Table 4 Overview of variables and how they are measured

Low literacy

 LEO

Alpha-level 3 and below (Hartig and Riekmann 2012), 5 variables derived from 5 plausible values for literacy skills for each respondent

 PIAAC

Proficiency level 1 and below (Yamamoto et al. 2012b), 10 variables derived from 10 plausible values for literacy skills for each respondent

 NEPS

Proficiency level 1 (Durda et al. 2020), 1 variable derived from 1 WLE value for literacy skill for each respondent

Gender

 LEO

Ascertained by information on the gender

Recoded into 0: female, 1: male

 PIAAC

 NEPS

Age

 LEO

Ascertained by information on the date of birth and the interview date For logistic regression recoded into: 1: 24 to 33 years, 2: 34 to 43 years, 3: 44 to 53 years, 4: 54 to 64 years

 PIAAC

 NEPS

Native language

 LEO

Which language did you learn first in your childhood? (1: German, 2: another language, 3: bilingual with German and other language, 4: bilingual, without German, 9: not specified)

Recoded into 0: German, 1: non-German

 PIAAC

What is your mother tongue? That is, the first language you spoke at home during your childhood and that you still understand (1: German, 2: Turkish, 3: Italian, 4: Polish, 5: Greek, 6: Serbian, 7: Croatian, 8: Russian, 9: another language); Recoded into 0: German, 1: non-German

 NEPS

What language did you learn as a child in your family? Recoded into 0: German, 1: non-German

Highest educational attainment

 LEO

International Standard Classification of Education (ISCED-97)

Recoded into 3: no (ISCED 1 and below), 2: low (ISCED 2), 1: medium (ISCED 3/4), 0: high (ISCED 5/6)

 PIAAC

 NEPS

Occupational status

 LEO

International Socioeconomic Index of Occupational Status (ISEI-08, Ganzeboom 2010)

For logistic regression recoded into: 4: low occupational status (11.56 to 30.91), 3: medium occupational status (30.92 to 50.25), 2: high occupational status (50.26 to 69.60) and 1: very high occupational status (69.61 to 88.96). Note: In NEPS, determined from the spell type (sptype) in the episode split of the biography data (for details, see Rompczyk and Kleinert 2017), in cases of multiple spells, employment spell with the longest duration was selected

 PIAAC

 NEPS

Employment status

 LEO

Are you personally at present … (1: full-time employed, 2: part-time employed, 3: unemployed, 4: on parental leave, 5: trainee in company apprenticeship, … 11: other, 99: not specified)

Recoded into 1: employed, 2: unemployed, 3: other

 PIAAC

Please have a look at this list and tell me which statement best describes your current situation. Decide which statement you think best suits you (1: full-time employed, 2: part-time working: 3: unemployed, 3: school pupil, student, …10: other)

Recoded into 1: employed, 2: unemployed, 3: other

 NEPS

Employment status is determined from the spell type (sptype) in the episode split of the biography data (for details, see Rompczyk and Kleinert 2017)

Recoded into 1: employed, 2: unemployed, 3: other

Sampling weights

 LEO

Nonresponse weights [pgew]

 PIAAC

Nonresponse weights [SPFWT0] and replicate weights [SPFWT1-SPFWT80]

 NEPS

Nonresponse weights [w_t3_cal or w_t5_cal]