European Institute for Gender Equality (EIGE)

EIGE’s Gender Mainstreaming programming area

Not applicable.

Not applicable.

European Institute for Gender Equality, Gedimino pr. 16, LT-01103 Vilnius, Lithuania

EIGE.SEC@eige.europa.eu

+370 5 215 7444

22/09/2020

In 2015-2016, the European Institute for Gender Equality (EIGE) conducted a restricted survey in the 28 EU Member States (EU-28) on the benefits of gender-sensitive infrastructure.

EIGE’s work in this field is based on the conceptual framework and methodology to assess the benefits of gender equality through the provision of public infrastructure, developed by Dr Gloria Alarcón García and Dr José Colino Sueiras (University of Murcia, Spain). EIGE conducted a restricted survey in the 28 EU Member States (EU-28) on the benefits of gender-sensitive infrastructure. Dr Gloria Alarcón García lead the survey design, and the EIGENET framework of contractors was engaged in the survey implementation, involving 23 organisations and consortiums. The later analysis of the survey results was carried out by a research group led by the University of Murcia, Spain. Authors include Dr Gloria Alarcón García (principal researcher and research group leader), Dr Tindara Addabbo, Dr Edgardo Ayala Gaytan, Dr Estela Fernandez Sabiote, Dr María Rubio Aparicio, Dr Andreu Castellet, Dr Angela O’Hagan, Dr José Manuel Mayor. EIGE’s Gender Mainstreaming team coordinated the work. Dr Vaida Obelene, Dr Paula Franklin and Dr Irene Riobóo Lestón provided quality assurance and carried out further analysis of data.

The survey asked 5,378 European women and men about nine infrastructure services and covered seven activities of daily life.

The infrastructure services covered were:

1. Nurseries 0-3 years
2. Nurseries 3+ years (until mandatory school age)
3. Health services and medical centres
4. Care for elderly people
5. Care for people with disabilities
6. Public transport
7. Sidewalks and footpaths
8. Parks
9. Streetlights

Activity domains:

1. Mobility
2. Education
3. Physical and mental health
4. Leisure
5. Employment
6. Domestic and care work
7. Social relations

The survey aimed to collect direct information on the importance of existing infrastructure services for everyday activities, and the level of well-being that public infrastructure provides

Results are presented at the EU28-level and where possible at the country-cluster level. The clusters are defined as follows:

• North-western: Austria, Belgium, France, Germany, Ireland, Luxembourg, United Kingdom
• Southern: Croatia, Cyprus, Greece, Italy, Malta, Portugal, Spain
• Eastern: Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Slovenia
• Nordic: Denmark, Finland, Netherlands, Sweden

Not applicable

Most questions in the survey ask respondents to assign a score between 1 and 10 to given aspects of their use of public services. For instance, the score represents the respondents’ level of satisfaction with a service, assessment of the quality of the services, and so on. For this type of questions, the indicators presented in the Gender Statistics Database are the average scores computed in each country-cluster or at the EU28 level.

For questions that require yes/no answers, the indicators are percentages of respondents answering “yes”.

All results are broken down by sex and whenever possible, by country-cluster (see point 3.2 for the definition of the country clusters). For selected indicators, a breakdown by age, education or hours of unpaid care work (per week) is presented.

Age groups:

• 18 years or over
• 18 to 39 years
• 40 to 64 years
• 65 years or over

Education groups:

• Less than primary, primary and post-secondary non-tertiary education (ISCED 2011 levels 0-4)
• Tertiary education (ISCED 2011 level 5-8)
• All ISCED 2011 levels

Hours of unpaid care work per week:

• Less than 21 hours per week
• More than 21 hours
• Total

The unit of analysis is individuals aged 18 years or older in each country

The universe comprised the total population, rather than the 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 survey. The share of people aged 18+ in the total population was very similar across the EU-28, accounting for approximately 80%. 

EU28   

2015-2016

Not applicable

Not applicable

Not applicable.

Not applicable.

EIGE’s survey on gender equality and public infrastructure does not contain any administrative information such as names or addresses that would allow direct identification. Anonymity is preserved in the aggregate data

Not applicable.

Not applicable.

The full dataset, with all survey results aggregated at the country-cluster level can be made available to users upon request to EIGE.

No regular new releases.

As part of the study, the following publications were produced:

• Technical report: Benefits of gender equality through infrastructure provision: an EU-wide survey
• Factsheet on gender equality and care infrastructure
• Factsheet on gender equality and urban mobility
• Factsheet on gender equality and health service infrastructure

Not applicable.

Individual level data are not available to external users. The full results of the survey aggregated at the country-cluster level can be made available to users upon request to EIGE.

Not applicable.

The methodology used to conduct the survey is described in the Technical Report “Benefits of gender equality through infrastructure provision: an EU-wide survey”, available at this link.

The quality checks performed on the survey data are described in the Technical Report “Benefits of gender equality through infrastructure provision: an EU-wide survey”, available at this link.

