246* (0.094) 1.12 3.48 W 0.028 (0.206) 0.673* (0.188) 1.03 1.96 M 0.141 (0.125) 1.436* (0.109) 1.15 4.20Age1 T Ref. Ref. Ref. Ref. W Ref. Ref. Ref. Ref. M Ref. Ref. Ref. Ref.Age2 T 0.138 (0.105) 0.007 (0.093) 1.15 1.00 W –0.064 (0.198) 0.031 (0.181) 0.94 1.03 M 0.192 (0.124) 0.002 (0.190) 1.21 1.00

The standard errors are in parentheses. What is significant in estimated parameters is also significant in the odds ratio.**significance level at 5%, *significance level at 1%.

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 12

Holth et al. 11

in the workplace is an important part of the career logic and affects the chances for pro-motion in an organization (Halrynjo, 2010). We discuss this further in the gender specific analysis section.

Regarding age, there are no significant results in this analysis. As shown in Table 2, nearly the whole population is in the age group 20–34 years.

The odds for reaching the highest position level increases greatly when we move from industry to the service sectors, private or public. In the public sector the odds ratio was 3.48 of changing to position 3 w.r.t. position 1. In summary, there was a greater chance that changes were made upwards in the hierarchy in the public service sector, such as education and research, the public sector or personal service (industry sector 3) than in the manufacturing industry. One explanation could be the regulations on career planning and salary mapping in public agencies (see Gonäs, 2005).

For the variable workplace gender segregation (qualindex), we have specified even gender distribution (qualindex2) as the reference category. The results from the model estimating level 3 with relation to 1 shows an odds ratio of 1.71 for qualindex1 (= male dominance). It suggests a much higher chance for people in an organization with male domination among the higher educated to make this transition, than in an evenly gender distributed setting. For people in organizations with a female dominance among the highly educated, the chance of changing to a higher position was considerably lower. It was easier to move upwards to occupations with requirements corresponding to the indi-vidual’s qualifications in organizations with a male dominance among the highly educated.

A gender specific analysis

The greatest difference between women and men is found in the higher chance of making an upward change at all with relation to a lower level. Gender plays a decisive role for promotion or changing to an occupation matching qualifications. MSE qualifications are important for both women and men in changing to the highest position with regard to the lowest position.

Parenthood is a significant factor for men in the transition from the lowest to the intermediate position, and from the lowest to the highest position, whereas parenthood for women was only significant in the changing to the highest position w.r.t. the lowest position. From quantitative as well as qualitative studies we know that men’s career development takes off in connection with starting a family. Granqvist and Persson (2004) show in their study that women with children have around 50% less chance of switching to a better job than men with children. The study was based on material from the Swedish Level of Living Survey (LNU) covering the period 1950–1991. Around 3800 randomly selected persons representing the Swedish population were interviewed in 1991 and asked to recount their occupational history. The difference between the sexes among the highly qualified was somewhat lower (40%) than for the survey popu-lation as a whole (see Halrynjo, 2010). The claim that differences in career prospects are related to the uneven distribution of unpaid household work was supported by that study, which showed that the duration of parental leave affected women’s career oppor-tunities negatively.

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 13

12 Economic and Industrial Democracy 0(0)

The greatest chance of changing to the highest position w.r.t. the lowest position for both women and men was in the public service sector. In regard to the gender distribution among the highly qualified (qualindex) in an organization, the chance of promotion to the highest qualification level is significantly higher for women within an organization with male dominance. This could be explained by the fact that male dominated organiza-tions have a network character, which means that the hierarchical levels are less rigid (Kvande and Rasmussen, 1993).

Discussion

Higher education in the area of IT does not contribute to changing the labour market gender segregation. A causal explanation is that women remain a minority group in the field and we know that gender stereotyped educational choices increase labour market segregation and impede matching (SOU, 2004). The key results of the study show that there is still a big difference between women’s and men’s chances of obtaining employ-ment in an occupation or position that matches their qualification level in the IT sector. Women and men find employment to the same degree after graduating but here end the similarities in terms of career opportunities. Men reach higher position levels to a greater extent than women. The gender segregation and the successive selection process for recruitment to higher positions have developed over time at the level of the organization, thus contributing to shaping women’s and men’s different career patterns. The systematic underutilization of women’s technical competencies becomes a form of structural dis-crimination. There is a risk that the problem will be individualized instead of treated as an organizational failure to take full advantage of women’s qualifications and competen-cies. This underutilization also runs the risk of reproducing and reinforcing perceptions of the strong link between men and technology in general and men and computer science in particular.

