Eurostat, the statistical office of the European Union

F4: Quality of life

2920 Luxembourg

The main purpose of the EU survey on gender-based violence against women and other forms of inter-personal violence (EU-GBV) is to assess the prevalence of violence in order to address the requirements of the Istanbul Convention. The survey covers psychological, physical and sexual violence by intimate partner, physical and sexual violence by non-partner, sexual harassment at work, violence experienced in childhood and stalking by any perpetrator.

The data collection for the first wave (year 2021) is planned to take place between 2020 and 2023 in the EU countries, based on their national timetables. This means that Eurostat will release the data for groups of countries as soon as data are available and have been validated.

The disseminated indicators focus on violence by perpetrator, disaggregated by type of violence, by time of occurrence, by age and by personal characteristics of the respondent; and on frequency, severity, seriousness and reporting of the experienced violence.

However, it is necessary to point out that survey data might only be a close proxy to real prevalence as survey data depends on the willingness of the respondent to disclose any violence experienced. Therefore, to understand the prevalence of violence and disclosure rates by survey respondents, it is important to take into account the extent to which violence is tolerated in the wider community. For example, in cultures where people are ready to talk about their painful experiences, their answers may reflect more accurately their own experiences rather than community norms. However, in less aware communities it may be that people might reflect socially desirable norms. To provide some background on country specific context, few indicators on commonness and awareness of support services are disseminated.

It is essential to avoid using sensitive terms that could cause anxiety or concern when introducing the survey. Accordingly, the general recommendation was that the survey name should be neutral when contacting the respondents. The aim was to avoid alerting any perpetrators of domestic violence to the nature of the survey or frightening off any victims of violence, in order to minimise non-response, as some respondents might be discouraged from taking part if the name of the survey included terms like ‘assault’, ‘sexual violence’, or ‘gender-based violence’.

Majority of countries have followed this recommendation and the title of the survey was translated as survey on health, safety or security and well-being or living conditions; quality of life or relationship survey. Only few countries (BG, SK) used gender-based violence in the title of the survey during data collection and explained that this decision was taken as no issue appeared during testing the survey using the word “violence”, or the word "violence" was used in order to avoid misunderstanding regarding the aim of the survey and to reduce non-response due to the fact that respondents were not aware of the real theme of the survey.

However, the pilot survey results indicate that respondents understood the rationale for the choice of neutral survey name once they had been given an explanation, and agreed that it was right. Due to the sensitivity of the topic, the participating countries were strongly encouraged to include experts on violence against women and/or gender-based violence as well as psychologists and psychotherapists in every step of the survey - from the preparation, through the field work to the data dissemination.

Majority of countries included experts on the topic in the project team: gender statisticians, gender-based violence or violence against women researchers, policy experts, psychologists, social workers, experts working on victim support or NGOs, experts on victimization surveys. External experts were included in the preparation of the survey, training of the interviewers and in order to provide support to the interviewers.

Few countries (MT, FI) established the focus group or expert group consisting of different experts in the field and providing the support to the survey during all phases.

The country codes conform to the ISO 3166 (International Organisation of Standardisation), and regional codes to the NUTS 2 classification. Educational level of respondent and her partner are based on ISCED 2011. Variables on respondent occupation and economic activity are optional, however, if collected, the classification of occupation uses ISCO-08 (two-digit) and the classification of economic activity uses NACE Rev. 2 (two-digit).

Statistical concepts and definitions are specified in the ‘Methodological manual for the EU survey on gender-based violence against women and other forms of inter-personal violence (EU-GBV)’.

Violence by type of perpetrator

Violence by intimate partner covers psychological, physical (including threats) and sexual violence. Intimate partners are persons with whom a respondent has or had an intimate relationship.

Prevalence of intimate partner violence is calculated out of women who have ever been in intimate partnership.

Violence by non-partner, domestic perpetrator and any perpetrator covers physical (including threats) and sexual violence. Non-partners are all perpetrators with whom a respondent does not have or has never had an intimate relationship. Domestic perpetrator includes intimate partners and family members, and additionally any other individual living or having lived in the same household as the respondent at the time of one or more violent events.

