The 2023 UNESCO Institute for Statistics (UIS) series OE.5T8.40540.GPIA describes the gender balance of outbound internationally mobile tertiary students from each origin country who are studying in sub-Saharan Africa. The extract contains 171 country and area observations. It uses the adjusted gender parity index, or GPIA, which is bounded from 0 to 2 and is centered on 1.00. A value of 1.00 means that the female and male values of the underlying indicator are equal; values below 1 point to a lower female value, while values above 1 point to a higher female value.
The striking feature is not a smooth ranking from low to high. Almost the entire dataset is concentrated at three exact values. Of the 171 observations, 106 are exactly 1.00, 37 are exactly 0.00 and 27 are exactly 2.00. Saudi Arabia is the only observation between those points, at 0.37436. That structure changes how the indicator should be read: the useful question is less about tiny country-to-country differences and more about what the endpoints, the parity reference and the absence of underlying student counts actually mean.
Table of Contents
GPIA measures a relationship, not the number of students
The adjusted gender parity index is designed to compare female and male values on a scale that is symmetric around 1. The conventional gender parity index divides the female value of an indicator by the male value. When the female value is higher, that ordinary ratio can extend without an upper bound and is not symmetric around equality. UIS therefore uses an adjusted form that places the result on a 0-to-2 scale. The agency commonly treats values from 0.97 to 1.03 as the practical parity range.
For this specific series, the underlying subject is internationally mobile tertiary students from an origin country who study in sub-Saharan Africa. The country dimension is therefore most naturally read from the origin side. A GPIA below 1 indicates that the female value for that origin country is lower than the male value; a value above 1 indicates the reverse. A value of 1 means equality in the underlying female and male indicator values.
None of those statements tells us how many students are involved. A country with a GPIA of 1 could have a large flow or a very small one. Likewise, an endpoint value of 0 or 2 does not by itself indicate a large international-student population. GPIA summarizes the balance between the two sex-specific values, so a separate count series is required to measure volume.
About 62% of observations are in the parity band
Using the UIS 0.97–1.03 range, 106 of the 171 observations fall within the parity band. That is 62.0% of the dataset. Another 38 observations (22.2%) are below 0.97, and 27 (15.8%) are above 1.03. The median is exactly 1.00, while the unweighted mean is 0.938.
| GPIA range | Observations | Share | Reading |
|---|---|---|---|
| Below 0.97 | 38 | 22.2% | Female value lower than male value |
| 0.97–1.03 | 106 | 62.0% | UIS practical gender-parity range |
| Above 1.03 | 27 | 15.8% | Female value higher than male value |
The quartiles are also unusual: the 25th percentile, median and 75th percentile are all 1.00. That means the central half of the distribution collapses to the same exact value. In a typical continuous country indicator, an average and a ranked list might reveal a gradual pattern. Here, the count of observations at 0, 1 and 2 is more informative than a long decimal ranking.
This is why the mean of 0.938 should not be interpreted as a typical country having a modest male-leaning imbalance. The mean is pulled by the 37 zeros and 27 twos, while the majority of observations are exactly equal to 1. The distribution is highly discrete rather than smoothly centered around 0.938.
Why the exact endpoints matter
Under the adjusted GPI formula, 0 is the lower bound and 2 is the upper bound. An exact 0 can occur when the female value of the underlying indicator is zero while the male value is positive. An exact 2 can occur when the male value is zero while the female value is positive. In this 2023 extract, 37 observations are exactly 0.00 and 27 are exactly 2.00.
Those endpoint values deserve caution because the data supplied for this comparison do not include the sex-specific student counts. If one origin country sent only a very small number of students to sub-Saharan Africa, a zero for one sex can generate an extreme GPIA even though the absolute difference is small. Without the female and male counts, the index cannot tell us whether the endpoint represents a handful of students or a much larger flow.
For that reason, it would be misleading to label 0.00 or 2.00 as a ranking of the ‘largest’ gender gap in practical scale. They are mathematical boundaries of the adjusted index. They show direction and an extreme ratio condition, but not the number of people affected. Any interpretation of magnitude should be paired with underlying counts.
Saudi Arabia is the only intermediate observation
Saudi Arabia, at 0.37436, is the only observation in the 171-row extract that is neither 0, 1 nor 2. The value is below 1, so the female value is lower than the male value for this particular mobility indicator. It is also below the 0.97 lower edge of the UIS parity band.
Even here, it is better not to translate 0.37436 directly into a simple statement such as ‘female students are 37.4% of male students’ without checking the unadjusted values. The adjusted index uses a transformation when the female value exceeds the male value so that deviations are symmetric around 1. For values below 1 the adjusted and unadjusted ratios coincide, but the present dataset still does not provide the underlying sex-specific counts needed to describe the actual flow.
