Left behind single on the partnering market? Entry into cohabiting unions by women and men with low educational attainment across regions of Europe, cohorts 1960 to 1985 By Nadia Sturm and Jan Van Bavel --------------------------------------------------------------------------------------------------------------------------------- Table of contents 1) Getting the data 2) Variables 3) Explanation of data files --------------------------------------------------------------------------------------------------------------------------------- 1) Getting the data The ESS data is available free of charge for non-commercial use and can be downloaded from the ESS DataPortal after a short registration: https://ess.sikt.no/en/data-builder/?tab=round_country The data is available either as a country-specific file or in an integrated international dataset, which includes data from all countries for a particular round. For this paper we downloaded all countries and rounds 1 to 10SC (round 10SC are only the self completion interviews of round 10, but they are excluded in a later step). Several countries are excluded in later steps, as outlined in more detailed in the paper as well as in the R Do Files. For the code, R version 4.3.0 is used. --------------------------------------------------------------------------------------------------------------------------------- 2) Variables We select the following variables for the analysis. Basic variables - cntry; essround: Country and Essround are downloaded autcomatically to the data set when downloading the data. - mode: Indicates interview mode. We use this to exclude the self-completed interviews in Round 10. - gndr: Gender - agea: Age at interview - yrbrn: Year of birth Education - edulvla: Education following ISCED Scale 1 to 5 for round 1 to 4 in the ESS. - eisced: Education following ISCED Scale from 1 to 6 for round 5 to 10 of the ESS. - edctn: Currently enrolled in education in the past 7 days. Parental education - edulvlfa & edulvlma: Father's and Mother's education, ISCED Scale, Round 1 to 2. - edulvlfb & edulvlmb: Father's and Mother's education, ISCED Scale, Round 3 to 4. - eiscedf & eiscedm : Father's and Mother's education, ISCED Scale, Round 5 to 10. Union formation - marital : Marital status Round 1 to 2. Includes also information on divorce. - maritala: Marital status Round 3 to 4. Includes also information on divorce. - lvgptne : Ever lived with a partner without being married, Round 1 to 4. - lvgptnea: Ever lived with a partner, without being married, Round 5 to 10. - dvrcdeva: Ever been divorced/had civil union dissolved, Round 5 to 10. - hhmmb: Number of people living in the household - rship2 to rship15: Relation to person living in the household, Round 1. - rshipa2 to rshipa18: Relation to person living in the household, Round 2 to 10. - rshpsts : Relationship with husband/wife/partner currently living with, Round 5 to 10. Religion - rlgblg : Belonging to particular religion or denomination at time of interview Migration background: - brncntr : Born in country - cntbrth : Country of birth - Round 1 - cntbrtha: Country of birth - Round 2 to 3 - cntbrthb: Country of birth - Round 4 to 5 - cntbrthc: Country of birth - Round 6 to 8 - cntbrthd: Country of birth - Round 9 to 10 Cohabition (only asked in Round 3 and 9 of the ESS; we use it to calculate the mean ages at first cohabitation) - lvptnyr : Year first cohabited with a partner for more than 3 months - evlvptn : Ever cohabited with a partner for more than 3 months. --------------------------------------------------------------------------------------------------------------------------------- 3) Explanation of data files There are 4 RMarkdown Files included in the Zip File. Below is an explanation of each data file and how to use it to replicate the data. For the code, R version 4.3.0 is used. a) Union formation_35to45_Sturm This file includes the main analysis and the barplots (Figure 1 to 4). The data file is currently selecting women (gndr == 2). To run the same results for men, simply run the code with gndr == 1. This data file includes the code for the logistic regression (Table 2 in the paper) and the code for the predicted probabilities (see graphs in Figure 5). To recreate the result for men (see Figure 6, and Appendix A2 to 6) run the code with gndr == 1. The file also contains the robustness check with a categorical measure of cohort instead of a continuous one. Explanations of the code are included troughout the document. b) Union formation_men and women_Sturm This file includes the results for a joint analysis for men and women. This file repoduces the predicted probabilites for both men and women after including a three-way interaction (Figure 7 and Table A7 in the Appendix) and the logistic regression on a two-way interaction for only the lower educated (Table 3). Explanations of the code are included troughout the document. c) Union formation_Robustness check_Sturm This file includes the robustness checks mentioned in the Note at the end of the 'Data and sample selection' section: Pooled sample of countries with controls for country, pooled sample with clustered standard errors, Multi level analysis with country fixed-effects. Explanations of the code are included troughout the document. d) Mean ages at first cohabitation_Sturm This code only includes ESS Round 3 and 9. It is used to calculate the mean ages at first cohabitation and shares cohabiting by age 35, between ages 35 and 45 and not cohabiting ever by that age. In the data, men and women are included jointly. Explanations of the code are included troughout the document.