SOEP-Core - 2018: Documentation of the interviewer dataset (1984 until 2018)

Gerike, Martin Research Report SOEP-Core - 2018: Documentation of the interviewer dataset (1984 until 2018) SOEP Survey Papers, No. 825 Provided in Cooperation with: German Institute for Economic Research (DIW Berlin) Suggested Citation: Gerike, Martin (2020) : SOEP-Core - 2018: Documentation of the interviewer dataset (1984 until 2018), SOEP Survey Papers, No. 825, Deutsches Institut für Wirtschaftsforschung (DIW), Berlin This Version is available at: https://hdl.handle.net/10419/214744 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-sa/4.0/ SOEP Survey Papers Series D – Variable Descriptions and Coding SOEP-Core – 2018: Documentation of the Interviewer Dataset (1984 until 2018) 825 SOEP — The German Socio-Economic Panel at DIW Berlin 2020 Martin Gerike SOEP-Core – 2018: Documentation of the Interviewer Dataset (1984 until 2018) Martin Gerike* February 24, 2020 DIW Berlin – Deutsches Institut für Wirtschaftsforschung e.V. Sozio-oekonomisches Panel (SOEP) Mohrenstraße 58 10117 Berlin *based on the 2017 version of Martin Kroh, Luisa Hilgert and Arne Langlet Contents 1 Overview 2 2 Contents of INTERVIEWER 3 2.1 Personnel Records .................................... 3 2.2 Interviewer Surveys 2006, 2012 and 2016 ....................... 3 2.3 Fieldwork Outcomes and Interviewer Activity .................... 4 2.3.1 Basic Information ................................ 4 2.3.2 Fieldwork: Period and Region ......................... 4 2.3.3 Workload and Modes .............................. 5 2.4 Variables from Multiple Sources ............................ 7 3 Descriptive Statistics 9 4 Matching Person and Interviewer Data 15 5 Appendix 17 5.1 Useful Links ....................................... 17 5.2 Questionnaires ...................................... 17 1 1 Overview The SOEP does not only aim at collecting high-quality data on the living conditions and wellbeing of households, but –as a by-product of internal quality assurance processes– it lends itself increasingly as an empirical source for survey research. The purpose of the INTERVIEWER file is to provide user convenient access to all available, longitudinal information on the SOEP interviewers. These information on interviewers may be linked to the wave-specific person and household files via the interviewer identifier (INTID) and the year of observation. Information stemming from different sources are brought together in this file. First, we derive basic demographical information and employment spells on interviewers from personnel data of the fieldwork organization. Since 2000, Kantar Public regularly updates these information. Second, in 2006, 2012 and 2016, more than 95 percent of all SOEP interviewers of these years completed a short version of the standard individual questionnaire themselves. This allows, among others, a one to one match between responses of interviewers and responses of interviewees. Third, we generated user-friendly information on the working experience and workload of interviewers from the SOEP person and household interviews conducted by them as well as the fieldwork information they report to Kantar Public. The INTERVIEWER file is available in long format (interviewer ×waves) and covers all interviewers who ever conducted at least one interview in the SOEP and a few more interviewers active in SOEP-related studies and pretests who participated in the interviewer surveys of 2006, 2012 and 2016. In 2018, the INTERVIEWER file contains a total of 2935 interviewers. Many of the variables of INTERVIEWER are by definition time-variant, such as the annual workload. Information on gender, year of birth, the start and end of being an interviewer at the SOEP is time-invariant as well as information on country of birth, mother tongue, and country of birth of the interviewer’s parents. The same holds true for classification of interviewers in the face-to-face fieldwork and those working in the centralized call center administrating mailed questionnaires. 2 2 Contents of INTERVIEWER This section provides an overview of information included in the INTERVIEWER file as well as the respective sources of these information. In the case of generated variables, this section also describes the coding of this information. The following pages give an overview of all variables included in the INTERVIEWER file. The Appendix provides some additionnal sources of information for different purposes such as useful links to the INTERVIEWER-related parts of SOEP Companion and Paneldata.org or the questionnaires. 