Occupations and survey research: methodological and substantive applications on the occupation-inequality link 2 |
|
| Chair | Dr Daniela Rohrbach-Schmidt (German Federal Institute for Vocational Education and Training ) |
| Coordinator 1 | Professor Christian Ebner (University of Cologne) |
In Switzerland vocational education and training (VET) diploma holders may, after some years of work experience, enter a vocationally oriented track of the tertiary education system, called professional education (PE). It accounts for one third of all tertiary-level students and includes Federal and Advanced PE Diplomas and Professional Education colleges. The few existing studies on the transition from VET to PE imply a weak role of socioeconomic family background (Kriesi & Trede, submitted/2016). Moreover women and people with a migration background have a lower probability to hold a PE certificate (SKBF 2014). Finally results of Buchmann et al. (2007; 2016) show that the probability of entering PE depends strongly on the type of the VET programme. This suggests that characteristics of upper-secondary education matter for the chance of entering PE. This contribution therefore examines the role of institutional characteristics of Swiss VET programs in explaining individual transitions from VET to PE. In order to formulate theoretical assumptions we draw on previous comparative research, which has shown that educational transitions depend on the level of standardisation, differentiation and vocational specificity of a countries educational system (Shavit & Müller 1998). Standardisation refers to the extent to which curricula or exams uphold the same nationwide standards. Differentiation taps the sorting of learners into horizontally or vertically different programs or tracks. Vocational specificity refers to the ratio of specific vocational versus general skills. We argue that these dimensions matter for the transition into PE because they affect VET learners’ human capital. Firstly, following Becker (1962) there are two ideal typical kinds of human capital: general or more vocation specific knowledge and skills. The former ease and support further learning throughout the career (Fazekas & Field, 2013), thus facilitating occupational mobility and the transition into further education. We argue that a low level of horizontal differentiation and a low ratio of occupation-specific curriculum content increase the chance of entering PE (Becker 1962). Secondly, quality and standard of upper-secondary education is likely to affect former VET learners’ human capital positively. We argue that quality and standard of education are higher in vocational programs with high vertical differentiation and high exam standardization. The former leads to more homogenous achievement groups. It sorts capable learners into tracks that are a prerequisite for admission and increases therefore their probability of entering PE (Shavit & Müller 1998). Exam standardization has been shown to increase the quality of education and to lead to higher student selection, thus increasing the level of students’ competences and human capital (Effinger & Polborn 1998). We argue that this increases students’ transition chances into PE due to their superior competences and because they face less competition. The analyses are based on the Swiss Labor Force Survey from 1991 – 2015. We use pooled data and a binary logistic regression to compare VET diploma holders with and without a transition to PE within the observation period.
The digitalisation of the economy, alongside with the globalisation of the markets and the demographic changes, is reported by several scholars and the relevant literature to be the one of the most important drivers of change behind the profound transformation of the labour market and the way people work, which is currently on going and which is thought to become even more significant in the years to come.
The digitalisation of the economy is seen in most of the literature as contributing to the polarisation of the labour market in many countries. Digitalisation is thought to have brought on the one hand to a significant increase over the time in the demand of high-skilled individuals, equipped with cognitive skills and technical knowledge to deal with tasks and procedures required by the new technologies, and on the other hand to a sharp decrease in the demand of the medium-skilled and – although to a lesser extent – the low-skilled or non-skilled workers. This will result in an increasing polarisation of remunerations and wealth and will therefore contribute to the increase of inequalities in the population.
In 2016 European Commission carried out the European Digital Skills Survey . The purpose of this survey is to collect information on the digital skills required by enterprises across the European Union. Specific objectives of the survey are to quantify the jobs that require digital skills; to provide evidence on the type of digital technologies use by workplaces; to measure the level/type of digital skills required by different jobs, to identify the main digital skills gaps in different occupational categories and collect information on how employers deal with digital skills gaps and the main bottlenecks/barriers to improved availability of digital skills.