The results indicate significant differences in upward mobility to the position levels matching qualification levels. It is less likely for women than for men that they will advance to the hierarchical level that corresponds to their qualification level. In IT organizations this is assumed to be related to the fact that women end up far from the technological core area and as a consequence they are not promoted (Guerrier et al., 2009; Peterson, 2007). This situation reveals the close connection between horizontal segregation and the processes generating the vertical segrega-tion. Women are to a greater extent than men in occupations requiring low or no academic qualifications. Men who are in low-qualification jobs have a better chance of leaving to enter more qualified occupations, whereas women tend to be ‘locked’ into these occupations. There may be many reasons for being locked into a low-qualification job. The regional labour market structure can be one. Employers’ reluc-tance to hire women for high positions can be another. Swedish as well as international studies point to the latter circumstances as restricting factors for women’s opportuni-ties to capitalize on their technological education (Gonäs and Rosenberg, 2012; Guerrier et al., 2009).

Parenthood increases men’s chance of upward mobility more than women’s, which supports previous research and the perception of women and men as potential workforce

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 14

Holth et al. 13

members. Acker’s (2011) discussion of men as the unencumbered employee can serve as an explanation for this pattern. The man is the norm and being a man involves having a family, which, from the employer’s perspective, means that he is to be seen as a stable employee, which in turn reflects the obsolete principle of the male breadwinner. The importance of working time is not possible to establish in this study, but it is likely to affect the role of parenthood and calls for further study.

Male dominated organizations provide the best opportunities for women and public service agencies give both women and men opportunities to obtain positions relevant to their qualifications to a greater extent than other sectors. These results need to be studied further. Regarding public service agencies, the gender-neutral recruitment strategies instead of network recruitment can be an explanation (see Knocke et al., 2003). According to Knocke et al., public agencies use formal recruitment strategies while IT organizations apply network recruitment and selection criteria such as how well the applicant would ‘fit’ into the environment, strategies that, as Acker (2011) points out, leave room for reproducing existing gender structures.

The long-term reconstruction of the labour market towards a higher degree of service production affects the occupational opportunities for IT engineers (Berner, 2003). Their entry into the financial sector in our material supports this conclusion. Our analysis shows the high possibilities of doing an upward career move in line with qualifications in the public service sector. Concurrently, there are still embarrassing differences between women and men in other sectors. Instead of referring to the concept over-education, it would be more reasonable to ask, from a resource perspective, how women’s qualifica-tions could be put to better uses and developed in the ICT field as a whole, particularly in view of the anticipated future shortage of qualified people in the field and the fact that even today computer experts are recruited from degree programmes other than the IT-related ones.

Summary

An important scientific contribution in this article is the study design as such. Data are longitudinal with repeated data collection for all individuals and we have used a multi-level approach when determining the significance of certain variables in the analysis. Further, our data contain all individuals that have been awarded a degree in our two selected educational groups during the specific time period. We studied the positions that the individuals had received during their entrance period into the labour market. There was a gender difference in the possibilities of reaching a job that requested the educa-tional qualifications that the single individual had. Men had reached these levels to a greater extent than women. In order to deepen the organizational perspective we also designed a variable that measures the workplace gender distribution. When it comes to the impact of gender structure in the organization our finding shows that technology jobs and male domination in organizations are still strongly related.

Acknowledgements

We would like to thank Dan Larsson at the department of Statistics at Karlstad University for his work on the database and the construction of the data files.

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 15

14 Economic and Industrial Democracy 0(0)

Funding

The article was written as part of the major research project JämVäxt 2.0 [Equal Growth 2.0] and centres on the transition from education to work. The project is funded by the Swedish Agency for Economic and Regional Growth (European Regional Development Funds), Region Värmland and Karlstad University.

