ranking literacy by using global library provision
The OCLC global library stats were an attempt by the OCLC to collect global stats for libraries. These include number of libraries, volumes, members, and librarians in public/academic/school/special libraries. The stats were manually collected and the relevant sources listed for each entry. The sources range from latest national stats, to decades-old library travel guides.
When counting libraries, the OCLC appreciate it's difficult to ensure standards.
For the sake of uniformity, the total number of libraries represents administrative units and not service points, since not all sources report service points consistently. Some of the country entries did not specify which unit they were reporting, in which case it was assumed that it was administrative units.
The administrative units figure is strange as it often doesn't represent what many would think of as a library. And for comparative stats, each country will have different systems of administration. In the UK the administrative unit is commonly the local authority, which can significantly vary in size, population, and number of libraries. It also seems odd to assume an ambiguous figure would be administrative units, and not branches.
Looking through the data it's possible to spot some inconsistencies:
An excellent site, Practical Statistics, looks at the accuracy of the data, particularly for the Norway data.
The results are not encouraging. The library statistics that OCLC offered for Norway could not be used for serious work. The quality was far too uneven. OCLC faced the same problem that Unesco faced. Countries with weak library systems have hardly any statistics at all. There may be some numbers floating around, but they do not reflect the situation on the ground. Countries with stronger systems have better statistics, but they do not publish them in a form that is convenient for outsiders.
A recent report on 'Literature Behaviour Characteristics' ranked Finland as the most literate nation.
The report is detailed at World's Most Literate Nations. It doesn't only use educational test results, but also cultural measures:
The data used for libraries turns out to be the OCLC global library statistics. The method used to rank the countries is to take those measures from the OCLC data, and adjust for population size. A key missing measure is number of librarians, which doesn't appear to be considered so important.
The data sources and methodology used is very open. The specific rankings for libraries are listed at Library rankings.
Looking through the rankings, the list becomes reflective of population size. China, with a population of over a billion, are at the bottom of the ranking. For example, if Estonia have 487 libraries with a population of 1.3 million, China (population 1.3 billion) would need to have around 487,000 library authorities to match Estonia. The OCLC stats say China has just 3073. But is libraries per population a fair measure? Commonly in library consultations distance to nearest library is used as a measure of provision, not the number of people a library serves.
As ever, it comes down to not having enough quality data with relation to libraries. Most countries don't seem to have enough data to accurately represent number of libraries in a consistent way, and the OCLC found that combining that data across countries was next to impossible. In this case taking that data further and constructing a ranking of literacy based on library provision seems an even greater stretch.