Implementation and Methodological Framework of the SHARE study in Croatia

SHARE – Istraživanje o zdravlju, starenju i umirovljenju u Europi je istraživačka podatkovna infrastruktura koju je Europska komisija pozicionirala kao jedan od prioritetnih projekata u europskom istraživačkom prostoru. U ovoj multidisciplinarnoj i transnacionalnoj bazi mikropodataka o zdravlju, socioekonomskom statusu te društvenim i obiteljskim mrežama nalaze se informacije o više od 140.000 osoba u dobi od 50 ili više godina iz 27 europskih zemalja i Izraela. Cilj ovog rada jest prikazati metodološki profil studije SHARE s posebnim osvrtom na razdoblje od priključenja Hrvatske 2015. U radu se analizira proces provedbe ankete SHARE u Hrvatskoj s detaljnim prikazom modula i pitanja iz šestog i sedmog vala studije. Svrha rada jest omogućiti hrvatskoj znanstvenoj zajednici bolji uvid u sadržaj ankete SHARE. U radu se daje i poseban osvrt na osmi val, u kojem je zbog pandemije koronavirusa prekinuto anketiranje licem u lice i napravljen metodološki zaokret, tj. prelazak na CATI anketu, tzv. SHARE Corona. Studija SHARE danas se suočava s brojnim izazovima, od kojih valja istaknuti zadržavanje europske pokrivenosti zbog nedostatnog financiranja, ali i budućeg načina provedbe ankete zbog utjecaja bolesti COVID-19.The Survey of Health, Ageing and Retirement in Europe (SHARE) is a multidisci¬plinary and cross-national panel database of microdata on health, socio-economic status and social and family networks of about 140,000 individuals aged 50 or older. SHARE covers 27 European countries and Israel. It is a research infrastructure that the European Commission has positioned as one of the priority projects in the Euro¬pean Research Area and one of the most significant panel studies in social sciences. This paper aims to present the methodological profile of the SHARE study with a specific reference to the period since 2015 when Croatia joined it. It also examines the process of SHARE study implementation in Croatia with a detailed presentation of modules and questions from the questionnaires in the sixth and seventh waves. The specific purpose of this paper is to emphasise the content and research opportunities of the SHARE study for the Croatian scholarly community. This paper highlights the specific aspects of the eighth wave of SHARE in which face-to-face interviewing has been suspended due to the coronavirus pandemic, and a methodological turn has been made, i.e. face-to-face interviews have been replaced with a short CATI (telephone) survey called “SHARE Corona.” As a panel study, SHARE collects data in waves, every two years since 2004, sup¬ported by CAPI (computer-assisted personal interviewing). A SHARE interview is quite long, with an approximate duration of one hour. Same respondents are inter¬viewed in regular waves, but new respondents can be added to refresher samples. New respondents help to maintain the core sample and deal with sample attrition. However, SHARE invests significant efforts in recovering panel respondents who have participated in the previous waves. The strategy to minimise the absence of panel respondents includes regular contact with respondents, i.e. panel care (send¬ing birthday cards, season’s greetings, brochures with selected SHARE results to re¬spondents or providing incentives). The SHARE study relies on ex-ante harmonisa¬tion and includes a core survey instrument that is common in all member countries. Strict comparability is crucial. Comparability is ascertained through identical ques¬tion design, a careful translation process with external certification, an electronic survey instrument (CAPI, CASE CTRL starting from Wave 8), and common training procedures ascertained by a train-the-trainer programme. In addition to the variety of electronic instruments, SHARE relies on several “physical” survey instruments, which are mainly used to obtain objective health measures. These instruments in¬clude dynamometers (to measure respondents’ grip strength) and peak flow me¬ters (to measure respondents’ lung capacity). Besides, SHARE collects data from the walking speed test and chair stand test, data on waist circumference, self-reported weight and height, and biomarkers from a sample of dried blood spots (HbA1c, total cholesterol, C-reactive protein and vitamin D). In the participating SHARE countries, the institutional conditions with respect to sampling are so different that a uniform sampling design for the entire project is infeasible. Good sampling frames for the target population of 50+ individuals and households with at least one 50+ individual do not exist or cannot be used in all countries. Most countries keep registers of individuals that enable stratification by age. In some of them, these registers are maintained at a regional level. In these cases, a two- or multi-stage design is needed, in which regions are sampled first, and then individuals are selected within these regions. As a result, sampling designs used vary from a simple random selection of households