During the implementation phase, data collection specifications were followed to ensure that the data collected were sound, robust and of the highest quality.

Quality assurance on comparability required clear definitions and statements to be followed:

• Coverage of sample, representing the target population regionally and across types and size and comparable across countries
• Non-response computed according to one internationally accepted standard across all participating countries
• Non-response bias estimated in a standard fashion across all participating countries
• List of sampling management systems used, indication of risk due to possible differences in fieldwork across countries due to use of different sampling management systems
• List of Compute Assisted Telephone Interview (CATI) systems used; indication of risk in using different CATI systems
• List of the interview staff composition (age, sex, education) for each country indication of risk due to different staff composition
• Documentation of translation process and outcomes; indication of risk involved due to translation or adaptation

Data checks were carried out on the pre-test data then on the soft launch. Completed interviews were checked during and after fieldwork. Phone files of the population under study had to meet established quality standards, including checking for missing or erroneous data, filtering, interview duration, duplicates, straight liners, cross-consistency, outliers and non-responses.

The data collection system was computerised. Fieldwork and sample quotas were closely monitored throughout, with regular data checks during the fieldwork.

Interviewers attended an in-depth briefing. Supervisors then conducted quality controls during the fieldwork by listening to interviews and providing the interviewers with feedback. Thorough data cleaning took place after the fieldwork.

An important indicator of the survey quality is the effective response rate (RR): higher RRs ensure more accuracy in the survey. The RR was satisfactory in most cases (see also point 13.3).

During the assessment stage, data were validated through an in-depth statistical analysis, focusing on the following aspects (see also point 13 for further details):

• Error detection and debugging data file
• Analysis of missing values
• Data filtering
• Analysis of reliability, validity and robustness.

Overall, the data from EIGE’s survey can be considered of good quality, based on the standard criteria outlined in points 12-15.

The results of the survey are relevant for the following categories of users of the Gender Statistics Database:

• Academic or policy researchers in gender equality. The results of the study can be combined with other characteristics at the country-cluster level to develop original research on gender equality and public infrastructure.

The final dataset for dissemination contains, besides cluster-level variables derived from the survey questionnaire, a System of Context Indicators (SCI) at the country-cluster level, belonging to the following areas: Income and income distribution, labour market, gender, transport, social expenditure, crime and corruption, environment, culture, violence against women. These could be used for further research.

• EU, national, regional and local policymakers involved in urban planning and the design of public infrastructure policies. The results of the study show, for instance, that the incorporation of a gender perspective in urban planning and public infrastructure policies can contribute to the advancement of gender equality, in line with one of the fundamental principles of gender mainstreaming. 
• Civil society organisations, e.g. women rights organisations. The results can be used to support communication and awareness-raising campaigns on the issue of the gender-sensitiveness of public infrastructure. 
• European social partners, who are active in promoting the improvement of working conditions for women and men. The results can be used to describe how public infrastructures can improve work-life balance
• EIGE can further use these data as a starting point for further studies and data collections on the same topic.

Not applicable.

Results are aggregated at the country-cluster level whenever possible. Indicators computed on a selected sample (e.g. indicators computed from filtered questions) are shown only at the EU28 level, as the cell size for the computation of sex-disaggregated indicators at the country cluster level was below the reliability threshold of 50 recommended by Eurostat (see https://circabc.europa.eu/sd/a/a31fe118-3b05-4d85-b5b2-a5c4b5511fc9/Guidelines%20for%20Publication.pdf).

Overall accuracy was checked in an in-depth statistical analysis conducted after the data collection. The analysis checked for:

• Errors and inconsistencies of responses (see also point 13.3): the data collected were debugged, verified, and edited to ensure that there was no error or inconsistencies.
• Percentage of missing values: The percentage of missing values was below 20% for most variables
• Sampling error due to data filtering (see also point 13.2): Data filtering increased sampling error, although this remained below 5% for most variables
• Reliability, i.e. the overall consistency of a measure, a measure is said to have a high reliability if it produces similar results under consistent conditions: Multi-item measure reliability (measured with Cronbach’s Alpha) was high. Following a well-known procedure, corrected item-total correlation was also checked. Based on the results, it can be concluded that the multi-item measures are highly reliable
• Content validity, i.e. the extent to which a measure represents all facets of a given social construct: As the questionnaire was developed by subject matter experts, the different concepts measured in the questionnaire could reasonably be expected to have content validity.
• Construct validity, i.e. the degree to which a test measures what it claims, or purports, to be measuring: Construct validity can be assumed to be high, as variables measuring similar concepts (e.g. happiness and satisfaction) were highly correlated
• Robustness of the results to a change of the sample: the results showed that the elimination of outliers from the sample did not change the results of the survey

The overall sampling error is the survey was 1.4%.

