Development of Sample Sizes

Individuals who refuse participation or are not available for an interview are kept in the so-called “gross” sample of the study as long as they continue to live in households with at least one participating person. Once the entire household declines to respond in two consecutive waves of data collection, all individuals from the household are removed from the SOEP. The table shows the starting sample sizes of samples A through M4, the years when the samples were first collected, as well as the percentage of those persons who were eligible for an interview but declined participation (“partial unit non-response”, PUNR) in the first wave. The figure illustrates the development of the number of successful person interviews since 1984. The reduction in the population size for all individual samples is mainly the result of person-level drop-outs, refusals, moving abroad, etc. However, due to new persons moving into already existing households, and children reaching the minimum respondent’s age of 16, and thereby increasing the sample size, this negative development is offset somewhat.

Starting Sample Size of the SOEP Samples

Sample Year Households (net) Persons(gross) Respondents (net) Partial Unit Non-Response (percent) Children (gross)
A 1984 4528 11422 9076 0.6 2290
B 1984 1393 4830 3169 0.7 1636
C 1990 2179 6131 4453 1.9 1591
D1 1994 236 733 471 2.9 248
D1/D2 1995 541 1668 1078 6.1 517
E 1998 1057 2446 1910 3.5 466
F 2000 6043 14510 10880 5.5 2991
G 2002 1224 3538 2671 6.1 693
H 2006 1506 3407 2616 6.0 623
I 2009 1495 3428 2432 13.4 620
J 2011 3136 6873 5161 9.9 1147
K 2012 1526 3286 2473 9.2 563
L1 2010 2074 7939 3770 6.7 3900
L2 2010 2500 9063 4227 5.1 4611
L3 2011 924 3645 1487 4.2 2092
M1 2013 2723 8522 4964 17.8 2481
M2 2015 1096 3048 1711 19.3 927
M3 2016 1775 4823 2351 22.0 1808
M4 2016 1779 7297 2465 27.1 3915

Cross-Sectional Development of Sample Size (Respondents)

../_images/crossdevel.PNG

Download Stata Code to create figure

This cross-sectional view is insufficient when examining the longitudinal development of the sample, which is influenced by different demographic and field-work related factors. As already shown, demographic reasons for entering the panel are birth and residential mobility. Analogously, the demographic reasons for a panel exit are death and moving abroad. Fieldwork related reasons are different, in that they relate to the interaction between the interviewer and the responding household. Respondents are either not reached for an interview (non-contact) or they decline to participate for the current year. The figure illustrates the longitudinal development of first-wave respondents in 1984, as well as their children, of samples A and B.

Longitudinal Development of the 1984 Population

../_images/where2.png

Download Stata Code to create figure