Ethics, participant safety and data in the WLCE Index

Who reviews the research before it starts, how consent works and how your answers are protected

See how to take part How you are kept safe

Three conditions before a single question is asked

Taking part in the Women’s Life-Cycle Economy Index (WLCE Index) is voluntary: you can skip a question or stop, and the sensitive questions open only with your separate consent.

The questionnaire opens in a country after three conditions: ethics review and permits, a data protection impact assessment and, for the questions about violence, a verified map of help services.

Slide 1/3

Ethics review and permits

Before any data is collected, the requirements for ethics review and permits will be defined and met for the protocol and for each country. The main protocol will be reviewed by an ethics committee independent of Womafreesm, such as a partner university’s committee. Fieldwork will also be reviewed in every country where it takes place, with the permits that the country requires. The Index Results Register will show the status of preparation and the decisions received.

Ethical approval comes from outside Womafreesm

The main study protocol will go to an independent ethics committee outside Womafreesm: either a partner university’s committee, acting for Womafreesm under a written agreement, or a recognized independent committee. The committee will be chosen in 2027, before data collection begins.

Fieldwork in each country will also go through that country’s own review, with whatever permit or local partner institution the country requires. The independent committee will be appointed before fieldwork begins, and its appointment will be recorded in the Index Results Register as a separate milestone.

Womafreesm’s own Ethics and Data Protection Board will write the documents and carry out the decisions. Approval rests with the external committee.

Index Results Register

The main questionnaire

Whether to answer the main questionnaire at all; the other seven decisions are additional. She can skip any question in it and stop at any point.

The sensitive questions

Whether to open the separate part about violence, health and money. That part has its own consent, its own risk review and its own access rules.

Later survey waves

Whether to answer again in a later wave, as part of the panel of returning participants. Retention periods for repeat waves will be set separately.

Keeping her contact details

Whether the study may keep a way to reach her for a follow-up, and which way is safe for her. Contact details are stored apart from her answers.

Recording an interview

Whether the interview, if there is one, may be recorded. She is asked this before the interview, at the moment when the question becomes relevant.

Quoting her words

Whether her words may be quoted in publications. Before a quotation is used, it is checked for the risk that someone could recognize her from it.

Future studies

Whether her answers may be used in future studies within the topics, aims and access rules of the consent text; any other use requires new consent.

Access for partner researchers

Whether partner researchers may work with her answers, with codes in place of names, in the secure research environment. They do not receive a copy.

Before any answers, the consent text says what stays confidential in her country and when the law requires otherwise: in some countries the law obliges researchers to report certain facts, and the text names those cases.

It also explains what happens if she later decides to leave the research: no one from the research will contact her again, her contact details will be deleted, and a limited administrative record of the withdrawal itself will be kept, with a stated basis and retention period.

Answers that can still be linked to her through the key are removed from the research dataset at her request; the rule for answers already included in verified calculations will be fixed before the pilot. Published combined results cannot be changed: individual answers are not distinguishable in them.

Sensitive questions follow their own rules

The sensitive questions cover violence, symptoms dismissed by doctors, control over her money, harassment, and situations where she felt unsafe. They form a separate part of the survey, the sensitive module, with their own consent, their own risk review, their own data class (see the table below) and their own access rules.

Neutral wording

Invitations and notifications look like an ordinary message, and the app has an ordinary name with no hint of the topic. For her this means: notifications do not reveal what the questionnaire is about. On a shared or someone else’s device, protection is limited by the device itself: whoever controls the phone can see what was opened on it.

A safety check first

Before the sensitive questions open, the survey asks three things: is she alone, is this a good moment, does she want to continue. For her this means: she answers only when she has chosen the time and the place herself. If the moment is wrong, the survey waits, and she can come back to it later.

Plain words, her own names for things

The questions ask what happened, in plain words. What happened is asked first; what the clinic, the employer or the police did about it is asked separately. For her this means: she describes events in her own words, without having to call them “violence” or any other legal term, and no one judges her words.

Skip and leave

Any question can be skipped. The section can be left at any point with one tap, and the questionnaire will not return to it. For her this means: she decides what to talk about. Skipping a sensitive question does not by itself exclude the rest of her answers from the analysis.

A quick exit

If someone enters the room, one tap instantly replaces the screen with an ordinary neutral page. For her this means: the person who came in sees an ordinary website. Quick exit switches the screen and does not protect against monitoring of the device or of internet activity; the browser keeps the list of visited pages, and it is worth clearing only where doing so is itself safe. What happens to answers already entered when quick exit is used will be stated in the conditions of participation.

Minimal technical data

The sensitive pages keep only what is needed to record an answer; tools that record clicks or the screen are switched off there. For her this means: what she typed and deleted is not kept. Which technical data is collected for quality control, such as completion time, will be listed in the conditions of participation.

If a question brings up something hard

A separate procedure is being developed for three formats: self-completion, a remote interview and an in-person interview. Online answers are not read in real time, and no one intervenes automatically: next to the sensitive questions there will be a verified contact of a help service in her country, and in an interview the interviewer knows where to stop and where to refer her. A service contact is a way to ask for help, not a guarantee of receiving it.

How the people who read her answers are protected

Interviewers, translators, the people who type up recordings, and the people who code and tag answers read the same material as the participant, and it can weigh on them too.

So they will be selected, trained through practice interviews, supported by an experienced supervisor, and given time to talk through hard interviews afterward. If the work becomes too heavy, any of them can step away from it at any time, and that is treated as a normal part of the procedure.

Every one of them will sign an agreement on how they handle the data. Before material passes from one person to the next, names are replaced with codes: that is what pseudonymized means on this page. A woman’s story in her own words is processed only in systems the program controls or has a signed data agreement with. That includes AI tools: the permitted tasks and data for them will be listed in the data agreement, and work with them takes place only inside the secure environment. Public chatbots and other outside services that may keep a copy of what is typed into them are not allowed.

