We use analytics cookies to understand how the site is used. They stay off unless you accept them. Cookie Notice
An open survey for reach and a separate random sample for the country score
For partner organizations Take part
The Women’s Life-Cycle Economy Index (WLCE Index) plans two ways of collecting answers.
The open survey: a woman answers online or in person through a partner organization, in a language available in her country, and her answer helps show the range of women’s experience. A separate random sample in each study country, about 2,000 women as a guide: only its answers will feed the survey part of the country score. The pilot will test this design.
| Reach | Estimates | |
|---|---|---|
| What it is | Open-survey questionnaires, online and offline, that passed quality checks | Countries where a separate random sample has been drawn, with a minimum size (counted after weighting) set in advance |
| What it gives | Rare situations, subgroups, problems women describe in their own words, and women who agree to be asked again in later years (the panel of returning participants) | The country’s score, its uncertainty interval and its range of possible positions |
| What it cannot show | A country’s place in the comparison, and how common a situation is among all the country’s women | Rare situations: a sample of about 2,000 holds too few of them |
The open survey gives reach: questionnaires that passed quality checks, which show a wide range of life situations. The random sample gives the country score its survey part. That part is then compared with the analysis of rules and with the international benchmarks that the methodology will fix (for example, at least 14 weeks of maternity leave under ILO Convention 183).
The estimate will be published with measures of its uncertainty and the limits of the data, and a separate check will show how much a country’s place would change under other acceptable ways of calculating the Index.
A country will enter the comparison once checks against criteria published in advance confirm that its data is of good quality, that there is enough of it and that it can be compared with other countries; until then it will receive a country profile and a map of Unconsidered Points (an Unconsidered Point is a point in a woman’s life cycle where the system did not see her when it designed the rule).
If some women find it easier to get into the survey than others, the result tilts toward their experience, and more answers do not by themselves remove that tilt.
In an open survey no one knows who saw the invitation and who did not.
The statistician Xiao-Li Meng estimated the selection bias in polls before the 2016 US election and showed by calculation that, with that bias, even 2.3 million answers, about 1 percent of eligible voters, would carry the same error as a simple random sample of about 400 people. Weighting (adjusting how much each group of women counts in the overall result) narrows the gap but does not close it.
For the WLCE Index this matters even more: the reason a woman answers or stays away (caring for relatives, discrimination she has lived through, access to the internet) is exactly what the Index measures. That is why the country score counts only the answers of the random sample.
Each study country will have its own random sample (a country probability sample). For each country the methodology will define which women the study covers and how selection is done; every woman in that group will have a known chance of being drawn, as in a lottery. As a guide, a country’s sample is about 2,000 participants; the exact size will depend on the selection scheme, the expected response rate and the precision needed. The limits of coverage and the sample size achieved will be published with the results.
Telephone interviews where enough people have a telephone; in-person interviews at randomly selected addresses or households in all other cases; online survey panels only where the panel members were themselves selected at random.
Sample participants are selected by a scheme set in advance. Open invitations through Womafreesm and partner organizations belong to the open survey; a partner organization interviewing for the sample follows the sample’s protocol.
The random sample answers the same core questions as the open survey, so in every study country the results will show how far the open survey’s answers diverge from those of the random sample, both before weighting and after it.
How many answers the sample is really worth after weighting, how the selection scheme and weighting affected precision, and what share of the women invited actually answered: all published before the country joins the comparison.
The model comes from programs with decades of experience: the European Social Survey, the World Values Survey, the Gallup World Poll and the Demographic and Health Surveys work with random samples whose size is set by the research task and the precision required.
The open survey does a different job from the random sample. Among its participants, it will help:
find and describe rare life situations that occur too seldom in a sample to be studied
see small groups: say, women returning to work after having a child, in one region and one industry
spot, through answers in women’s own words, problems that were not among the questionnaire’s original assumptions
invite women who agree to be contacted again into the panel of returning participants
build the base of women’s experience at scale
For each analysis the composition of the group, the way it was recruited and the limits of the conclusions will be stated: the number of such answers does not by itself show how common a situation is among all the women of a country. Where a group is too small, “insufficient data for a reliable estimate” will be published.
Data will come through two channels: online through the program’s platform and offline through partner organizations, women’s associations, health centers and local venues.
Offline collection is planned so that the chance to take part depends less on internet access. Such answers will feed the survey part of the country score only if the women were selected by the random-sample scheme; answers from women a partner invited on its own will add to the base of women’s experience.
Safety measures will be set for both online and offline collection, taking the conditions of participation into account; for questions about violence, the safety rules will draw on the World Health Organization’s recommendations for research on domestic violence against women (2001). For in-person interviews, a separate offline collection protocol will set how sites are chosen, how interviewers are prepared and cleared to work, how the privacy of the conversation is ensured, how data is encrypted and consent is obtained, and when an interview is stopped if safe participation is not possible.
A language is chosen by asking who would be left out without it.
At a minimum, every language native to at least 5 percent of a country’s population is included (the threshold the European Social Survey applies); a smaller group’s language is included too if without it a significant share of women in a particular life situation would be missed. Every national version will be translated by a team, reviewed and approved, and tested with women in that country, and the decisions on key terms will be written down. The check must establish whether women understand the meaning of the questions and answer options in the same way across languages. Before the survey starts, each woman will see the languages available in her country.
A questionnaire counts toward the program’s total when all the conditions are met together: the woman is 18 or over, she gave consent, she reached the end of the main questionnaire, and the questionnaire passed quality checks, including the checks for duplicates, bots and fraud.
The program is designed for 2,000,000 questionnaires, and questionnaires from all collection paths count toward it: the open survey, the random samples and the repeat waves.
A participant’s questionnaire in a later wave is not a duplicate, so the number of women who answered is published next to the number of questionnaires. Skipped questions do not stop a questionnaire from counting: declining the sensitive questions is recorded as the woman’s decision, and rules to be published will set how gaps are handled in each analysis. Unfinished questionnaires do not count; within the participant’s consent, they help show where the questionnaire needs improvement. The rules for checking and exclusion will be published; how answers are protected is described on the ethics page.
The pilot will also show whether the collection plan is feasible: the pace of recruitment, the cost, the share of completed questionnaires, the reach across different groups of women and the safety of participation. The scaling plan will be refined based on its results. The pilot tests the instrument and is not the first edition of the Index.
Status of the WLCE Index and the pilot
Before a country enters the comparison, the program will publish how the data was obtained, how it was checked and how precise the estimates are.
who was invited, how the random selection was done and how the selected women were contacted
the number of participants and how much the selection method and weighting affect the precision of the estimate; the technical report states the effective sample size and the design effect
what share of those invited answered in the random sample; for the open survey, separately: how many women opened, started and completed the questionnaire
by what rules the contribution of different groups of women to the overall result was adjusted (this is called weighting); the rules are fixed before the relevant stage of collection begins
which checks the questionnaires went through and by what rules answers were included in the analysis or excluded from it
how the data was collected and processed, which methods were applied and which limits matter when reading the results
No, and the Women’s Life-Cycle Economy Index (WLCE Index) does not claim it. The program is designed for 2,000,000 quality-checked questionnaires across all collection paths: they will make rare situations and different life paths visible. The survey part of a country’s score is computed from a separate random sample, because in an open survey no one knows who saw the invitation.
Select your Telegram language