Significant flaw in new library eBook study
/On Wednesday, September 3, 2026, the AAP and Authors Guild announced a new study they commissioned to analyze the impact of library eBooks and eAudiobooks on the retail market. However, their study authors failed to understand the limitations of the datasets they were using. The report is seriously flawed in its calculation of the number of library eBooks in the market, improperly counting shared consortia collections as if they were offered by individual libraries. To be clear, this is not an error within a few percentage points. They overstate eBook availability by multiple times.
The study authors use the IMLS Public Library Survey (PLS), a valuable tool for understanding traditional library statistics, but with limited usefulness for understanding digital library data because it has no way for libraries to indicate when they share collections as part of a consortia.
For example, the state of Rhode Island has 49 libraries, all of which participate in Ocean State Libraries, a consortium that contracts with OverDrive for ebooks and eAudiobooks. In the 2023 PLS, each of the 49 libraries reports that it offers its patrons an eBook collection with ~186,000 items. If we used the study authors' methodology to determine the total number of eBooks available to residents of Rhode Island, we would total all of the numbers each of the 49 libraries reports and come up with 8.96 million eBooks. But in fact, the total number of eBooks available to Rhode Island residents in 2023 was ~186,000 and the study methodology overreports by a factor of 49.
Ebook consortia are the norm in the US, not the exception. Every state has at least one consortium. At least eight consortia are run at the state level like Rhode Island. 37 of the 50 states have at least one consortium with more than 20 libraries, and some are much larger. The Ohio Digital Library, for example, includes 180 libraries.
To find the total number library eBooks available in a given year, we need to determine the total offered by each consortia in the country and then add that to the totals for non-consortial libraries. To our knowledge there is no single list of all consortia and the libraries they include, so it is not possible to calculate from public records the number of eBooks available through public libraries.
Unfortunately for the study authors, their numerous arguments about the economic impact of library eBooks rests on the claim that 1.5 billion library eBooks were available in 2023. Because this number is wildly inflated, all conclusions they draw about the availability of library eBooks and its impact on the retail market must be summarily rejected.