Notes

1. SCB has developed a system for external access to the micro database MONA (Mikrodatorer Online Access). MONA provides access to the databases. SCB is responsible for the design and confidentiality of the data volume. The present study has been approved by the ethics committee.

2. The record of the Swedish Agency for Higher Education consists of several sections of accu-mulated data for the academic years 2000/2001–2006/2007. The record covers individuals enrolled in first- and second-cycle education. It provides information on academic year, higher education institution, degree title and degree area.

3. RAMS contains information on the regional labour market, workplace and employee residence.

4. The analysis was carried out with MLWin 2.10.

References

Åberg R (2002) Överutbildning – ett arbetsmarknadspolitiskt problem? In: Abrahamsson K, Abrahamsson L, Björkman T et al. (eds) Utbildning, kompetens och arbete. Lund: Studentlitteratur.

Acker J (2011) Why such persistent gender gaps in working life? A short history of theorizing gen-der inequality in work. In: Symposium on Conditions for Egalitarian and Sustainable Growth, Karlstad University, 30 September 2011.

Arbetsförmedlingen (2012) Var finns jobben? Bedömning för 2012 och första halvåret 2013. Stockholm: Arbetsförmedlingen URA 2012:4

Berggren C (2007) Arbetsmarknaden och efterfrågan på högre utbildning – ett köns- och klasspers-pektiv. In: Olofsson J (ed.) Utbildningsvägen – vart leder den? Om ungdomar, yrkesutbildning och försörjning. Stockholm: SNS Förlag, pp. 55–69.

Berggren C (2008) Matchande jobb för akademiker? Arbetsmarknad & Arbetsliv 14(1): 65–75.Berner B (2002) Kunskapssamhällets paradoxer; Utbildningens betydelser och könsrelationernas

roll. In: Sturesson L et al. (eds) Spänningsfält; Tekniken, Politiken Framtiden. Stockholm: Carlssons Bokförlag.

Berner B (2003) Kvinnor i ingenjörsarbete. Kön, makt och ingenjörskulturer. In: Berner B (ed.) Vem tillhör tekniken? Kunskap och kön i teknikens värld. Lund: Arkiv Förlag, pp. 175–200.

Bettio F and Verashchagina A (2009) Gender Segregation in the Labour Market: Root Causes, Implications and Policy Responses in the EU. EU Expert Group on Gender and Employment (EGGE). Luxembourg: European Commission.

Bihagen E and Ohls M (2006) The glass-ceiling – Where is it? Women’s and men’s career pros-pects in the private vs the public sector in Sweden 1979–2000. The Sociological Review 54(1): 20–47.

Bills DB (2004) The Sociology of Education and Work. Malden, MA: Blackwell.Cockburn C (1983) Brothers: male dominance and technological. London: Pluto.Computer Sweden (2012) Available at: computersweden.idg.se/2.2683/1.433629/ensamt-pa-top-

pen-for-kvinnliga-vdar (accessed 1 March 2012).

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 16

Holth et al. 15

Crompton R (2006) Employment, flexible working and the family. In: Gonäs L and Karlsson JC (eds) Gender Segregation: Divisons of Work in Post-Industrial Welfare States. London: Ashgate.

Crompton R, Lewis S and Lyonette C (eds) (2007) Women, Men, Work and Family in Europe. London: Palgrave Macmillan.

European Commission (2009) Women and ICT Status Report 2009. Brussels: European Commission, Information Society and Media.

Faulkner W (2000) The power and the pleasure? A research agenda for ‘making gender stick’ to engineers. Science, Technology, and Human Values 25(1): 87–119.

Gelman A and Hill J (2007) Data Analysis Using Regression and Multilevel/Hierarchical Models. New York: Cambridge.

Goldin C and Katz L (2008) The Race between Education and Technology. Cambridge, MA: Harvard University Press.

Gonäs L (ed.) (2005) På gränsen till genombrott? Om det könsuppdelade arbetslivet. Stockholm: Agora Förlag

Gonäs L and Rosenberg K (2012) Conditional equality – When the higher educated meet the labour market. In: Jansdotter Samuelsson M, Krekula C and Åberg M (eds) Gender and Change. Karlstad: Karlstad University Press, pp. 95–111.

