UNESCO Institute for Statistics (UIS) indicator `FHLANGILP.LOWERSEC.LOWSES` measures the percentage of lower-secondary students from a very poor socioeconomic background whose first or home language is the language of instruction, for both sexes combined. It is an ordinary percentage, not a parity index. The population is deliberately narrow: the denominator contains lower-secondary students classified in the very disadvantaged socioeconomic group, not all students in the country.
The 2022 dataset contains 75 official national observations. The mean is 81.49% and the median 89.27%. The maximum is 99.26%, followed by El Salvador at 98.97%, Brazil at 98.92%, Japan at 98.91%, and Chile at 98.67%. At the other end, the Philippines records 4.32%, Brunei Darussalam 5.60%, Malta 9.25%, and Morocco 11.84%. The wide spread makes the lower tail as important as the high national values.
Table of Contents
The socioeconomic denominator defines the indicator
The percentage should be read as a conditional measure. It asks: among lower-secondary students in the very poor socioeconomic group, what share have a first or home language that is also the language used for instruction? A value of 80% does not mean 80% of all lower-secondary students in the country experience a language match.
The socioeconomic category should not be converted into a single monetary poverty threshold. International student assessments often construct socioeconomic background from household possessions, parental education or occupation, learning resources, and other information. The current data do not provide one comparable income cutoff for every country.
Language match describes an instructional condition
A student’s first language may be the language learned earliest, while home language can refer to one or more languages regularly used in the household. Schools may teach in a national language, regional language, minority language, or another instructional language. The indicator records whether a first or home language matches the language of instruction.
That makes the percentage an indicator of linguistic alignment in the learning environment. It does not tell us how proficient the student is in either language, how frequently each language is used, or whether a multilingual student receives bilingual instruction.
The highest observations are close to 100 percent
The top of the distribution is tightly clustered. The maximum is 99.26%, followed by El Salvador at 98.97%, Brazil at 98.92%, Japan at 98.91%, Chile at 98.67%, Poland at 98.23%, Colombia at 98.11%, Cambodia at 97.99%, Hungary at 97.58%, and Romania at 97.18%.
There are 21 observations at 95% or higher. In these systems, most observed lower-secondary students from the very disadvantaged socioeconomic group have a first or home language matching the language of instruction. This does not establish that schooling is otherwise equitable, high quality, or universally accessible.
| Country or territory | 2022 language-match share |
|---|---|
| Republic of Korea | 99.26% |
| El Salvador | 98.97% |
| Brazil | 98.92% |
| Japan | 98.91% |
| Chile | 98.67% |
| Poland | 98.23% |
| Colombia | 98.11% |
| Cambodia | 97.99% |
| Hungary | 97.58% |
| Romania | 97.18% |
The lower tail is exceptionally wide
The Philippines has the lowest observed share at 4.32%. Brunei Darussalam is at 5.60%, Malta at 9.25%, Morocco at 11.84%, Indonesia at 36.29%, Paraguay at 43.86%, and the United Arab Emirates at 49.43%. Singapore follows at 52.28%, Switzerland at 59.40%, and Austria at 63.97%.
Seven observations are below 50%. A low percentage means many students in the specified socioeconomic subgroup are being taught in a language other than their first or home language. The series does not determine why. Multilingual social structures, migration, education policy, language history, and school choice could matter, but country-specific explanations require separate evidence.
| Country or territory | 2022 language-match share |
|---|---|
| Philippines | 4.32% |
| Brunei Darussalam | 5.60% |
| Malta | 9.25% |
| Morocco | 11.84% |
| Indonesia | 36.29% |
| Paraguay | 43.86% |
| United Arab Emirates | 49.43% |
| Singapore | 52.28% |
| Switzerland | 59.40% |
| Austria | 63.97% |
The mean is well below the median
The mean is 81.49% and the median 89.27%. The first quartile is 78.22% and the third quartile 95.53%. Half of the observed countries and territories therefore fall between roughly 78.22% and 95.53%.