2.1 Personnel Records Kantar Public personnel data1are updated every year as of 2000. They contain information on the working experience of interviewers within the fieldwork company, education, and marital status (see section 2.4). 2.2 Interviewer Surveys 2006, 2012 and 2016 In 2006, 2012 and 2016, the interviewers filled in (mailed) questionnaires themselves. The interviewer questionnaire mimics for the most part the regular SOEP person questionnaire. Hence, we obtain not only information on basic demographics of interviewers (gender, year of birth, etc.), but also their occupational history, health status, denomination, family status, country of birth, and mother tongue. While in the surveys of 2006 and 2012 the interviewers were asked about their gender, marital status and education, these questions were omitted in the 2016 survey. Moreover, we collect standard SOEP data on social and political attitudes, life satisfaction, and the interviewer’s worries. Additionally questions relate to their personality drawing on the "Big Five"-inventory, risk aversion, impulsiveness, and trust (see questionnaire at the end of the document). Besides these SOEP-items, interviewers are asked to answer some questions which directly relate to their work. Thus the interviewers are asked about their motivation to work as an interviewer and whether these expectations are fulfilled and for how long they plan to work as an interviewer for the SOEP. Furthermore, the interviewers are asked to estimate the willingness of different given social groups to participate in the SOEP. Finally, the interviewers report on items of an inventory of social desirable responding (Winkler et al. 2006).2In 2012, interviewers also report on their participation in interviewer trainings and working satisfaction. Furthermore, interviewers are asked about their competence in a set of languages and about the country of birth of their parents. Finally, interviewers are asked about their actual and preferred working hours. Country of birth, mother tongue, and country of birth of the interviewer’s parents (all binary "Germany" vs. "not Germany") is time-invariant and therefore extended to all years. The INTERVIEWER file lists all time variant information in the respective survey years 2006, 2012 1These data were provided to SOEP users until data version v30 in form of the INTVIEW-files 2Winkler, Niels, Kroh, Martin, and Martin Spiess. 2006. Entwicklung einer deutschen Kurzskala zur zweidimensionalen Messung von sozialer Erwünschtheit. DIW Discussion Paper 579. 3 and 2016 only. In case of conflicting information between personnel data of the fieldwork organization and the interviewers’ self-reports provided in the surveys, we document this in form of a flag-variable. 2.3 Fieldwork Outcomes and Interviewer Activity Information concerning the work as an interviewer in the SOEP is derived from the $p-, $h-, $pbrutto, and $hbrutto-files for the years 1984 till 2018, the $lela-files for the years 1995 till 2018, and the $youth-files for the years 2000 till 2018. Particularly, we use meta-data from these files on, for instance, the number of contacts before interview, the mode of data collection, date of the interview, duration of the interview, and of course the interviewer who conducted the interview. We aggregate the person and household level data, for instance, to the amount of interviews conducted in one year by an interviewer, the pace of interviewing, the mean number of contacts with households, the region of the primary interviewer allocation, and the size of the communities of the primary interviewer allocation. The following paragraphs document the generation of these variables: 2.3.1 Basic Information TYPINT (interviewer works in the face to face field or in the centralized call center) The variable TYPINT indicates whether an interviewer works in the face to face field or if they administers mailed questionnaires from a centralized call center. EXPERIENCE The variable EXPERIENCE gives the cumulative number of years an interviewer worked for the SOEP so far. EXPERIENCE is given for each interviewer and for each year from 1984 until 2018. The variable increases by 1 if the interviewer conducts at least one person, household, or youth questionnaire in one year. 2.3.2 Fieldwork: Period and Region FIRSTINTM (month of first interview in respective year) The variable FIRSTINTM provides the month of the first personal, youth, and household interview for each interviewer for each year from 1984 until 2018. The generation of the variable FIRSTINTM is based upon the variables $pmonin from the $p-files, $jmonin from