Type of perpetrators of non-partner violence are defined according to relationship:

• ‘Family member or relative’ includes blood relatives like parents and children, and other blood relatives that can be cohabitating or non-cohabitating, as well as other household members or relatives by marriage or adoption (e.g., siblings, grandparents, aunts, uncles, cousins, nephews, nieces, in-laws, etc.);
• ‘Non-family or relative but other known’ as friends, family friends, schoolmates, colleagues, co-workers, supervisor, boss, professor, teacher; any person with some authority as army or police officer, priest, religious leader, judge, doctor; any other person known;
• ‘Any known’ includes family member or relative or other known explained above;
• ‘Stranger’ is someone completely unknown to the respondent.

Type of violence

Psychological violence by intimate partner includes a range of behaviours, encompassing acts of emotional abuse and controlling behaviour towards the respondent: belittling and humiliating; forbidding the respondent to see friends or family, or to engage in hobbies or other activities; tracking the respondent via GPS, phone or social network; forbidding the respondent to leave the house without permission or locking the respondent up; constantly accusing the respondent of being unfaithful or getting angry if the respondent speaks to another person; forbidding the respondent to work; controlling the finances of the whole family and the respondent’s personal expenses; keeping or taking away the respondent’s ID card/passport to control the respondent; yelling and smashing things or behaving in a certain way with the aim of scaring or intimidating the respondent; threatening to hurt the respondent’s children or other people close to the respondent; threatening to take away the respondent’s children or to deny custody; and threatening to harm them self if the respondent leaves them.

Physical violence refers to a range of violent types of behaviour or acts involving harm and fear, such as threatening to harm the respondent; pushing or shoving the respondent, pulling their hair, slapping or throwing something at them; punching the respondent or beating them with an object; kicking; burning (with fire or acid or by some other means); trying to choke or strangle the respondent; threatening to use or actually using a knife, gun, acid or something similar; or using force against the respondent in some other way with the aim of hurting them.

Sexual violence includes unwanted sexual intercourse through force or physical violence or by exploiting a situation in which the respondent is not able to refuse sexual intercourse because they are under the influence of alcohol or drugs. It also includes unwanted sexual intercourse which the respondent is too afraid to refuse and cases in which the respondent is forced into unwanted sexual intercourse with another person or persons. Attempts to carry out any of the above acts or any other unwanted sexual behaviour that the respondent finds degrading or humiliating are included as well. Finally, this type of violence also covers unwanted sexual touching by non-partners. 

Sexual violence in the childhood covers experiences before the age of 15 perpetrated by any person, such as posing naked in front of another person, unwanted sexual touching or sexual intercourse. Type of perpetrators are defined as for non-partner violence.

Sexual harassment at work covers the following unwanted behaviours with sexual connotations that happen in a work context: inappropriate staring or leering; being exposed to sexually explicit images or videos; indecent sexual jokes or offensive remarks about a person’s body or private life; inappropriate invitations to go out on a date or suggestions for sexual activity of any kind; unsolicited physical contact; inappropriate advances on social networking websites or sexually explicit emails or text messages; threatening with unpleasant consequences if sexual proposals or advances are refused; and any other similar behaviour with sexual connotations that took place at work or work-related settings and that offended, humiliated, or intimidated the respondent. Types of perpetrators are defined according to relationship and authority/power:

• ‘Co-worker’ is a person with whom one works, typically someone in a similar role or at a similar level within an organisation: colleague or fellow worker.
• ‘Boss’, is the job title of a management position that is primarily based on authority over a worker or being in charge of a workplace: supervisor, overseer, facilitator, monitor or area coordinator, etc.
• ‘Other work related person’ is someone with whom the respondent has a professional relationship but who does not work in the same workplace as the respondent (e.g. client, customer, patient, student, passenger, etc.).

Prevalence of sexual harassment at work is calculated out of women ever worked.

Stalking includes a range of offensive or threatening forms of behaviour or acts repeated in the course of the respondent’s life: sending unwanted messages, including messages on social media, emails and letters, or gifts; making obscene, threatening, nuisance or silent telephone calls; trying persistently to contact with the respondent, waiting or loitering outside the respondent’s home, school or workplace; following or spying on the respondent in person; intentionally damaging the respondent’s things (car, motorbike, letterbox, etc.) or the belongings of people close to the respondent, or harming animals belonging to the respondent; making offensive or embarrassing comments about the respondent in public, including comments on social networks; publishing photos, videos or highly personal information about the respondent. Types of perpetrators are defined according to relationship as intimate partners and non-partners.