Results from the European Digital Skills Survey seem to confirm trends of polarisation related to the use of digital technologies. Findings display that European employers believe that a certain level of digital skills is required to workers in all the occupations and across a range of sectors. Nevertheless, the analyses clearly indicate that the vast majority of employers require high-skilled employees to possess digital skills. On the other hand, only a limited proportion of workplaces seem to require medium-skilled and low or non-skilled employees to possess digital skills to perform their job tasks. Applying a set of statistical methods to the survey data, our presentation will display how digital technologies are contributing to the polarisation of the labour market and to the increase of inequalities.
In social science, the individual’s occupation is used as a proxy variable for various indicators, e.g. status measures, class schemes or task specialization. Depending on the research question it makes also sense to consider people within an occupation or occupational group as homogenous in regard to the specific indicator of interest. Transformed in a statistical model, this suggests a multi-level design because of the nested structure of individuals within occupational groups. However, independent of the statistical model, it is often assumed for the analysis that the occupation is a constant measure of the interesting social context. Whereas this assumption is certainly weak in cross sectional designs, it can be questionable if the research aims on comparisons over time. This might especially be the case if one’s analyses technological impact on task change with occupational codes as proxies or matching variable for certain activities, as it is known that task changes not only take place between but also within occupational categories (Autor 2013). However, data sets on task developments within occupations are very rare. For Germany until now, only the BIBB-IAB/BIBB-BAuA employment survey has been used to explain task changes over time. In my presentation, I will introduce the German Microcensus (Labour Force Survey) as an alternative data source to measure task developments between and within occupations in Germany since 1973.
I will demonstrate that relevant key variables can be harmonized over 16 cross section surveys from 1973 to 2011 by using weak assumptions. Differentiating by harmonised 10 tasks items, it can be shown that one third of the decline of extraction/manufacturing and production tasks in Germany happened within occupations, whereby low skilled workers were more affected from within change than medium skilled workers.
The strength of the Microcensus based time series lies in a more consistent measurement of tasks over time than in other available data sources. Furthermore, task information can also be obtained for several groups with different sex, age, place of work, nationality or working hours. Due to the provision of a harmonized occupational classification that can be reconstructed with every German classification of occupations it is also possible to match the Microcensus task information to every time series data on occupational level. However, the Microcensus information has also its weakness. Due to the focus on the main activity, only substantial (and no peripheral) changes of task shifts are observed and - maybe more important - the differentiation of the ten harmonized task items is not in line with any known theoretical explanation. To overcome this challenge, I will show how additional occupational information from other data sources like the BIBB-IAB/BIBB-BAuA employment surveys can be used to give the harmonized task items a meaningful interpretation in terms of task biased technological change by calculating interaction effects between the task shares and different clusters of occupations.
Reference: Autor, David H. (2013): The "task approach" to labor markets: an overview. Journal for Labour Market Research (46): 185-199
Categorizing jobs available in modern labor market is a daunting task. One of the tools most widely used for this purpose in survey research is International Standard Classification of Occupations (ISCO), managed by International Labor Organization, and last updated in 2008. This hierarchical classification scheme is utilized in renowned research projects both at international (e.g. European Social Survey, World Value Survey), and national level (e.g. in Poland: Diagnoza Społeczna, Bilans Kapitału Ludzkiego). Data thus obtained are then used to analyze country’s labor market in terms of occupational composition, mobility, interplay between supply and demand for particular kinds of jobs, etc.
In the first part of the paper, I focus on the possible pitfalls of gender comparisons based on the ISCO data. The main difficulty lies in the inherent heterogeneity of occupational categories, readily revealed when higher level categories are split into sets of lower-level categories. In particular, the intensity of job segregation depends on the ISCO level at which the analysis is performed. This also poses a serious difficulty for pay comparisons between the sexes with “type of job under control” as we face inevitable analytical trade-off between the precision of job description and the sample sizes of particular occupations. In the second part, I show how occupational segregation rankings depend on the level of aggregation employed for international comparisons.
To be sure, the problems discussed aren’t specific for ISCO only – other job classification schemes are vulnerable to the same problems as well.