Each data class has its own access rules

Class	What it holds	Who can access it
Operational	Name, contact details, consent records, permission to contact her again, checks against fake entries	The operations team only. Always stored apart from her answers
Linkage key	The only object that connects a person to her answers	One named keeper; two people must act together to use it. The most protected class of all
Sensitive	Sensitive answers, interview audio recordings and free-text accounts, locked with their own separate keys; before material passes to the next specialist, names, places and other direct pointers to a person are removed from it	Named analysts; each sees only the columns approved for her task and works inside the secure research environment, where local download and outside services without a data agreement are blocked
Pseudonymized research data	Answers with names, contact details and other direct identifiers removed	Approved researchers, inside the secure research environment, after the Data Access Committee approves the request and a data access agreement is signed
Public aggregates	Combined results for countries and regions, checked for the risk of identifying individuals	Everyone

The data is divided into five classes by use. Access is determined by the specialist’s role, the approved task and the amount of data needed; additional restrictions apply to sensitive information. The rules are the same for everyone.

Each class will have its own retention period, written down and built into the system. The periods for contact details, the linkage key, answers, audio recordings and service logs are set separately, for one-time participation and for repeat waves; they are among the rules to be settled before data collection, listed below.

Everything that protects the woman stays with Womafreesm: the store of identity details, the consent registry, the keys that link a person to her answers, the sensitive answers, the checks before publication and the secure research environment.

Why only combined results are published

A combined result is one figure for many women at once: for example, the share of women in a country who returned to the same position after parental leave.

One woman’s answers, even without her name, are still a record about a person. That is why the study calls such data pseudonymized rather than anonymous, and why what leaves the study is combined results, with quotations only under her separate consent and after a check of the risk of recognition.

Why a record without a name can still reveal who a woman is.

Picture one answer: country, age 52, midwife, small town, returned to work after caring for her mother. There is no name, but in a small town there is one such woman, and anyone who knows her will recognize her from those details. In a study about violence, menopause and money, that is enough to do harm. The rule below protects against exactly this.

It has happened to others.

In 2016, researchers posted 68,371 profiles from a dating site online, reasoning that since people had filled in the profiles themselves, the data was “already public.” The first version of the dataset held usernames and cities; they were later removed from it because of privacy concerns. In 2018, the Australian privacy regulator found that in a government health dataset released as de-identified, the identification numbers of health providers could be recovered: the encryption could be reversed. Both times the dataset had been judged safe to publish.

That is why only combined results are published, and every figure is checked before publication: could it point to one person? If a group is too small, for example a few women with a rare situation in one region, the figure is merged with a neighboring one or held back. Individual women’s records stay inside the secure research environment.

What will be published openly

Everything needed to check the Index will be open.
One thing stays closed: the answers of individual women.

Methodology and research protocol

How the Index is built and how the study is run, step by step. Published before data collection begins, so it can be criticized before there are results to defend.

Questionnaires

Every question, in every language version. Anyone can see exactly what women were asked and how, and compare the wording across languages.

Codebook

A dictionary for the data: what each column of the table means and what each answer code stands for. Without it, a table of numbers is unreadable, even for a specialist.

Calculation code

The program that turns data into the Index. Open code lets anyone study the algorithm; a full recalculation is possible only with access to the data in the secure research environment, and the practice dataset is used to test the code itself.

Combined results

Figures for countries and regions, each checked beforehand so that it cannot point to one person. This is the only form in which results leave the study.

Practice dataset (synthetic data)

The same columns and similar answers as in the real dataset, but every row is invented and there are no real women in it. It is used to learn to work with the data and to test the code, and it goes through the same disclosure-risk check as the combined results.

Researchers who need individual participants’ data can apply for access to it in the secure research environment.

Access opens after the Data Access Committee approves the request and a data access agreement is signed.

For researchers

What will be settled before data collection begins

Retention periods

Periods for each kind of data, from contacts to service logs, will be set before the pilot. Deleting the link after quality checks and keeping a coded dataset for repeat waves and reproducing calculations are compatible: the rule will say what is kept and how long.

The list of pilot countries

The pilot countries will be determined; the list will be published with the pilot methodology note. In each country on the list, ethics review and permits come before the first question; the questions about violence wait for a verified, dated map of help services.

What happens when answers are hard

For three formats (self-completion, a remote interview and an in-person interview), it will be written down before the pilot: whether answers are read, how to ask for support, who responds and how quickly, where to refer a woman and where the interviewer’s role ends.

Rules for in-person collection

The in-person collection protocol will be written and tested before fieldwork begins: how sites are chosen, how interviewers are selected, trained and cleared, how both the woman and the interviewer are protected and when the local copy of the data is deleted.

Partner researchers’ access and publication of quotations

Before the pilot, the terms on which partner researchers work with coded answers in the secure research environment will be fixed, along with the rule for publishing quotations: her own separate consent and a check of the risk that someone could recognize her.

Data location and the vendor

Before the pilot, the survey-service vendor, the geography of storage and access and the procedure for access from other countries will be determined, and vendor access, backups, deletion, logging and incident response will be fixed in the system requirements.

Several rules are being written now.

Each will be fixed before the step it is needed for, and each is recorded in the Index Results Register.

Index Results Register

FAQ

1. Who reviews the ethics of the study?

The main protocol will be reviewed by an independent ethics committee outside Womafreesm, to be chosen in 2027, before data collection begins; fieldwork in each country will go through national review. Womafreesm’s internal board will prepare the documents and carry out the decisions; approval stays with the external committee.

2. Can I skip a question or leave the sensitive module?