Repeated violence

Repeated violence (series of episodes) refers to similar violent episodes repeated by the same person(s), during which similar thing(s) are done under the same circumstances more than once. For instance, a woman might be beaten by her partner in several episodes over a period of three years.

Prevalence of one time violence and prevalence of repeated violence is calculated for each type of violence by intimate partner (out of women who have ever been in intimate partnership); by non-partner; by any perpetrator. Frequency of violence (one time or repeated) is also calculated for sexual harassment at work (out of women ever worked). As stalking is defined as repeated offensive or threatening behaviour, repetition (once or more) is not relevant.

Reporting or telling to other people about violent experiences

Estimating unreported violence is an important aspect in defining targeted policies to eradicate violence. Indicators on reporting of violence show whether the respondent talked with anyone about the experienced violence.  Reported violence experienced by intimate partner or by non-partner (to whom it was reported) includes:

• ‘Close person’ that may be family or friend and should be understood in its widest meaning.
• ‘Health or social service’ includes reporting or talking to any doctor or nurse, or any other health, medical or social worker.
• ‘Support service’ means any organisation or official body providing help to victims of violence, e.g. support services, helplines, shelters.
• ‘Police’ covers the situation when respondent herself reported to the police any episode of experienced violence.

For sexual harassment at work and sexual violence in childhood to whom it was reported includes:

• ‘Official body’ covers reporting or talking to any support, medical or social service or worker; police; official at work (sexual harassment at work) or someone official in the school as teacher, social worker, psychologist (sexual violence in childhood).   
• ‘Unofficial’ means talking to family and friends; colleague (sexual harassment at work) or classmate (sexual violence in childhood).

For experiences on stalking, reporting covers only offensive or threatening forms of behaviour done by the last perpetrator and to whom it was reported includes:

• ‘Police’ covers the situation when respondent herself reported to the police any episode of stalking by last perpetrator.
• ‘Legal or victim support service’ includes asking help from lawyer or contacting victim support service.
• ‘Support service or police’ covers both above: police, legal or support service.

Share of women reported or talked about experiences of violence by to whom is calculated out of women experienced this type of violence.

Consequences of experienced violence

The consequences of violence are an important consideration both for the victim and for society. Regarding violence experienced by intimate partner or non-partner, indicators for following consequences are disseminated:

• ‘Physical injury’ is damage to the body caused by external force as cuts, scratches, bruises, burns, broken bones or fractures, head, internal or genital injuries, miscarriage, pregnancy due to rape or any other type of injury due to physical or sexual violence.
• ‘Felt that their life was in danger’ means that the respondent felt that her life was in danger when any of violent episodes happened.

Share of women experienced this type of violence and consequences due to this is calculated out of all relevant women: for intimate partner violence out of women ever been in intimate partnership; for non-partner and any perpetrator out of all women.

Regarding the consequences due to stalking, the focus is on more serious actions that were taken by the respondent due tostalking:

• ‘Changed telephone number/email address or closed social network’ covers the situations when respondent had to change telephone number, change email address or close social network account.
• ‘Implemented protecting measures’ means that respondent stopped going out alone, changed the usual route to get to work, school or university, started to take something with her to protect herself, such as scissors, knife, pepper spray.
• ‘Moved to another residence or changed working or studying’ covers the situations when respondent had to change her place of residence, job, school, university or had to stop working or studying.

Share of women had to do take this kind of measures due to stalking is calculated out of women experienced stalking.

Awareness of support services

Awareness of support services includes knowledge of specific free of charge helpline available 24/7, other helpline, specific shelter free of charge, other shelters, or any other support service available for the victims of violence.

Awareness of free legal aid means the knowledge about this service available for persons experienced sexual abuse.

Countries were requested to assess the availability of national support services and provide the exact name of services in the question.

Personal characteristics

Indicators are disseminated by following personal characteristics:

• Disability status is measured through a concept of general activity limitation (limitation in activities people usually do because of health problems for at least the past 6 months).
• Educational attainment level refers to the highest level in the international standard classification of education (ISCED) which an individual has successfully completed.
• Degree of urbanization of the area where a person has their usual residence is classified in local administrative units at level 2 as cities, towns and suburbs, or rural areas, based on the share of local population living in urban clusters and in urban centres.
• Country of birth is defined as the country of usual residence of the respondent’s mother at the time of the respondent’s birth, (according to the current national boundaries and not to the boundaries in place at the time of birth). The main categories include: native-born (born in reporting country); foreign-born in EU country; and foreign-born in non-EU country or country of birth unknown.