the $youth-files and upon the following variables from the $h-files: $hmonin for the years 1984 until 1988, as well as $hhmonin from 1989 onwards. FIRSTINTD (day of first interview in respective year) The variable FIRSTINTD provides the day of the first personal, youth, and household interview for each interviewer for each year from 1984 until 2018. The generation of the variable FIRSTINTD is based upon the variables $ptagin from the $p-files, $jtagin from the $youth-files and upon the following variables from the $h-files: $htagin for the years 1984 until 1988, as well as $hhtagin from 1989 onwards. LASTINTM (month of last interview in respective year) The variable LASTINTM provides the month of the last personal, youth, and household interview for each interviewer for each year 4 from 1984 until 2018. The generation of the variable LASTINTM is based upon the variables $pmonin from the $p-files, $jmonin from the $youth-files and upon the following variables from the $h-files: $hmonin for the years 1984 until 1988, as well as $hhmonin from 1989 onwards. LASTINTD (day of last interview in respective year) The variable LASTINTD provides the day of the last personal, youth, and household interview for each interviewer for each year from 1984 until 2018. The generation of the variable LASTINTD is based upon the variables $ptagin from the $p-files, $jtagin from the $youth-files and upon the following variables from the $h-files: $htagin for the years 1984 until 1988, as well as $hhtagin from 1989 onwards. ENDINT (year of Last SOEP-Interview) The variable ENDINT provides the year of the last SOEP-interview for each interviewer. It is constant over time. The generation of the variable ENDINTis based upon information from the SOEP fieldwork data that indicates when the interviewer’s last year was when he or she worked for the SOEP. MODBULA (primary interviewer allocation, federal state) The variable MODBULA provides the mode of the German Federal Land in which the household-interviews were conducted. It is given for each interviewer and for each year from 1984 until 2018. MODGGK (primary interviewer allocation, size of community) The variable MODGGK provides the mode of the population size of the communities where the household-interviews were conducted. It is given for each interviewer and for each year from 1984 until 2018. 2.3.3 Workload and Modes MEANCONTACTHH (mean number of HH contacts made (gross HH data)) The variable MEANCONTACTHH reports the average number of contacts before interviewers realize the interview (or the household finally refuses to participate). The variable provides the mean per interviewer for each year between 1985 and 2018. It is the mean of the variable $intk from the $hbrutto-file. Note that this mean may underestimate the true value as the maximal category of the number of contacts is "9 or more interviewer contacts". RESPONSERATE (HH Response Rate according to AAPOR definition rr6) The variable RESPONSERATE gives the (longitudinal) response rate as described by the American Association for Public Opinion Research (AAPOR). This means that RESPONSERATE is defined as the share of the realized interviews in one wave divided by the number of eligible households in interviewers. Eligible households are all responding households from the previous wave, split-offs from these households, and temporal drop-outs from previous waves. We define the resolution of households by death and emigration as ineligibility, and these cases are therefore excluded from the calculation of RESPONSERATE. It is given for each interviewer and for each year from 1985 until 2018. For the generation of the variable RESPONSERATE, we use the variable $hergs from the $hbrutto-files. 5 Figure 3: Median Age and Mean Experience of Interviewers by Wave Figure 3shows that the strong commitment of interviewers to the SOEP survey leads to an increasing median age of the interviewers over time. In 1984, the median age is about 42 years and it reaches 67 years in 2017. The general trend of aging interviewers is interrupted in years with a substantial influx of new interviewers as in 2000. At the same time, we observe an increase in interviewer experience, measured in the cumulative number of years in which an interviewer conducted interviews5. Starting at one in 1984, the mean experience reaches, on average, 11,7 years of experience in 2018. 5This measure ist not always equal to the number of years since the start of working as an interviewer, as interviewers might pause in some years. 