Girls Leadership Index: a real-life example of a composite index development
Plan International is an independent development and humanitarian organisation that advances children’s rights and equality for girls. In 2019, they commissioned the development of the Girls Leadership Index (GLI) for 19 countries in the Asia region. In 2021, the index for Asia was updated and a new one for the Pacific (14 countries) was developed. This year, the indices for both the Asia and Pacific regions are being updated. The main objective for the development of the GLI was to use it as an instrument to influence policy makers, donors and members of the public to invest more in girls’ development, empowerment and ability to realise and exercise their rights at regional and national levels. I have been one of the three members of the team that has developed and updated the GLI. The three of us have different skills and knowledge, and I am the statistician of the team.
Why develop an index? The general idea behind the development of a (composite) index is to summarise a complex, multi-dimensional concept into one single adimensional value, which is easier to interpret than a battery of many individual indicators and simpler to analyse. We can then use the index to rank units (countries in this case), compare their performance over time, or compare them to other well-established indexes such us the Human Development Index (HDI), the Gender Development Index (GDI) or the Gender Inequality Index (GII).
Plan International originally defined six domains for the Asia Index: Education, Economic opportunity, Safety, Health, Political participation and Civil liberty, and National law and policies. In 2020, we updated the index for Asia and developed a new one for the Pacific to include Climate as a seventh domain. Finally, in 2022 the Climate domain will also be included in the Asia region.
The first part of the process was the desk research stage, which involved a comprehensive review of existing literature and research on girls’ and young women’s empowerment and leadership, related key concepts and theories, and the enabling and constraining factors. The review considered international and regional evidence, research and studies carried out in the past 10 years by research institutions, multi-lateral agencies, and international and national non-governmental organisations.
The statistics work included the selection of indicators in each domain, obtaining the data from reliable secondary sources, dealing with missing data and imputation, weighing, computing the normalised indicators, the domain indices and the overall index and data visualisation.
The results can be analysed from different points:
1. As a picture of the current state of the girls’ and adolescent women empowerment and ability to exercise their rights
Figure1: 2021 Pacific GLI
2. Looking at each domain
Figure2: Education and Economic Opportunities domain indexes for 2021 Pacific
3. Looking at trends
Figure3: GLI values for 2019 and 2021 Asia
Debates and technical challenges
Though the construction of a composite index requires sound statistical principles, it is largely determined by the particular application, including both formal and heurist elements, and incorporates expert knowledge on the matter. Some decisions must be made that require both kinds of experts: statisticians and subject matter. Here are some of the ones we found most challenging:
- • Choosing the indicators. Once the main dimensions, or domains, are defined, which indicators should be chosen to measure them? Let’s think of Education or Economic opportunities. There are lots of possible indicators to be chosen, some measure similar things, some different, some are available and periodically updated, while others are not. And then some can be found in reliable databases such as the World Bank, WHO or UNICEF, while others cannot. A few may have missing values, and some may not.
- • Dealing with missing values: sometimes we can impute or estimate the value using other variables, but at other times this won’t be possible.
- • Weighting: Once we have decided which indicators will be used, should we assign each one with the same weight or a different weight? Likewise, should we assign the same weight to all domains or a different weight?
Finally, some may ask whether an index like this is useful at all. We may see that the index does not change, but this may mean that nothing has changed or that some indicators have experienced an improvement, while others have gotten worse. We still think that the index is useful, obviously it summarises a lot of information into a single value and not all details can be identified. But when constructing the index, we also have six/seven domain indices to look at, in addition to a number of individual indicators.
Plan International has presented the index in a number of regional forums as an advocacy tool. It raises awareness of, and advocates for actions to support girls’ development, empowerment and their ability to realise and exercise their rights. The Index is an instrument that can be used by national governments and their development partners to monitor and measure the advancement of girls’ leadership and development as part of the sustainable development agenda.
From a technical perspective, we still don’t know how fast the index will detect abrupt changes. For instance, two events happened in 2021 that should have had a large effect in the GLI, or at least in some domain:
1- The Taliban takeover of Afghanistan with immediate effect in adolescent girls and young women, who have been barred from education and jobs, with changes in laws
2- The tsunami in Tonga, which should affect the Climate domain
We already know that none of these will be detected in the 2022 GLI. We are looking forward to knowing when this will happen.
References: links to the 2020 and 2021 PLAN GLI reports
• 2021 Asia – Pacific Report: https://plan-international.org/2021-asia-pacific-girls-report
Author: Alex Riba
Alex is an engineer with over 20 years of experience teaching statistics and conducting research. Having worked as a statistician on a wide range of projects, he is particularly interested in processes that let data speak for itself, especially in meaningful ways for non-statisticians.
0 comments for "Girls Leadership Index: a real-life example of a composite index development":
Add a comment:
We run an anonymous commenting system. If you are not logged in, we do not collect any information on who you are when you leave a comment. This means we manually confirm comments before they appear on the site.
If you want to have a comment you submitted deleted, please contact us, giving the date of the comment and name of the article.