Indicators are calculated as share of women with certain personal characteristic (e.g. severely limited) who experienced violence (e.g. by intimate partner) out of all women with this personal characteristics (e.g. severely limited).

Timeframe of violence

Violence experienced in adulthood covers violence experienced by a non-partner since the age of 15 and intimate partner violence during a person’s lifetime.

National difficulties of implementation of the common EU-GBV questionnaire

Countries were requested to translate and adopt the common EU-GBV survey questionnaire in their national context: relevant languages to be used and software used to collect the data. The translation and adaption of the questions, terms, definitions and concepts was done via pretesting. In general, countries did not report particular difficulties in translating the definitions and concepts. However, few countries had to find proper translation for the word ‘episode’.

Issues appeared with implementation of the questionnaire:

• BG used ‘112 EU Emergency number’ in the question on knowledge of specific free of charge helpline available 24/7, accordingly, indicator ‘Women as victims/non-victims of violence, by awareness of the existence of support services (gbv_awr_serv)’ is disseminated with flag ‘d’ (definition differs).
• BG informed that around half of interviews were conducted in private, that may impact disclosure the violence by respondents and therefore, the prevalence rate of violence.
• DK had an issue with implementation of the routing and therefore not all relevant questions were asked when applicable regarding repeated violence by intimate partner. Therefore, indicators are not disseminated for reporting and consequences of intimate partner violence, and accordingly of violence by any perpetrator.
• ES explained that there is no free legal aid for victims of sexual abuse available. Therefore, indicators on awareness of free legal aid are not available.
• FR had an issue with implementation of the routing and therefore not all relevant questions were asked when applicable regarding repeated violence by non-partner. Therefore, indicators are not disseminated for reporting and consequences of non-partner violence, and accordingly of violence by any perpetrator.
• PL had an issue with implementation of the routing and therefore not all relevant questions were asked when applicable regarding non-partner section. Therefore, indicators by occurrence, frequency and consequences are not disseminated for non-partner violence, and accordingly for violence by any perpetrator.
• PT had an issue with implementation of the routing to the question on consequences of stalking and accordingly, the indicator is not disseminated.
• SK had an issue with implementation of the routing and therefore not all relevant questions were asked when applicable regarding in depth questions of intimate partner and non-partner violence, sexual harassment at work and sexual violence experienced in the childhood. Therefore, indicators are not disseminated for occurrence of the last episode by non-partner; frequency, consequences and reporting for intimate partner, non-partner and violence by any perpetrator.
• XK explained in the report that one reason for drop out the survey was due to husband or another household member did not allow to participate in the survey. This may increase the underestimated prevalence rate.
• XK used ‘helpline’ and ‘shelter’ instead of real list of existing national services for the questions on awareness of support services. As the indicator takes into account knowledge of any service, this should not have significant impact.
• XK territory is divided to Urban and Rural area by degree of urbanization, therefore indicators on prevalence of violence by personal characteristics (degree of urbanization) is not disseminated.

Statistical unit of EU-GBV survey is individual. Only one person per household can be interviewed and proxy is not allowed. The reason is the safety of the respondent and the interviewer, as it is not uncommon for the abuser to be part of the victim’s household.

The EU-GBV results cover individual persons, residing in the territory of the Member States and participating countries at the time of data collection. Persons living in collective households and in institutions are excluded from the target population as individuals living in institutions (e.g., hospitals, prisons or religious institutions) are often not listed in the household/person survey frames and this makes it difficult to contact them.

The target population of the EU-GBV survey is persons aged 18-74 who live in private households, with a focus on women. However, men can be included in the target population by countries willing to do so. As a result, around half of countries included men, but indicators are calculated and disseminated only for women.

Table 1. Gross sample size and included population.

* Information for Italy is not provided as aggregated data are provided based on national violence against women survey.

In order to get EU-wide results, survey has to cover all EU-27 Member States. 18 Member States implement the EU-GBV survey nationally (BE, BG, DK, EE, EL, ES, FR, HR, LV, LT, MT, NL, AT, PL, PT, SK, SI, FI). Italy agreed to share the data based on its national Violence Against Women survey. For the remaining eight Member States (CZ, DE, IE, CY, LU, HU, RO, SE), the European Institute for Gender Equality (EIGE) and the European Union Agency for Fundamental Rights (FRA) are implementing a separate data collection following the Eurostat methodological manual.

