You are here

Home

Open retractions data: podcast

Conversations with COPE 

Ivan Oransky co-founder of Retraction Watch and Ed Pentz the Executive Director at Crossref join Trevor Lane to discuss the acquisition of the Retraction Watch database by Crossref. They talk about the benefits this brings to the community and how this partnership improves access to the data by authors, readers and publishers. The conversation moves on to the wider picture of researcher incentives and trust signals.

Summary

Describe the new partnership between the two nonprofit groups and tell us how it came about?

Crossref acquired Retraction Watch's database of article retractions, aligning with their shared mission. The partnership means Crossref have much stronger retraction data from Retraction Watch and the benefits of the technological expertise at Crossref will improve the database and enable the data to be open to the community. The Retraction Watch taxonomy gives reasons for a retraction which is valuable to the community. Both Ed and Ivan look forward to the potential community impact and future technological improvements. 

How can people expect to use the database?

The open dataset can be downloaded, and is integrated into 3rd party systems making it easier to check for retracted articles. Ed anticipates the integration of retraction data into existing systems like discovery services, reference managers, and publishers' platforms. He envisions broader dissemination through Crossref's API and anticipates future enhancements aligning with NISO's recommendations.

Is there a parallel system for expressions of concern?

While Retraction Watch's database primarily focuses on retractions, it also includes expressions of concern and limited corrections. However, the handling of expressions of concern poses greater challenges due to varying publisher practices.

On the journalism side how do you select which cases to report in Retraction Watch given that retractions could be either misconduct or honest mistake? 

Retraction Watch will continue with their independent journalism, separate from the database acquisition. The selection process for reporting on retractions has evolved due to the significant increase in retractions over the years. Cases are prioritised based on factors like paper prominence and underlying story interest which goes beyond retractions but is related. 

How do you think the database could be used to improve generative AI tools by ensuring that they include only sound work that's unretracted in either the training or the searching processes?

Regarding the database's potential impact on generative AI tools, Ivan and Ed acknowledge its utility for specialised services but it would be very difficult for general large language models as they are trained on data from across the web without reference to provenance. 

What will happen after the 5 year partnership period is over

The plan is for an ongoing relationship and expected renewal after 5 years, with a revised and updated agreement to take account of how the landscape changes over those 5 years.

Is the database archived?

Crossref makes an annual public access file hosted on a 3rd party server of all their metadata, ensuring continued access.

What else do you think needs to happen in the research community for better preservation of the Integrity of the research record?

Ivan emphasises the need to address research integrity issues upstream by tackling the incentive structure in academia, aiming to prevent pollution of the literature rather than just cleaning it up. He advocates for a holistic approach focused on openness, fair assessment, and shifting away from the publish-or-perish mentality. Ed echoes this sentiment, highlighting efforts to connect research metadata and assess quality and impact more comprehensively. 

Useful links