Sampling

• Pages: 17-19

Sampling

Benefit s of gender equality th rough infrastruct ure provision: an EU-wid e survey 17

4. Sampling

4.1. S c o p e

Respondents from the EU-28 took part in this

study. The design was based on one sample of

5 378 respondents (residents over 18 years old).

The sample was designed to be representative

at EU level.

Daily and monthly updates are published on the

Eurostat website and the study data on Euro-

pean population were taken from Eurostat on

15 September 2014. Although these data rep-

resent a temporal estimation at the time of

extraction, this does not result in a methodo-

logical problem because dierences in times of

extraction are not signicant when designing

a thorough sample.

The expected nal sample was 5 378 respond-

ents but was increased to 5 385 to accommo-

date four additional respondents in Germany,

two in Portugal and one in Poland.

The universe comprised the total population,

rather than the 18-and-over (≥ 18) population,

as infrastructure and/or services are used by

the general population. However, ethical and

legal reasons required respondents to be over

18 years of age to participate in the sur vey. The

share of people aged ≥ 18 in the total popula-

tion was very similar across the EU-28, account-

ing for approximately 80 %.

4.2. Sample design

The sample in every Member State (MS) was

the general population of both sexes, aged 18

and above (≥ 18), and residing in the EU Mem-

ber States. In order to address the sur vey to the

selected populations, a stratied probabilistic

sample (populations/regions, sex and age) was

carried out, with a stratication of units from

the rst and second stages, selecting an inde-

pendent sample within ever y population (MS).

The units of the rst stage were populations/

regions, while those of the second stage were

sex and age variables.

In every region (organisational level above the

local entity of every MS, nomenclature of ter-

ritorial units for statistics (Eurostat) (NUTS) 2),

the units of the rst stage were stratied by

age and sex. The allocation among stages and

strata was strictly proportional. The strategic

variables of stratication for the sample were

determined as follows:

• Stratum 1: Regions (NUTS 2) of the MS

under consideration.

There are more than 200 MS regions across

the EU. Given that the regional dimension is

not as important in all MS, a feasible alternative

could include the regional dimension in Member

States with a federal or highly (de jure and/or de

facto) decentralised state s tructure, e.g. Spain,

Belgium, United Kingdom, Germany and Austria.

• Stratum 2: Sex and age.

The stratication of age distribution was 18-39,

40-6 4, and ≥ 65.

Although not stratication criteria, account was

taken of the following.

• At least 25 % must live in rural areas. This

can vary depending on the rural population

weight within the total population in every

MS, with a maximum uctuation of 10 %.

• At least 35 % must be employed workers

(self-employed and employees).

4.3. Data collection

Data were collected through a telephone-based

survey addressed to the general popula-

tion of the Member States. The information