TNS Gallup Channel: Description of Sample

Sample description 1993–1998

The sample of the years 1993–1998 is demographically and regionally representative of all Finnish households, excluding only the Swedish-speaking Åland Islands. The sample composition was continuously controlled and updated. For a nation-wide sample, multistage stratified sampling was used. The first stage consisted of dividing the number of households required by the sample regionally by province. In the next stage, the number of households in each province was divided by municipality type. The municipality type classification was the following: urban municipalities were divided into big and small cities according to population size, and rural communities into three categories according to their industrial structure. The selection of municipalities was representative. The household sample can be said to correspond to a sample of persons.

When selecting households for the sample, Income Distribution Statistics of the year 1988, provided by Statistics Finland, were used as comparison material. The observation unit in the Income Statistics is household. The selection and updating criteria for the sample was household size (number of persons residing in the same dwelling), and socio-economic status of the head of household, in the following manner:

Households by size, whole country
Age under 65 %
1 person480,19728
2 persons459,56327
3 persons330,40119
4 persons302,78718
5+ persons128,1448
Total1,701,812100
Households by the socio-economic status of the household head
Age under 65%
Self-employed or entrepreneur213,14213
Upper-level employee with administrative, managerial, professional or related occupations274,34916
Lower-level employee with administrative and clerical occupations388,44623
Worker561,08133
Pensioner160,4119
Other104,3836
Total1,701,812100

Sample description 1999–2010

In the years 1999–2010 the GallupChannel sample represented the population of Finland aged 15 or over, excluding the Åland Islands. The sample composition was continuously controlled and updated. Samples were selected using three-stage stratified sampling.

The first stage consisted of dividing the number of households required by the sample regionally by province. In the next stage, the number of households in each province was divided by municipality type (urban municipalities, semi-urban municipalities, rural municipalities). Furthermore, urban municipalities were divided into big and small cities according to population size.

All big cities in Finland, i.e. cities with over 30,000 inhabitants, were automatically included in the sample. The cities in question are Helsinki, Espoo, Vantaa, Kauniainen, Pori, Turku, Hämeenlinna, Lahti, Tampere, Kotka, Lappeenranta, Kuopio, Jyväskylä, Vaasa, Joensuu, Oulu, Rovaniemi and Mikkeli.

The selection of municipalities was representative. For each municipality (except for urban municipalities), the difference from a hypothetical average municipality was calculated using the following variables:

  1. Support (%) for the four major political parties (calculated from the number of inhabitants entitled to vote).
  2. Average income of the household, and mean age of family members aged 15 or over.
  3. The share in the population of the municipality of farmers, workers, upper- and lower-level employees, pensioners, and persons with upper secondary level education at minimum.

The household sample can be said to equal a sample of persons. In the third stage of the sampling process, Channel households were selected from each sampled municipality in proportion to population size.

When selecting households for the sample, Income Distribution Statistics of the year 1998, provided by Statistics Finland, were used for municipality level comparison, as they have households as observation units. The selection and updating criteria for the sample were: household size (number of persons residing in the same dwelling), and socio-economic status of the head of household, in the following manner:

Households by size, whole country
%
1 person36.5
2 persons30.8
3 persons14.1
4 persons11.7
5 persons4.7
6+ persons2.1
Total100
Households by the socio-economic status of the household head
%
Farmer2.2
Self-employed or entrepreneur4
Upper-level employee with administrative, managerial, professional or related occupations11.5
Lower-level employee with administrative or clerical occupations17.1
Worker19.9
Student3.1
Pensioner30.3
Other11.9
Total100
Population by municipality type
%
Urban municipalities60.1
Semi-urban municipalities16.5
Rural municipalities23.4
Total100.0

The selection criteria for a GallupChannel household: household size and the occupation, level of education and age of the head of the household (=person with the highest income). Households were recruited both by telephone based on a random sample, and by mail using a sample drawn from the Population Register. The Population Register contains information which enabled sampling persons with specific characteristics. In the Helsinki metropolitan region, households were recruited in proportion to the number of inhabitants in the province of Uusimaa with the telephone area code 09.

Distribution of responses on the TNS Gallup Channel on different days 1999–2010:

Time% of households responding on average
Friday~15%
Saturday~20%
Sunday~33%
Monday~24%
Tuesday~8%

The household sample was constantly controlled and updated. To replace households that had withdrawn from the sample, new households were recruited weekly. The background information of new households was collected immediately. Background information of the other households was updated once a year (in August). About one third of the respondent group changed annually. The response rate was kept at 80% of the sampled households. If the rate fell below 80%, a reminder phone call was made to non-respondent households. If notified in advance, a respondent household was able to temporarily withdraw from the sample. Therefore only those households which were reachable in a particular week were counted in the number of households in the sample. Because of the special nature of the data collection system, response rates were not calculated for respondents.

Sample description 2011–

From 2011 onwards, the TNS Gallup Channel and later from 2023 onwards, the Kantar Forum Channel sample represents the population of Finland aged 15 or over, excluding the Åland Islands. The sample composition is continuously controlled and updated. Samples are selected using three-stage stratified sampling.

The first stage consists of dividing the number of persons required by the sample by major region (NUTS2). In the next stage, the number of individuals in each major region is divided by municipality type (urban municipalities, semi-urban municipalities, rural municipalities). Furthermore, urban municipalities are divided into big and small cities according to population size.

All big cities in Finland, i.e. cities with over 30,000 inhabitants, are automatically included in the sample. The cities in question were Helsinki, Espoo, Vantaa, Kauniainen, Pori, Turku, Hämeenlinna, Lahti, Tampere, Kotka, Lappeenranta, Kuopio, Jyväskylä, Vaasa, Joensuu, Oulu, Rovaniemi and Mikkeli. In the third stage of the sampling process, persons are selected from each sampled municipality according to gender and age, in proportion to population size.

When selecting individuals for the sample, the most recent population statistics, provided by Statistics Finland, are used for municipality level comparison. The selection and updating criteria for the sample are a person's age, socio-economic status and education level.

The basic selection criteria for a Kantar Forum Channel member: major region, age, gender, socio-economic status and level of education. Members are recruited both based on a random sample and by using a sample drawn from the Population Register that allows for correcting bias in the Channel sample. Recruitment is carried out via telephone and recruitment campaigns on the Internet.