Granqvist L and Persson H (2004) Kvinnors och mäns karriärvägar på den svenska arbets-marknaden. Bilaga 2 till SOU 2004: 43. Den könsuppdelade arbetsmarknaden. Fritzes: Stockholm.

Guerrier Y, Evans C, Glover J and Wilson C (2009) Technical, but not very … : Constructing gendered identities in IT-related employment. Work, Employment and Society 23(3): 495–511.

Hacker S (1989) Pleasure, Power and Technology. Boston, MA: Unwin Hyman.Halrynjo S (2010) Mödre og fedre i möte med karrierelogikkens spilleregler. Institutt for sosiologi

og samfunnsgeografi. Oslo: Oslo universitet.Hebson G and Rubery J (2004) Könssegregering – Förändring och kontinuitet. Bilaga 1 i SOU

2004:43. Den könsuppdelade arbetsmarknaden. Stockholm: Fritzes.Hertzberg V (1989) Kvinnliga diplomingenjörers väg från studier till yrkesliv. In: Silius H (ed.)

Kvinnor i mansdominerade yrken. Åbo: Institutet för kvinnoforskning vid Åbo Akademi.Holt H, Gerdsen LP, Christensen G et al. (eds) (2006) Det kjønsopdelte arbejdsmarked. En kvan-

titativ och kvalitativ belysning. Köpenhamn: Socialforskningsinstituttet.Holter Ø G, Svare H and Egeland C (2009) Gender Equality and Quality of Life: A Norwegian

Perspective. Oslo: University of Oslo.Holth L and Mellström U (2011) Revisiting engineering, masculinity and technology studies: Old

structures with new openings. International Journal of Gender, Science and Technology 3(2): 313–329.

HSV (2008) Women and Men in Higher Education. Report 2008 :48R. Stockholm: The Swedish National Agency for Higher Education.

Hox J (2002) Multilevel Analysis: Techniques and Applications. Mahwah, NJ: Lawrence Erlbaum.JÄMO (Jämställdhetsombudsmannen. JämO:s IT-granskning) (2002) Utvärdering av 1999–2000

års granskning. StockholmJohansson M and Katz K (2007) Wage Differences between Women and Men in Sweden: The

Impact of Skill Mismatch. Working paper 2007:13. Uppsala: IFAU.Kivinen O, Hedman J and Kaipainen P (2007) From elite university to mass higher education:

Equality of opportunity and returns to university education. Acta Sociologica 50: 231–247.Knocke W, Drejhammar I-B, Gonäs L and Isaksson K (2003) Retorik och praktik i rekrytering-

sprocessen. Arbetsliv i omvandling 2003:4. Stockholm: Arbetslivsinstitutet.

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 17

16 Economic and Industrial Democracy 0(0)

Kvande E and Rasmussen B (1993) Nye kvinneliv: kvinner i menns organisasjoner Oslo: Ad Notam Gyldengal.

Lagesen VA (2005) Extreme Make-over? The Making of Gender and Computer Science. Doktorsavhandling. STS-Rapport 71. Trondheim: NTNU.

Le Grand C, Szulkin R and Tåhlin M (2005) Education and Inequality in Sweden: A Literature Review. Sociologiska institutionen. Stockholm: Stockholms universitet.

Löfström Å (2009a) Gender Equality: Economic Growth and Employment. Report. Umeå: Umeå University, Department of Economics.

Löfström Å (2009b) Utbildning – ett sätt för kvinnor at närma sig männens löner? In: Egerö A-M (ed.) Nio perspektiv på jämställdhet – en antologi om kvinnor och mäns löner. Stockholm: Medlingsinstitutet.

Mellström U (1995) Engineering Lives, Technology, Time and Space in a Male-centred World. Studies in Art and Science, nr 128. Linköping: Linköpings universitet.

Mellström U (2009) Är IT manligt? Teknik och genus i tid och rum. In: Karlsohn T (ed.) Samhälle, teknik och lärande. Stockholm: Carlssons bokförlag, pp. 66–85.

OECD (2009) Education at a glance 2009: OECD indicators. Available at: www.oecd.org/edu/eag2009.Peterson H (2007) Gendered work ideals in Swedish IT firms: Valued and not valued workers.