to rather complicated multi-stage de¬signs. Taking into account the size of the population in each participating country, SHARE calculates weights to reduce the potential selection bias associated with non-response errors. In the Croatian wave six, the sampling of potential respondents was based on probabilities from the administrative register of age-appropriate individu¬als. From the database of insured persons of the Croatian Health Insurance Fund (HZZO), 4,990 persons born in 1963 and earlier were randomly selected. Each person received an invitation letter for participation in the SHARE study in Croatia. The response rate in SHARE wave six was 43.7% at the household level, while the indi¬vidual response rate was 41.9%. These rates resulted in 2,495 individual surveys con¬ducted in 1,588 households in Croatia. In all countries that had refresher samples, the response rate was 51.3% at the household level and 46.8% at the individual level. The seventh SHARE wave, called SHARELIFE, was mainly retrospective, accomplishing a full EU coverage. As a part of the seventh wave, a relatively small refresher sample was selected in Croatia i.e. 346 interviews were conducted in 234 households. The minimum satisfactory response rate of 30% was achieved at the household level. In Croatia, the retention rate of respondents in the seventh wave was 84.6%, which was the highest retention rate of respondents between waves six and seven. In Slovenia, it was 82.9%, in Greece 82.8%, in Estonia 82.2%, in Belgium 70.4%, in France 64.9%, and in Italy 62%. That retention rate in Croatia resulted in 2,062 SHARELIFE inter¬views with an additional 101 end-of-life interviews (interviews about the last year of life of a deceased respondent). The sampling procedure for the refresher sample in wave eight in Croatia followed the standard phases of two-stage sampling. Primary sampling units were polling stations selected based on a probability proportional to the number of voters aged 50+ at each polling station. The sample was stratified by counties and by settlement size. In the second phase, the gross sample of individuals aged 50 or older was selected randomly. The representativeness of the sample was achieved by weighting a set of eight calibration variables (men and women in age groups 50–59, 60–69, 70–79, and 80+). Following the spread of the coronavirus pandemic across Europe, in March 2020, all SHARE countries suspended field surveys. By that date, 1,279 panel interviews (including end-of-life interviews) and 835 refresher interviews had been collected in Croatia. In response to the pandemic, the SHARE Central, in cooperation with national teams, created the SHARE Corona questionnaire, designed for a computer-assisted telephone interview (CATI), lasting about 20–25 minutes. This survey, con¬ducted in 27 European countries and Israel from June to August 2020, included panel respondents only. The SHARE study is a prime example of a truly European research infrastructure that exists largely because of its European dimension. It is crucial to point out that the data collected by the survey questionnaire are harmonised ex-ante across Europe, which significantly contributes to the improvement of international comparative re¬search. The main value of this project lies with the diversity of collected data, with each participating country contributing to this diversity with data on living condi¬tions, health, pension and social policies. Therefore, SHARE is much more than just a group of national surveys. The SHARE study today faces a number of challenges, the most prominent ones being the retention of European coverage due to lack of fund¬ing and the future method of surveying in the light of COVID-19. In Croatia, joining the SHARE study was marked by significant challenges. Firstly, SHARE is the first longitudinal study on demographic ageing conducted in our country. Secondly, the SHARE survey requires ample financial resources, so the size of the Croatian sam¬ple had to be adjusted accordingly and to meet high scientific standards set by the SHARE study. The third significant challenge that the SHARE research team faced were barriers to accessing the sampling framework. Substantial efforts were under¬taken to demonstrate that the SHARE study adheres to all ethical standards and regulations related to the protection of the personal data of respondents. Another is¬sue was the limited number of survey agencies in Croatia that can conduct demand¬ing surveys of this type. This was especially evident in the sixth wave, when certain logistical issues arose because the fieldwork phase started quite late (end of June 2015), leading to the stagnation of surveys in some parts of the country (Dalmatia, Istria). Difficulties related to the recruitment of interviewers and their withdrawal in the early stages of the survey required significant efforts during the last month of the fieldwork. However, experiences from the sixth wave contributed to the extremely successful implementation of the seventh wave (SHARELIFE)