Funding Data from Publication Acknowledgments: Coverage, Uses, and Limitations

Abstract

This article contributes to the development of methods for analysing research funding systems by exploring the robustness and comparability of emerging approaches to generate funding landscapes useful for policy making. We use a novel data set of manually extracted and coded data on the funding acknowledgements of 7,510 publications representing UK cancer research in the year 2011 and compare these “reference data” with funding data provided by Web of Science (WoS) and MEDLINE/PubMed. Findings show high recall (around 93%) of WoS funding data. By contrast, MEDLINE/PubMed data retrieved less than half of the UK cancer publications acknowledging at least one funder. Conversely, both databases have high precision (+90%): That is, few cases of publications with no acknowledgment to funders are identified as having funding data. Nonetheless, funders acknowledged in UK cancer publications were not correctly listed by MEDLINE/PubMed and WoS in around 75% and 32% of the cases, respectively. Reference data on the UK cancer research funding system are used as a case study to demonstrate the utility of funding data for strategic intelligence applications (e.g., mapping of funding landscape and co‐funding activity, comparison of funders' research portfolios).

Who Bears the Burden of Long-Lived Molecular Biology Databases?

“…Funding statements are not always available. Even when they are present, the context in which authors reference funding must be interpreted (11). While it is fortunate that funders could be attributed to a high percentage of databases covered here, the data should not be over-interpreted.…”

“…Although sub-organizations were often reported for both funding or operation, e.g., the National Library of Medicine for the US National Institutes of Health or the Interdisciplinary Nanoscience Center at Aarhus University, this granularity was inconsistent, a complication noted in previous efforts to evaluate research funding (11,12). Thus, the highest level of the organization is reported here.…”

“…Through review of acknowledgments and funding statements within articles, funders were attributed for 63 (94.0%) of the 67 databases. This percentage is on pare with evaluation of life science journals, where results show ~90% of articles report funders (11). A total of 89 unique funding organizations were identified 176 times (supplemental Table 1).…”

“…Since a database's most recent update article may have been published prior to widespread indexing of funder metadata circa 2009, the full text was manually reviewed to capture funding organizations. Following the heuristics of Grassano et al (11), organizations had to be explicitly named within the article to be included and were not inferred from other context. Organizations were only counted once per database, regardless if mentioned multiple times within statements (e.g., for distinct grants).…”

Assessing the uptake of persistent identifiers by research infrastructure users

“…In some similar studies, acknowledgments are considered to be part of the article full text, such as in the data citation analysis of Mayo, Vision, and Hull [24], but for this study acknowledgments were given the second priority because they have been examined as an interesting aspect of scholarly communication themselves in numerous studies (e.g. [25,26]), and some work has been done to develop automated acknowledgment analysis tools [27,28]. In-text references were thus assigned the lowest priority.…”