Concerning the survey implementation in other countries, Iceland implements the survey nationally while, under the Instrument for Pre-Accession Assistance (IPA) as part of the multi-beneficiary statistical cooperation, four countries (ME, MK, RS an XK) implement the EU-GBV survey and two other countries (BA and AL) conducted some methodological work.

Current dissemination includes the countries for which data collection was coordinated by Eurostat. The data for entire EU is planned to be disseminated at the later date, including countries covered by the EIGE/FRA data collection and national data collection. Given the current state of play, it is likely that results for full EU will be disseminated during first half of 2024.

Data collection for the first wave covers the 2020-2023 period. The duration of the fieldwork was recommended as four-months period based on national timetables (wave 2021). However, the sensitive nature of the survey means that staff may encounter many refusals and may need to make return visits, ring people back, or even do re-sampling and therefore, fieldwork period could be prolonged in order to ensure high quality data.

Table 2. Fieldwork period.

 *Information for Italy is not provided as aggregated data are provided based on national violence against women survey.

Results of the EU-GBV survey are disseminated in prevalence rates (e.g. % of women in relationships who experienced intimate partner violence).

The reference time of experienced violence is divided into the last 12 months, 1-5 years ago or earlier than 5 years ago, according to the occurrence of the last episode.

Data covering experiences over the last year and the last five years can give an indication of the extent and the nature of current levels of violence and an estimate of the number of people who may require help. Lifetime experiences, by contrast, provide an indication of the total number of people ever affected by such forms of violence.

To develop efficient and effective policy and legal responses to limit/stop gender-based violence, accurate data are necessary. Thus, the Gender Equality Strategy of the European Commission for 2020-2025 calls for comprehensive, updated and comparable data for policies on combating gender-based violence. The EU-GBV survey is not covered by an official EU legislation and the first wave of the data collection in 2020-2023 takes place on a voluntary basis.

Data collection on gender-based violence is implemented based on the Commission implementing decision on the financing of the Programme for single market, competitiveness of enterprises, including small and medium-sized enterprises, and European statistics and the adoption of the work programme for 2021-2024. Even more, Regulation (EU) 2021/690 of the European Parliament and of the Council of 28 April 2021 requests to provide high quality, timely and reliable statistics to support the European Pillar of Social Rights and the Union Skills Policy, including statistics on the labour market, employment, education and training, income, living conditions, poverty, inequality, social protection, gender based violence, undeclared work, and satellite accounts on skills.

Data on GBV are collected in order to address the requirement of Istanbul Convention (Council of Europe, Convention on preventing and combating violence against women and domestic violence, Istanbul 2011). The Gender Equality Strategy announced measures to obtain the same objectives as the Istanbul Convention. Therefore, on 8 March 2022, the Commission adopted a legislative proposal on combating Violence against Women and Domestic Violence.

Regulation (EC) No 223/2009 on European statistics (recital 24 and Article 20(4)) of 11 March 2009 (OJ L 87, p. 164) stipulates the need to establish common principles and guidelines. It ensures the confidentiality of data used for the production of European statistics and the access to those confidential data, with due account for technical developments and the requirements of users in a democratic society.

Publication rules:

• estimates should not be published if based on fewer than 20 sample observations or if the non-response for the item concerned exceeds 50%;
• estimates should be published with a flag if based on 20 to 49 sample observations or if non-response for the item concerned exceeds 20% and is lower or equal to 50%;
• estimates shall be published in the normal way when based on 50 or more sample observations and non-response for the item concerned is lower than 20%.

Flags:

u     unreliable (i.e. due to small sample size)

d     definition differs, see metadata

Flag ‘d’ is used:

• in table ‘Women as victims/non-victims of violence, by awareness of the existence of support services (gbv_awr_serv)’ for BG as country used ‘112 EU Emergency number’ in the question on knowledge of specific free of charge helpline available 24/7 and therefore, definition is not fully comparable with other countries
• data for IT are based on national violence against women survey.

Indicators based on EU-GBV survey data are published currently for 18 Member States (BE, BG, DK, EE, EL, ES, FR, HR, LV, LT, MT, NL, AT, PL, PT, SK, SI, FI) and three other countries (ME, RS, XK) where data collection was coordinated by Eurostat. Data for IT are based on national violence against women survey. Given the current state of play, it is likely that results covering full EU will be disseminated by second half of 2024.