The mean is pulled downward by a small number of very low values. Observations below 12% in the Philippines, Brunei Darussalam, Malta, and Morocco have a substantial effect on the arithmetic average. The median is more representative of the middle of a distribution that is concentrated at relatively high values but has a long lower tail.
Five bands show the shape of the distribution
There are 7 observations below 50%, 10 between 50% and under 75%, 22 between 75% and under 90%, 15 between 90% and under 95%, and 21 at 95% or higher. In total, 36 of the 75 observations are at least 90%.
The pattern is therefore not a simple divide between high and low countries. A large cluster is near the top, a broad middle spans 75–95%, and a small set of countries sits far below the rest. Those lower observations strongly affect averages and deserve separate attention in any cross-country reading.
This is not a measure of lower-secondary access
The denominator includes students who are already participating in lower-secondary education and belong to the specified socioeconomic group. Children who are out of school or never reached lower secondary are not represented in the same way. A high language-match share therefore does not mean that disadvantaged adolescents have high enrolment or completion.
Access requires separate measures such as enrolment ratios, attendance, out-of-school rates, and completion. The current indicator is better understood as a feature of the learning environment among students who are observed in school.
Language match should not be substituted for learning outcomes
Instruction in a familiar language can be an important educational condition, but this percentage is not a reading score, mathematics result, or minimum-proficiency indicator. A country can have a very high language-match share and still face learning challenges.
Learning achievement must be assessed with direct outcome measures. If researchers want to study whether language alignment is associated with proficiency, they need both sets of data and an analysis that separates correlation from causation.
National values can conceal internal linguistic diversity
A national percentage can average over major differences by region, migration background, ethnic or linguistic community, and school type. A country above 90% can still contain groups whose home language is rarely used for instruction.
The current observations are national in geographic scope. They are useful for identifying large cross-country differences, but they cannot tell us which subnational communities experience language mismatch. That requires more disaggregated data.
One disadvantaged subgroup cannot quantify socioeconomic inequality by itself
Because this series contains only the very poor socioeconomic group, it does not directly measure the gap between disadvantaged and affluent students. A country with an 85% value could have almost no socioeconomic difference or a very large one; the comparison group is missing from this indicator alone.
A socioeconomic inequality analysis would need corresponding indicators for other background groups under the same definitions and assessment design. The current dataset is best used to compare the disadvantaged subgroup across countries, not to claim that poverty itself causes language mismatch.
Missing observations remain missing
There are 75 official national observations for 2022. Countries absent from the dataset do not have a language-match share of zero; they simply do not have a reported observation in this indicator-year combination. None of the 75 observed values is an official zero.
Replacing missing values with zero would create false extreme cases and lower the mean. Correct international comparison preserves the distinction between ‘no data’ and an observed 0%.
A 2022 snapshot cannot establish the trend
The cross-section shows differences at one point in time but not the direction of change. Language policy, migration patterns, school supply, and the composition of the socioeconomic subgroup can evolve between assessment cycles.
Trend analysis should follow the same `FHLANGILP.LOWERSEC.LOWSES` series across comparable years and verify that the socioeconomic classification, sampled population, and assessment design remain sufficiently comparable.
Official source and interpretation
The official source is the UNESCO Institute for Statistics. This comparison uses 75 NATIONAL observations for 2022 from indicator `FHLANGILP.LOWERSEC.LOWSES`, reported in percent for both sexes combined. The measure is the share of lower-secondary students from a very poor socioeconomic background whose first or home language is the language of instruction. Official values are available through the UNESCO UIS Data API.
The most important reading rule is the population scope. This is not a national percentage for all lower-secondary students. It describes the instructional-language environment of a particularly disadvantaged socioeconomic subgroup, and it does not directly measure access, learning achievement, or educational quality.
Frequently Asked Questions
Does this indicator cover all lower-secondary students?
No. It covers only lower-secondary students classified as having a very poor socioeconomic background.
Does a high language-match share mean high enrolment?
No. It describes students already in the observed education group and is not an access measure.
Is language match the same as learning achievement?
No. It describes an instructional-language condition rather than proficiency or test performance.
Can countries with no observation be treated as zero?
No. Missing official observations remain missing and must 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.
