This dataset gives an early global picture of adult literacy, using data mostly collected between 1945 and 1954. UNESCO estimated illiteracy rates for people aged 15 and over. In many cases, the numbers came from recent censuses. Where data were missing or outdated, estimates were made using past trends, school or military records, or general information about each country.

Because the data was not always complete, most illiteracy rates are shown in 5% steps. Even in countries with high literacy, a small number of people—about 1–2%—were assumed to be unable to learn to read or write due to disabilities. The estimates use a simple definition of literacy: being able to read or to read and write.

This dataset presents UNESCO’s first global survey of adult illiteracy, based on census and official government data from 26 countries. Countries were included if they had illiteracy data for at least three census years since around 1900. The dataset uses the most recent available figures and shows how literacy levels changed over time in the first half of the 20th century.

The study notes that countries defined and measured literacy in different ways. Some counted people as literate if they could read or write; others required more, like writing a letter or reading a passage. In some places, being able to sign one’s name was enough; in others, it was not. The age at which people were counted also varied—from as young as 5 to as old as 15.

These differences make it hard to compare literacy rates across countries or even across time within a country. People who could only read or only write were classified differently depending on local rules. Other issues, like missing data and unclear definitions, add to the difficulty. Despite these limits, the dataset offers important insights into global literacy patterns during this period.

This paper presents estimates of the production of manuscripts and printed books in Western Europe over a span of thirteen centuries. To estimate literacy levels, the authors first calculate book consumption per capita, then convert these figures into literacy rates using two key inputs: historical trends in book prices and an assumed elasticity of demand for books. The book price data are sourced from the work of Jan Luiten van Zanden and Gregory Clark, while the elasticity of demand—reflecting how book consumption responds to price changes—is set at 1.4, based on contemporary economic literature. This method assumes that lower book prices made books more accessible, which in turn helped to increase literacy rates.

The data comes from the How Was Life? Global Well-being Since 1820 report by the OECD. It notes that the concept of literacy has changed over time, which affects how it is measured. In earlier periods, literacy was broadly understood as being “educated” or “learned.” However, by the late 19th century, the definition narrowed to mean the ability to read and write simple texts. This shift in meaning was driven by the spread of mass education, which emphasized basic reading and writing skills for large populations. As a result, historical estimates of literacy are based on differing definitions over time, making it important to interpret the data within its historical context.