Please refer to the Release calendar - Eurostat (europa.eu) publicly available on the Eurostat’s website.

In line with the Community legal framework and the European Statistics Code of Practice, Eurostat disseminates European statistics on Eurostat's website (see section 10 - 'Accessibility and clarity'), respecting professional independence and in an objective, professional and transparent manner in which all users are treated equitably. The detailed arrangements are governed by the Eurostat protocol on impartial access to Eurostat data for users.

Not applicable.

The following publications are disseminated in the EU-GBV dedicated section: https://ec.europa.eu/eurostat/web/gender-based-violence:

Methodological manual

Statistical report 

Please consult: Database - Eurostat (europa.eu)

Please consult: Database - Gender-based violence - Eurostat (europa.eu)

Not applicable.

Website: http://ec.europa.eu/eurostat

Information is provided in the Methodological manual for the EU survey on gender-based violence against women and other forms of inter-personal violence (EU-GBV) (available here) as well as in Statistics on crime and criminal justice dedicated group in CIRCABC.

Not applicable.

Common questionnaire and the methodology have been developed for the EU-GBV survey that are followed by the countries in order to implement the survey. Please see more details in the Methodological manual for the EU survey on gender-based violence against women and other forms of inter-personal violence (EU-GBV).

EU-GBV survey questionnaire and methodology are developed based on the results of the pre-testing and pilot survey. Testing was conducted in 14 countries over the period 2018–2019.

Standardisation is achieved with the common questionnaire and common list and content of target variables. This is complemented by Eurostat consistency and integrity checks on the micro data. In addition, countries should report to Eurostat any deviation from the standard.

EU-GBV statistics have overall high quality. National surveys are considered as reliable sources applying high standards with regard to the methodology. However, the EU-GBV, like any survey, is based upon a sample of the population. The results are therefore subject to the usual types of errors associated with random sampling. Based on the sample size and design in the various Member States and participating countries, Eurostat implements basic guidelines intended to avoid publication of figures that are unreliable and to give warning for low reliability.

The relevance of an instrument has to be assessed in the light of the needs of its users. As for the EU-GBV survey results, the main users are the following:

• institutional users like DG JUST of the Commission and other Commission services;
• statistical users in Eurostat or in Member States and participating countries;
• end users - including the media - interested in gender-based violence against women

Eurostat carries out a general Eurostat User Satisfaction Survey every year to collect feedback on the quality of its statistics. The survey is usually addressed to the registered Eurostat users who are mainly students, academic, private users, business, government and international organizations.

EU-GBV results cover national territories of the participating countries.

Data were to be based on a nationally representative sample of the population residing in private households within the country, irrespective of language, nationality or legal residence status. The sampling frame and methods of sample selection were left to the countries’ experiences with other population surveys in which they could ensure that every individual in the target population was assigned a known and non-zero probability of selection. The main condition that needed to be fulfilled is that only one person can be interviewed per household.

The indicative effective sample size defined for the EU-GBV is 5000 respondents per country. However, countries are allowed to increase or decrease the sample size in order to collect high quality data.

Table 3. Gross and net sample size for women.

 *Information for Italy is not provided as aggregated data are provided based on national violence against women survey.

Standard errors of key indicators are commonly used as a measure of the reliability of data collected through sample survey. EU-GBV survey was designed to provide measure of prevalence of violence during lifetime by women aged 18 -74 years. The indicative effective sample size was defined taking into account this accuracy requirement.

A non-sampling error is an error in survey estimates which cannot be attributed to sampling fluctuations. Such errors can either be coverage errors, measurement errors, non-response errors, processing errors or model assumption errors.

Coverage errors arise due to divergences that exist between the target population and the sampling frame.

Measurement errors occur at the time of data collection. There are a number of sources for these errors, such as the survey instrument, the information system, the interviewer and the mode of collection.

Data collection is based on the common questionnaire, definitions and concepts. Due to sensitivity of the topic, safety and well-being of survey participants and interviewers was highlighted in the methodological manual and countries were requested to report methods used in the metadata.

It was highlighted in the methodological manual that only one member of the household should be interviewed alone in safe and secure conditions in order for both the respondent and interviewer to be safe and have no distractions while conducting the interview. Not feeling safe or being in the same place with the potential abuser makes the interview impossible to conduct and therefore leads to refusals to participate to the survey.