12 Table 3.2: Interviewer Attributes Year 2000 2006 2012 2016 number of observations 770 624 729 653 male 452 367 404 369 female 307 252 293 277 secondary school degree 155 98 105 58 intermediate school degree 285 219 248 163 technical school degree 54 49 upper secondary school deg. 99 199 163 63 graduate degree 179 17 63 138 other degree 12 8 unwed 71 57 65 58 divorced 63 54 widowed 111 55 56 66 married 574 429 433 305 Table 3.2 gives an overview of some interviewer attributes in 2000, 2006, 2012 and 2016. Please note that due to more detailed response categories, the information on marital status and education is more detailed in years of the interviewer survey, namely 2006 and 2012. On average, interviewers working in the SOEP have about average education and marital status compared to the German population, though the number of male interviewers exceeds the number of females. 13 Figure 4: Workload household-level Figure 5: Workload personal-level Figure 4 shows the median number of household-interviews per interviewer and the median of the mean interview length in minutes for each year between 1985 and 2018. The number of interviews increases clearly over the years and has almost doubled since 1984. The median length of the interviews conducted by interviewers is relatively stable and ranges typically between around 15 and 20 minutes. In 2010 however, an extreme increase in length of interviews can be observed. Figure 5 shows the same variables only on the person-level. Just as in the previous graphic, you can observe a massive increase of numbers of interviews while the median of the mean interview length is, with a few exceptions, mostly between 30 and 40 minutes. 14 4 Matching Person and Interviewer Data The interviewer identifier INTID that is recorded in both the INTERVIEWER file and all person and household files of the SOEP allows to directly link information of interviewers with information on interviewees.6The STATA code on the next page illustrates how to match the "political party supported" of respondents with the "political party supported" of their interviewers in 2016 (Wave BG). In the first line the unique ID for individuals "persnr", the survey year "syear", the unique ID for interviewers "intid" and a variable for party preference "bgp14501" are loaded into Stata from the person dataset of the wave BG (2016). Next, the intid is used to merge the variable ipol3, the interviewer’s party preference. In line four, the value labels for the newly merged variable are taken from the existing one for the purpose of easier readability. In the last step, a cross tabulation for the values 1-8 of the two variables is displayed (assuming both labels are identical as well). The cross-tabulation of party identification in respondents and interviewers indicates a statistically significant association (Cramer’s V = .0993825). Please note that this association may not necessarily reflect causal influence attributable to the social interaction between interviewer and respondent. It may also result from regional differences in party affiliation and the local interviewer allocation in the face-to-face fieldwork. In Table 4.1, interviewers might be counted several times. 6Please note that the INTERVIEWER file also contains a second identifier, hhnr. This second identifier does not refer to any household of SOEP respondents, but serves as an internal case identifier of the SOEP data bank at DIW Berlin. 15 1 use p e r s n r s yea r i n t i d bgp14501 us i n g "$DATA/ bgp . d ta " , c l e a r 2 3 merge m: 1 i n t i d s y ea r u s ing "$OUT / i n t e r v i e w e r . d ta " , keepus ( i p o l 3 ) keep ( match ) nogen 4 l a b e l v al ue s i p o l 3 bgp14501 5 t ab i p o l 3 bgp14501 i f bgp14501 <9 & bgp14501 >0 & i po l3 <9 & ip ol 3 >0 , V Table 4.1: Party Identification of Respondent and Interviewer in 2016 Parteipraeferenz Favored party (interviewer) SPD CDU/CSU FDP Greens The Left Right-Wing Parties Other Total N % N % N % N % N % N % N % N % SPD 370 30 499 26 35 29 211 32 66 22 5 17 18 40 1204 28 CDU/CSU 496 41 823 43 31 26 237 35 100 33 12 40 10 22 1709 40 FDP 72 6 127 7 15 13 53 8 12 4 0 0 6 13 285 7 Greens 170 14 256 14 29 24 115 17 29 10 1 3 8 18 608 14 The Left 77 6 158 8 8 7 34 5 68 23 8 27 2 4 355 8 Right-Wing Parties 24 2 23 1 1 1 18 3 23 8 3 10 0 0 92 2 Other 7 1 7 0 0 0 0 0 1 0 1 3 1 2 17 0 Total 1216 100 1893 100 119 100 668 100 299 100 30 100 45 100 4270 100 Cramer’s V = .0993825 16 5 Appendix 5.1 Useful Links For further information please use SOEP Companion and Paneldata.org. 5.2 Questionnaires The German questionnaires to the different interviewer surveys can be accessed under the following links: 2006 2012 2016 17