The variable PGSNDJOB represents the imputed current gross labor income from second job generated for all SOEP respondents in each respective wave. Income details are consistently provided in euros for all waves. Item nonresponse is imputed in a two-stage procedure: first, with the “Row-and-Column” method of Little und Su (1989) using individual longitudinal data as well as cross-sectional trend data (cf. Joachim R. Frick and Markus M. Grabka (2005): Item-Non-Response on Income Questions in Panel surveys: Incidence, Imputation and the Impact on the Income Distribution. Allgemeines Statistisches Archiv (ASTA) 89, 49–61). Alternatively, if no individual longitudinal information is available, we base the imputation on a regression using a subset of different Mincer covariates. Information about gross income from second job is firstly asked in wave 1995 (L). In 2013 information about income from second jobs was not collected for sub-sample M. We use information from the the subsequent survey year (2013) about income from second job in the previous year (variable BEP2C03). If persons from the sub-sample M stated that he/she had income in the previous year from second jobs, then PGIMPSND in year 2013 was set to -1 and PGSNDJOB had been imputed. Imputed values are flagged (PGIMPSND).
The corresponding varialbles in file PL are PLC0062_V1 and PLC0062_V2. The original variables coming from the $P files are: lp7702, mp5802, np5802, op4902, pp6602, qp6302, rp6602, sp6602, tp8402, up67, vp79, wp70, xp81, yp76, zp78, bap69, bbp79, bcp67, bdp85, bep64, bfp113, bgp97.
The variable PGSNDJOB represents the imputed current gross labor income from second job generated for all SOEP respondents in each respective wave. Income details are consistently provided in euros for all waves. Item nonresponse is imputed in a two-stage procedure: first, with the “Row-and-Column” method of Little und Su (1989) using individual longitudinal data as well as cross-sectional trend data (cf. Joachim R. Frick and Markus M. Grabka (2005): Item-Non-Response on Income Questions in Panel surveys: Incidence, Imputation and the Impact on the Income Distribution. Allgemeines Statistisches Archiv (ASTA) 89, 49–61). Alternatively, if no individual longitudinal information is available, we base the imputation on a regression using a subset of different Mincer covariates. Information about gross income from second job is firstly asked in wave 1995 (L). In 2013 information about income from second jobs was not collected for sub-sample M. We use information from the the subsequent survey year (2013) about income from second job in the previous year (variable BEP2C03). If persons from the sub-sample M stated that he/she had income in the previous year from second jobs, then PGIMPSND in year 2013 was set to -1 and PGSNDJOB had been imputed. Imputed values are flagged (PGIMPSND).
The corresponding varialbles in file PL are PLC0062_V1 and PLC0062_V2. The original variables coming from the $P files are: lp7702, mp5802, np5802, op4902, pp6602, qp6302, rp6602, sp6602, tp8402, up67, vp79, wp70, xp81, yp76, zp78, bap69, bbp79, bcp67, bdp85, bep64, bfp113, bgp97.