Gender, Work and Organisation 14(4): 333–348.SCB (2010) Yrkesspridning och utbytbarhet på arbetsmarknaden 2010:4. Tema Arbetsmarknad.

Stockholm: Statistiska Centralbyrån.Smith M and Bettio F (2008) Analysis note: The economic case for gender equality. European

Commission, Directorate General for Employment, Social Affairs and Equal Opportunities. August 2008.

SOU (2004) Den könsuppdelade arbetsmarknaden. Betänkande av utredningen om den könsseg-regerade svenska arbetsmarknaden 2004:43. Stockholm: Fritzes.

Tillväxtanalys (2012) Uppföljning av handlingsplanen Jämställd IT-utveckling för ökad tillväxt. Working paper/PM 2012:05. Östersund: Tillväxtanalys.

Wahl A (2003) Könsstrukturer i organisationer – kvinnliga civilekonomers och civilingenjörers karriärsutveckling. 2:a upplagan. Lund: Studentlitteratur.

Wahl A, Holgersson C, Höök P and Linghag S (2001) Det ordnar sig. Teorier om organisation och kön. Lund: Studentlitteratur

Appendix 1: Operationalization of variables from the LISA database

The highest completed education, Sun2000Grp;

55E = Master of Science in Engineering with a major in electrical engineering, physics engineer-ing, or computer science, MSE in IT55J = Bachelor of Science in Engineering with a major in electrical engineering, physics engineer-ing, or computer science, BSE in ITGender: 0 = woman, 1 = manChildren: 0 = no children, 1 = one or several childrenAge groups: 1 = 20–34, 2 = 35+

Swedish occupational classification according to Ssyk 3,Position:

1 = Ssyk3 400≤999, occupations with low or no qualification requirements2 = Ssyk3 300≥399, occupation with short-cycle academic qualification (engineers, technicians)3 = Ssyk3 ≤299, occupations with long-cycle specialist qualification requirements or management experience

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from

Page 18

Holth et al. 17

Employment status:

1 = employed, 5 = not employed but with statement of income for the year, 6 = not employed and no statement of income for the yearSegregation quota based on the proportion of women and men in the organization.SegQuot = (Org_KvEGymnL)/(Org_MenEGymnL+Org_KvEGymnL).

Workplace segregation index:

Qualindex: 1 = SegQuot <= 0,4; 2 = SegQuot >= 0,6; 3 = 0,4 < SegQuot < 0,6Classification of workplace by industry sector AstSNI2002G = Industry sector, broad level, 10 groups

00 = Non-specified operation 06 = Financial & business services01 = Agriculture, forestry & fishing 07 = Education & research02 = Manufacturing & extraction 08 = Health and social care03 = Energy production, water supply & waste disp 09 = Personal & cultural services04 = Construction 10 = Public administration, etc.05= Commerce & communication

Reference categories

Gender Woman = 0, man = 1; Age group 1 = 20–34, 2 = 35+; Education MSE in IT: Sun 2000grp 55E = 1, BSE in IT: Sun 2000grp 55J = 2; Children 0 = no children, 1 = one or several children; Workplace gender distribution = Qualindex 1 = male dominance among the highly qualified in the organization, 2 = even gender distribution among the highly qualified in the organization, 3 = female dominance among the highly qualified in the organization; Industry sector 1 = SNI 20002 G 0–4 (e.g. manufacturing and construc-tion), 2 = SNI 20002 G 5–6 (private services), 3 = SNI 20002 G 7–10 (public services, personal and cultural services); Time year 0, year 1, year 2, year 3, year 4, year 5, year 6 after graduation.

Author biographies

Line Holth is PhD student in working life science at Karlstad University. Her research area is gen-der and work, gender segregation processes, gender and technology with a focus on higher educa-tion and labour market.

Abdullah Almasri is Associate Professor of Statistics at Karlstad University. His research interests focus on time series and wavelet analysis, multilevel analysis and forecasting.

Lena Gonäs is Professor in Working Life Science at Karlstad University and affiliated to the section for Insurance Medicine at Karolinska Institutet in Stockholm. Her research focus has been on gender relations in working life, labour market segregation, structural change and labour market policies.

at Karlstad Universitet on August 23, 2013eid.sagepub.comDownloaded from