The results selected for publication in the Gender Statistics Database may have higher sampling error (especially those based on filtered questions) but always below the maximum threshold of 5%.

The study checked for non-sampling errors as part of the quality control measures described in points 11.1 and 13.1. As a general indication of accuracy of the survey results (and indirectly of low non-sampling error), the technical report provided an analysis of effective response rates (RR).

RR was computed using two measures.

The first measure (RR1) focused on the Member States where the interviews were conducted, and it is the ratio of completed interviews to the total number of respondents eligible to take part in the survey in each Member State.

RR1 = [(Complete)/ (Complete + Refusal + Person not available + Partly completed)] * 100

Complete - subjects that completed the survey successfully

Refusal - those who refused to participate, either immediately after picking up the phone or right at the beginning of the questionnaire

Person Not Available - eligible respondents who were not available at the moment of the survey

Partly Completed - incomplete questionnaires due to an interruption or the respondent’s refusal to continue or request to be contacted later

RR1 ranged between 81% and 4% across countries.

The other way to calculate the effective RR focused on the questions of the survey. Here, RR was calculated by dividing the number of completed interviews by the number of interviews attempted, to obtain.

RR2 = [(Number of interviews completed)/(Number of interviews)]*100

This formula was applied to each question of the survey in order to determine the success rate of each question. RR2 was very high for all survey questions (over 85%)

Not applicable.

Not applicable.

The underlying survey data were produced using a common questionnaire across 28 EU Member States, hence the results can be considered comparable across countries.

Not applicable as the survey was conducted only for one year.  

The results obtained in this survey were compared to those obtained from similar questions in the fourth European Quality of Life Survey (EQLS) carried out by the European Foundation for the Improvement of Living and Working Conditions (Eurofound). Focusing on how respondents rated health and transport surveys in the two surveys, EIGE’s survey provides higher rates, but the ranks by sex and cluster are exactly the same in both sources. Therefore, their consistency is assured.

The estimates based on EIGE’s survey data have full internal coherence, as they are based on comparable microdata and they are calculated using the same estimation methods.

Results from the Study have been thoroughly quality-assured by the study team and validated through the consultation meeting (9-10 March 2016, in Vilnius). Further data analysis and revision (in depth statistical analysis mentioned in point 13 above) was conducted upon the completion of the study.

There is no fixed revision schedule.

Data were collected through a telephone-based survey addressed to the general population of the Member States. The information was obtained through a telephone conversation between the interviewer and the eligible respondent. The programme chosen, Computer Assisted Telephone Interviews (CATI), makes random calls and allows the interviewer to manage and complete the questionnaire with computer aid. This type of telephone-based survey provides an economic advantage by requiring fewer interviewers, appropriate monitoring, providing fast and efficient access to substantial population diversity. All versions of the CATI-system were valid for the data collection.

Not applicable

The information was obtained through a telephone conversation between the interviewer and the eligible respondent. The programme chosen, Computer Assisted Telephone Interviews (CATI), makes random calls and allows the interviewer to manage and complete the questionnaire with computer aid. This type of telephone-based survey provides an economic advantage by requiring fewer interviewers, appropriate monitoring, providing fast and efficient access to substantial population diversity. All versions of the CATI-system were valid for the data collection

Data validation procedures have been described in point 11.1 and 13.1. They are also detailed in the methodological report available at this link.

Scope of the survey

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 of the overall EU28 population (as reported by Eurostat as for 15 September 2014).

The expected final sample was 5,378 respondents but was increased to 5,385 to accommodate four additional respondents in Germany, two in Portugal and one in Poland.

The universe comprised the total population, rather than the 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 survey. The share of people aged 18+ in the total population was very similar across the EU-28, accounting for approximately 80%. 

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 countries. In order to address the survey to the selected populations, a stratified probabilistic sample (populations/regions, sex and age) was carried out, with a stratification of units from the first and second stages, selecting an independent sample within every population (MS).

The units of the first 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, Unit NUTS 2), the units of the first stage were stratified by age and sex. The allocation among stages and strata was strictly proportional. The strategic variables of stratification for the sample were determined as follows:

• Stratum 1: Regions (NUTS 2) of the MS under consideration. MS regions can be more than 200 across the EU. Given that the regional dimension is not as important in all MSs, a feasible alternative could include the regional dimension in countries with a federal or highly decentralised state structure, e.g. Spain, Belgium, UK, Germany and Austria.
• Stratum 2: Sex and age. The stratification of age distribution was 18-39, 40-64, and 65+. Although not stratification 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 fluctuation of 10%. 
  • At least 35% must be employed workers (self-employed and employees).

Weighting

In this survey-based research, the selected sample size was proportional to the size of the adult population in every participating country. In particular, in each country, the sample size for the survey was 0.0001 % of the population size. The sample size of each country thus acts as an implicit weighting factor in the analyses and it is not necessary use additional weights in the microdata to ensure that the sample is representative of the EU.

Not applicable