The fact that only one observation lies between the endpoints and equality also shows why a finely graded choropleth would add false precision. A map with many color classes would imply a continuum that the source data simply do not contain. A distribution chart communicates the structure more directly.
Examples of countries at 0.00 and 2.00
Examples at exactly 0.00 include Andorra, Bulgaria, Chile, Costa Rica, Czechia, Ethiopia, Ghana, Greece, Jordan, Nepal, Niger and South Africa. Examples at exactly 2.00 include Albania, Armenia, Azerbaijan, Bahrain, Georgia, Hungary, Liberia, Malaysia, Poland, Serbia, Uruguay and Uzbekistan. These are examples rather than a complete ranking.
The geographic spread of those examples matters because it cautions against attributing the endpoint pattern to one region alone. Countries from Europe, Asia, Africa and the Americas appear in both endpoint groups. The index alone cannot establish why a particular origin country has that result. Potential influences on international student mobility can include language, institutional ties, scholarships, cost, immigration rules, subject choice and historical connections, but none of those causes can be identified from the GPIA series by itself.
Why a distribution chart is more useful than a world map here
Country-level data often invite a world map, but geographic area is unrelated to the number of internationally mobile tertiary students. A large country filled with a strong color can dominate a map even when its underlying student flow is tiny, while a small island or territory can nearly disappear from view. That visual problem is especially serious when most values are exact endpoints or equality.
The distribution chart avoids that distortion. It shows immediately that 106 observations are in the parity band, 38 are below it and 27 are above it. It also keeps the focus on the number of observations rather than land area. For this indicator, that is a closer match to the information actually present in the source.
A map could still be useful as a locator if combined with student counts, destination detail or a richer set of intermediate GPIA values. With the current 2023 extract, however, the primary analytical value is the discrete distribution and the interpretation of its endpoints.
What the indicator can and cannot answer
The indicator can answer a narrow but useful question: for each reported origin country in 2023, was the female value lower than, roughly equal to, or higher than the male value among outbound internationally mobile tertiary students studying in sub-Saharan Africa? The GPIA scale and parity band provide a consistent way to classify that relationship across observations.
It cannot answer how many students went abroad, what share of all outbound students chose sub-Saharan Africa, which destination country they attended, which universities or fields of study they selected, or how long they stayed. It also cannot serve as an overall score for gender equality in a country’s education system. It summarizes one gender relationship for one very specific mobility indicator.
The distinction between a missing observation and an observed zero is also essential. The 171 rows used here are observations reported in the source. Other countries or territories absent from the extract should not be treated as zero. A missing value means that this comparison does not have a reported observation; an observed 0.00 is a real endpoint value published for the indicator.
How to use these results in a broader analysis
A stronger follow-up analysis would pair GPIA with sex-specific student counts. That would separate an endpoint produced by a very small flow from an endpoint involving many students. It would also make it possible to calculate the absolute difference in numbers and to see whether parity at 1.00 reflects substantial mobility or only a tiny count on both sides.
Destination-country detail would add another layer. The current series aggregates the destination region as sub-Saharan Africa, so it does not show whether students from an origin country are concentrated in South Africa, another major destination, or spread across several systems. A country-to-country mobility matrix would be needed for that question.
A time series would also help distinguish persistent patterns from one-year fluctuations. International student flows can be small for some origin–destination combinations, which means a change of only a few students may shift a parity index sharply. Comparing several years while keeping the same indicator definition would therefore be more informative than treating a single endpoint as a stable national characteristic.
Source and interpretation rules
The 2023 values are from the UNESCO Institute for Statistics API for OE.5T8.40540.GPIA using NATIONAL observations. The definition of the adjusted gender parity index and the 0.97–1.03 parity convention follow the UIS methodology explanation for GPIA. UIS describes the adjusted form as bounded between 0 and 2 and symmetric around the equality reference of 1.
All 171 reported observations were retained as published. Missing catalogue units were not filled with zero. The counts, percentages and mean in this comparison were calculated directly from the reported 2023 values. Because the data are unusually concentrated at exact 0, 1 and 2 values, interpretation should emphasize categories, endpoints and data limitations rather than small numerical rankings.
Frequently Asked Questions
What does a GPIA of 1.00 mean?
It means the female and male values of the underlying indicator are equal. UIS commonly treats 0.97–1.03 as the practical gender-parity range.
Does a GPIA of 0 or 2 mean there are many students?
No. Those are endpoints of the adjusted index, not measures of student volume. Sex-specific student counts are needed to judge the size of the flow.
Should countries missing from the extract be treated as zero?
No. A missing observation is different from an observed value of 0.00 and should not be zero-filled.
Related Articles
Green Map creates custom-edited map images using open geographic data sources such as geoBoundaries, Natural Earth, OpenStreetMap, and government open data. These maps are edited visual materials, not raw data files, and are provided for education, documents, presentations, and graphic reference.
