COUNTER filter
COUNTER developed a ‘unique article’ filter to compensate for the inflation of usage statistics caused by certain vendor interfaces and did a survey of current vendor practice on implementing unique article identifiers.
Full Project Title
Improving the Comparability of Usage Statistics and the Implementation of Unique Article Identifiers
Summary
COUNTER is an international initiative to improve the reliability and comparability of usage statistics for online publications. Its Codes of Practice set standards for recording, reporting, and delivering usage statistics from vendors (publishers, aggregators, etc) to libraries. COUNTER is always seeking to improve the quality of the statistics generated according to its Codes, and this was the starting point for the PALS 2 project. An independent study by Davis and Price (see reference below) suggested that the design of a vendor’s electronic interface could have a measurable effect on the usage statistics generated. For example, if a vendor requires users to view an HTML version before viewing the PDF, this could inflate the number of full text downloads. COUNTER felt that it would be useful to investigate this potential problem and propose solutions.
During the project, COUNTER worked with publishers to develop and test two data filters – an ‘unwanted HTML’ filter and a unique article filter. Project partners EBSCO and Elsevier Science developed a spec for each filter, created the filter, and tested it out on their own data. These initial tests indicated that the ‘unwanted HTML’ filter was not viable, so work on it was terminated. However, tests of the unique article filter looked promising, so this filter was improved and tested by five publishers. The unique article filter compensates for the inflation of usage statistics by providing a new metric – the number of successful unique article requests in a session. COUNTER will recommend that vendors use the new unique article filter and that the new metric is included in the next Code of Practice.
The project also conducted a survey of current vendor practice on implementing unique article identifiers. EvidenceBase conducted the survey, gathering data from 11 publishers. As there was great variation in practice, the project has recommended some best practice guidelines to include in the next Code of Practice.
Outcomes
This project is the first to test data filters on usage statistics from different publishers. This would not have been possible pre-COUNTER, and the project has explored new territory. The two main outcomes are:
-
Unique Article Filter – Including this in the COUNTER Code of Practice will provide librarians with a very useful new metric – ‘Number of Successful Unique Article Requests in a Session’. This will let them know how many unique articles are used from a given journal in a session, and provide a valuable check on the total full text usage in all formats provided in the existing COUNTER Journal Report 1.
-
Unique Article Identifiers – Their survey has shown that there is a wide diversity of practice among publishers regarding the implementation of unique article identifiers. In an online environment a more standardised approach is needed, if usage is to be measured credibly and different versions of articles are to be clearly identified.
The project has also learned some important lessons about the practical challenges involved in attempting to extract more information from vendor usage data. While in principle it is possible to further enrich the COUNTER usage data by using appropriate data filters, there will be limits to this in practice. These are determined on the one hand by the limitations in the data generating process, and on the other hand by the unwillingness of vendors to invest in these processes unless there is a clear benefit in terms of the quality and value of the data generated.
Outputs
Reports
The project’s final report (PDF) to JISC explains the project approach and methodology, summarises the results, and includes supporting data in appendixes. The appendixes are listed here in order and by topic under relevant headings below.
-
Appendix A (PDF) – HTML Filter Enhancement Methodology
-
Appendix B (PDF) – Unique Article Filter Testing Methodology
-
Appendix C (Excel) – EBSCO Results for 'Unwanted HTML' Filter Tests
-
Appendix D (Excel) – EBSCO Results for Unique Article Filter Tests
-
Appendix E (Excel) – Publisher A Results for Unique Article Filter Tests
-
Appendix F (Excel) – Publisher B Results for Unique Article Filter Tests
-
Appendix G (Excel) – Publisher C Results for Unique Article Filter Tests
-
Appendix H (Excel) – Publisher D Results for Unique Article Filter Tests
-
Appendix I (Excel) – Publisher E Results for Unique Article Filter Tests
-
Appendix J (PDF) – Unique Article Identifier Survey Briefing Paper
-
Appendix K (PDF) – Unique Article Identifier Survey Results
Research on Data Filters
- Technical model for the ‘unwanted HTML’ filter – see Appendix A (PDF) of final report
- Technical model for the unique article filter – see Appendix B (PDF) of final report
- EBSCO test results for both filters – see Appendix C (Excel) & Appendix D (Excel) of final report
- Publisher test results for unique article filter – see Appendixes E-I of final report
Research on Unique Article Identifiers
The Final Report (PDF) summarises the results for each question and the full data is presented in the appendixes:
- Briefing paper and questions for publishers participating in the survey – see Appendix J (PDF) of final report
- Tabulated survey results – see Appendix K (PDF) of final report
Presentations
Further Information