Mixed probability and non-probability sampling (DDI Alliance Controlled Vocabulary for SamplingProcedure)
Multistage (4-stage) sampling design.
First stage. The territory of Lithuania was divided into 4 strata.
a) Biggest cities: Vilnius, Kaunas, Klaipėda, Šiauliai, Panevėžys, Alytus, Kapsukas (currently Marijampolė).
b) Lithuanian districts were divided into 6 strata according to the ethnographic regions of Lithuania: The first stratum – Žemaitija (Samogitia), which included the following districts: Akmenė, Joniškis, Kelmė, Klaipėda, Kretinga, Mažeikiai, Plungė, Skuodas, Šilalė, Šilutė, Tauragė, Telšiai. The second stratum – Aukštaitija: Anykščiai, Biržai, Ignalina, Kupiškis, Molėtai, Panevėžys, Pasvalys, Rokiškis, Ukmergė, Zarasai districts. The third stratum – Central Lithuania: Jonava, Kaunas, Kaišiadorys, Kėdainiai, Pakruojis, Radviliškis, Raseiniai districts. The fourth stratum – East Lithuania: Trakai, Šalčininkai, Širvintos, Švenčionys, Vilnius districts. The fifth stratum – Dzūkija: districts of Alytus, Varėna. The sixth stratum – Suvalkija: Lazdijai, Kapsukas (currently Marijampolė), Prienai, Šakiai, Vilkaviškis districts.
c) Resort towns: Birštonas, Druskininkai, Neringa, Palanga.
d) Specific town: Sniečkus (currently Visaginas).
The sample was divided into strata proportionally to their population.
Second stage. In each stratum a number of clusters (cities and districts) was selected.
a) All the biggest cities were included into the sample.
b) From each stratum of districts, districts were selected on the basis of the share of population 60 years and older in the district. After calculating squared deviations from the mean, the most typical 10 districts (clusters) – deviating the least from the mean – in each stratum of districts were selected: In the stratum of Žemaitija (Samogitia): the districts of Tauragė, Telšiai and Skuodas. In the stratum of Central Lithuania: the districts of Kėdainiai and Radviliškis. In the Aukštaitija stratum: the districts of Kupiškis and Molėtai. In the East Lithuania stratum – Šalčininkai district. In the Dzūkija stratum – Varėna district. In the Suvalkija stratum – Šakiai district.
c) From the strata of resort towns, on the basis of the same criterion (least squared deviation from the city proportion of the population aged 60 and over), one cluster was selected – Druskininkai.
d) Sniečkus (currently Visaginas).
Third stage. In each cluster of the second stage, a certain number (proportionally to the population) of third stage clusters, identified as electoral districts, was chosen.
Electoral districts belonging to 19 clusters – selected on the second stage (7 biggest cities, 10 districts, the resort town of Druskininkai, and Sniečkus (currently Visaginas)) based on the administrative unit boundaries – were selected. Prior to the selection of electoral districts in Vilnius, Vilnius was divided into 7 strata, from each of which a cluster (district) was selected (7 districts in total); Kaunas was divided into 5 strata and in each of them 1 district was selected (5 in total); the other major cities were divided into 3 strata and in each of them a cluster (electoral district) was selected (3 districts in total). In the biggest cities stratum, electoral districts were systematically selected from the numbered lists of electoral districts. In the district strata, the selection criterion was the distance of the electoral districts from the center of the administrative units. The furthest, the average and the closest electoral districts were selected. The sample was proportional to the population of rural areas, towns, administrative units’ centers and cities. A total of 76 electoral districts were selected.
Fourth stage. In each cluster of the third stage, a certain number (proportionally to the population) of voters – respondents was selected.
The potential respondents were systematically sampled from the electoral rolls of the electoral districts. The sampling interval was calculated by dividing the total number of voters in the electoral district by the number of respondents expected to be surveyed. Electoral rolls were usually sorted by street (not alphabetically), so this method of selection ensured a random selection of respondents. The names and addresses of the persons to be interviewed were written on special cards. |