A 20th century version of the occupation multilingual ontology

Abstract

Task 3.2 Selected SSH Ontologies and Vocabularies in SSHOC Work package 3 Lifting Technologies and Services into the SSH Cloud aimed to foster the use of selected global ontologies in the social sciences and humanities, regarding occupational titles, educational categories, sectors of industry, geographical regions, food items, and religions. These ontologies service the usage of vocabularies for classifying text corpora and predefined response categories to facilitate self-identification in survey questions. Deliverable D3.10 A 20th century version of the occupation multilingual ontology focuses on occupational ontologies in the 20th century. As a database for self-selection of occupational titles in any labour market can easily include thousands of titles, it should allow for filtering in case the survey question addresses parental or previous occupational titles which might be in existence for the relevant decades but are no longer so in current times. In Section 2 two approaches were applied to build an ontology for the first half of the 20th century occupations. First, the occupational titles found in the documents of the 1909, 1920, and 1930 Censuses in the Netherlands were merged, leading to the conclusion that these lists most likely did not provide a full overview of all titles present in those years. This resulted in a list of 1,071 occupational titles which mostly did not overlap in the three years. Second, the newspaper archive Delpher in the Netherlands was explored whether it could be used to indicate in which decades the occupational titles were reflected in the Newspapers. A list of 26 titles existing only in the first half of the 20th century were matched with the texts. For almost half of these occupations 80% of the matches indeed were found in the years between 1900-1940, indicating that the archive is a useable tool to indicate in which decades the occupations were present in the Dutch labour market, but for more than half the titles the matches were below 80%. In Section 3 the international ISCO classifications for 1958, 1986, 1988, and 2008 were used for the ontology reflecting a multilingual occupational ontology for the second half of the 20th century. The available translations of the ISCO 1958, 1968, 1988, 2008 coding indexes were merged. ISCO 1958 and 1968 were available at 5-digit level and in the languages English, French, and Spanish. ISCO 1988 is available at 4-digit level and in the languages English, French, Spanish, and German. ISCO 2008 is available at 4-digit level and in 51 languages plus for another 2 languages not all entries have translations. The file ISCO58_68_88_08_4_5dgt_en_fr_es_many_20190916.xlsx is downloadable1 . In Section 4 the occupational titles throughout respondents’ entire working careers were coded in the Survey of Health, Ageing and Retirement in Europe (SHARE) waves. A so-called job-coder was used to classify the respondents’ current or past jobs as well as their parents’ jobs in the ISCO-08 classification directly at the time of the interview. The job-coder was developed by SSHOC partner SHARE/CENTERDATA based on the occupation database in surveycodings. SSHOC partner SHARE/UNIVE recovered the job codification for the retrospective careers and parents’ occupations. The effort allowed to observe that the distribution over occupations changed gradually, moving from those with low and medium-low level of qualification towards the more qualified ones over the years 1970, 1980, 1990 and 2000.