Summary

  • HoPEc asked researchers to select their own works from suggested records and confirm changes by email, helping distinguish authors with similar names.
  • In 2009, RePEc encouraged authors to maintain their profiles for better visibility; its automatic search took close to three months to cycle through registered people.

Michael B. Devereux and Michael P. Devereux were two different economists. A catalogue handed the name “Michael Devereux” had a small administrative problem with considerable potential for mischief.

The team behind HoPEc used this example in its paper dated July 2000. Initials, spelling variations and records from different archives complicated the sorting. Its answer was to enlist someone with a useful qualification: the person who had written the paper.

HoPEc’s registration service began in 1999. Researchers supplied personal details, then selected their works from suggestions using checkboxes. If a work was missing from the suggestions, they could enter its record identifier. The work had to exist in RePEc already. Changes required a four-digit confirmation sent by email.

The designers hoped to avoid forgotten-password administration and the duplicate accounts it could produce. They still expected some manual checking for duplicates. Even a system designed to save clerical work needed a clerk somewhere.

An appeal to scholarly self-interest

By April 2009, RePEc’s advice on improving an author’s ranking read partly like instructions for tidying the catalogue.

Claim every listed work. Keep the working-paper version in the profile after journal publication, so versions can be linked. Add the different forms in which publishers print your name. Check citations that the system had left for human approval because the match was uncertain.

There was even a way back from an earlier mistaken refusal: authors could restore the item from their refused list and run the search again. The service needed help remembering what belonged to whom. It had found an audience with a professional interest in the answer.

Your paper arrived before its reminder

Automation supplied a particularly good reason to take part. A May 2009 explanation said papers added to an archive were usually listed by the main discovery services within 24 hours and available for authors to claim the next day.

The automatic search for authors moved at a different pace. It rotated through 27,000 registered people, taking close to three months to complete a cycle. The searches burdened the machine and usually found nothing new. Authors who knew a paper was available were encouraged to log in and search manually.

The paper could be sitting in the catalogue while the reminder was still making its rounds. An author who remembered a new publication had a reason to check before the scheduled sweep.

The Christmas Eve division of labour

A 24 December 2009 post drew the boundaries plainly. Missing or incorrect publication listings belonged with the relevant archive maintainer: publishers controlled those records. Authors could maintain the works and affiliations in their own profiles.

The team described roughly 300 newly indexed works a day and nearly 23,000 registered authors, supported by spare-time labour. A request to correct a profile could therefore return to the person who sent it.

RePEc had made scholarly discovery depend on an unusually well-informed filing team: the scholars themselves. They knew which papers were theirs. The catalogue gave them somewhere useful to say so.