Different data collection modes were allowed for this data collection, however, face-to-face and computer-assisted methods were most recommended and possibility to provide self-completion was highlighted. Advantages and disadvantages for each data collection mode were provided in the methodological manual. Countries were recommended to use most appropriate method in national context, combined methods were allowed.

Table 4. Data collection mode.

 *Information for Italy is not provided as aggregated data are provided based on national violence against women survey.

Non-response errors arise due to an unsuccessful attempt to obtain the desired information from an eligible unit. Two main types of non-response errors are considered:

• unit non-response, which refers to the absence of information of whole units (households and/or persons) selected into the sample;
• item non-response, which refers to the situation where a sample unit has been successfully enumerated, but not all the required information has been obtained.

All surveys have to deal with non-response, i.e. information missing for some of the sample units.

Based on the national metadata, total response rate (not taking into account impact of data collection mode) is calculated as share of completed and accepted interviews (with women) out of gross sample size (of women).

Table 5. Response rate for interviewer assisted and not assisted modes.

 *Information for Italy is not provided as aggregated data are provided based on national violence against women survey.

Notes:

FR: response rates are provided based on phase 2;

NL: explained that CAWI interruption includes also cases when respondent just opened the questionnaire;

PT: due to sample of dwellings, non-response rates are calculated based on total respondents (male and female);

Item non-response typically happens to questions the interviewee does not answer because they consider them personal or not easily understandable. Regarding EU-GBV survey, several countries (DK, FR, PL, PT, SK) had issues with implementing the routing properly in the interviewing software due to complexity of the questionnaire. In case the question was applicable but not asked due to routing issues, the value 9 – not known is used in the variable.

In the EU-GBV survey, non-response was not allowed for several questions:

• most important background questions of respondent and her partner
• questions that were used to consider if section of the questionnaire was applicable, e.g. current activity status (for section on ‘Sexual harassment at work’) or actual marital status (for section on ‘Intimate partner violence’)

However, due to nature of the national survey, Portugal allowed non-response for all questions.

Secondly, as the main purpose of the survey is to measure the prevalence of violence, the set of screening questions for each type of violence covered in the survey could not be completely unresponded (at least one screening question had to be responded as ‘yes’ or ‘no’ in each set of screening questions). In case the rule was not met, the interview was not accepted in the database. Therefore, item non-response regarding screening questions is not relevant when disseminating the results. However, item non-response is provided in order to illustrate the sensitivity of the screening questions.

Item non-response is provided for the variables used to calculate disseminated indicators. Composite variable used for calculation of the indicators is considered as not responded if all relevant and applicable variables are not responded. For example, ‘occurrence of intimate partner violence’ is considered as not responded if respondent experienced violence only by current partner and occurrence of last episode by current partner is not responded or respondent experienced violence only by former partner and occurrence of last episode by former partner is not responded or respondent experienced violence by both partner and occurrence of last episode by both partners are not responded. Item non-response is calculated out of all applicable respondents. For example, ‘occurrence of intimate partner violence’ is calculated out of all respondents who experienced violence by intimate partner.

Information for Italy is not provided as aggregated data are provided based on national violence against women survey.

Processing errors occur in post-data-collection processes such as data entry, keying, editing and weighting.

Not applicable.

Not applicable.

Not applicable.

A high level of comparability across the participating countries is namely ensured by:

(a)    the use of the same definitions for all countries;

(b)    the transmission to Eurostat of the same list of variables with the same coding;

(c)    the use of common classifications;

(d)    the central validation of data done by Eurostat.

EU-GBV survey data collection follows international standards: ISCO, NACE, ISCED, degree of urbanisation, standardised variables for social surveys.

Not available

Not applicable.

Not applicable.

The EU-GBV is a random sample survey of persons living in private households. The sampling units are dwellings, households or individuals depending on the sampling frame. Different schemes are used to sample the units, ranging from the simple random sampling method to complex stratified multi-stage sampling methods of clusters.

Participation in the survey is voluntary for all participating countries.

Not yet agreed.

There are no constraints on the most appropriate mode of data collection, which can vary from country to country. Nevertheless, due to the sensitive nature of this survey, and research and data-based findings, most recommended mode is face-to-face and computer-assisted methods over other methods. The possibility of self-completion is also supported.

Prior to the dissemination of transmitted national data, Eurostat checks the data quality and consistency. Validation rules are provided in the Eurostat methodological manual.

Not applicable.

Not applicable.

No notes.