RePEc in July 2021

August 7, 2021

Everyone seems to be on vacation in July, as we are seeing a serious drop in activity across the board. We recorded 417,185 file downloads and 2,245,986 abstract views and welcomed the following new archives: University of Manchester (III), Idea Publishers. And we reached the following milestone:

3,700,000 indexed research items
2,400,000 indexed articles

RePEc in June 2021

July 6, 2021

IDEAS has added a new feature: if an item has a DOI, an alternative link to full text has been added that uses LibKey. This allows users to leverage their library subscriptions. We also welcomed the following new RePEc archives: Nature, Université Mohammed VI Polytechnique, Transition Academia Press, Journal of Management Practices, Humanities and Social Sciences, Journal of Advanced Research in Social Sciences and Humanities, Journal of ICT, Design, Engineering and Technological Science. We counted 545,984 file downloads and 2,926,106 abstract views. And finally, we reached the following milestones:

3,600,000 indexed items

The importance of name variations in author profiles

July 4, 2021

An importance project in the RePEc landscape is the RePEc Author Service. It allows economists to create an online profile of their works, built as their “recognize” the works they have authored that have been indexed in RePEc by the respective publishers.

Of course, as there are over 3.6 millions works indexed in RePEc at the time of this writing, the authors need to be helped in sorting through this large amount of data. This is done on the basis of name variations. The name is not sufficient. Indeed, publishers have many ways to name authors and it is easier to offer suggestions for authoring claims based on author names. During the the registration process on the RePEc Author Service, authors are asked to provide a list of name variations (which they can amend later) to facilitate this. But it is important to be comprehensive.

For example, say William Smith registers. The system will make automatically the following suggestions:

  • William Smith
  • W. Smith
  • Smith, William
  • Smith, W.

If the author approves this list of name variations as is, he will be a list of suggested works, first with those having authors with names corresponding to the variations. Not all may be his, as there could be homonyms. The author will also get an email whenever a new work was added to the database with one of those names.

However, this short list of names may not be sufficient. Indeed, publishers may be recording a middle name or at least a middle initial. If this is not captured in the list of name variations, the author may miss out on potential works to add to the profile. The same applies for a suffix (Jr., Sr., III). What if the author is also known as Bill or Will, which may be recorded as well, in particular for working papers?

There may be further complications due to local naming conventions. For example the French first name Jean-Claude may be abbreviated as J.-Cl. or J.-C., and similarly in Russian (Yuri, Yu.). Hispanic names often have two last names, with the second one sometimes omitted or abbreviated (Sánchez Gómez, Sánchez G., Sánchez), while in Portuguese the first of the two is sometimes omitted. Publishers sometimes includes accents and sometimes not. For all these cases, it is important to help the RePEc Author Service by providing a complete list of variations that publishers could be using.

And then there are those cases where authors change names, particularly through marriage. Anglicizing a name, adopting a religion or gender change can also be reasons to adapt the list of name variations. If one has published in a different script (Arabic, Cyrillic). that should be reflected, too.

This may sound like a lot of work for some authors, but this allows RePEc to provide more accurate suggestions. But we have noticed that many authors have a list of name variations that is too small. For this reason, we had to expand the email alerts beyond the strict name variations, or authors would not get the notices they would be expecting.

RePEc in May 2021

June 7, 2021

The big news of the month is that we reached major traffic milestones, with 120 million downloads and 500 million abstract views. We also had a good number of new RePEc archives: Asia-Pacific Applied Economics Association, International Journal of Business and Economic Affairs, International Journal of Business Research and Management, TAF Publishing, Polish Economic Society Johns Hopkins University Press, European Fiscal Board, Red Nacional de Investigadores en Economía (RedNIE), International Journal of Business and Administrative Studies. We also counted 639,773 file downloads and 3,233,905 abstract views. And here are the milestones for the month:
500,000,000 cumulative abstract views
120,000,000 cumulative downloads
150,000 indexed book chapters
700 RePEc-wide h-index

RePEc services reach major traffic milestones

June 3, 2021

The cumulative number of downloads through RePEc services has just surpassed 120 million, and the numbers of views of abstract pages has reached 500 million. Before discussing how we reached these milestones, let us understand what these numbers mean.

Lots of websites use RePEc data. A list of those we know about is found on the RePEc homepage. Some of them provide detailed traffic data to LogEc. The reporting sites are currently EconPapers, IDEAS, NEP and Socionet. Previously, NetEc and Economists Online would report until they shut down. LogEc takes all that data and analyzes it to remove anything that does not look like human traffic and delete any repeats. Doubtful cases go through additional human review. Typically much over 90% is thrown out in this process.

Two metrics are used: abstract views, that is, viewing by a human of the abstract page of a paper or article on the websites of the reporting RePEc services, and downloads, that is, any clicks downloading full texts from the publishers through the reporting RePEc service, whether the download is successful or not (bad links, gated material).

There are about 3.6 million items indexed in RePEc. RePEc has existed for 24 years. If we assume that on average an item has been indexed in RePEc for 10 years, we have that on average an item’s abstract is viewed just under 14 times a year and downloaded about 3.3 times a year. Or 140 cumulative abstract views and 33 cumulative downloads.

How did RePEc traffic evolve over time? The graphs below shows that it has increased fast at the start, then ebbed considerably and is picking up again. A 2014 blog post has some conjectures about why this prolonged dip happened. With a few more years of observations since that post, we believe the cannibalization hypothesis is the most likely one: the reporting RePEc services have been there from the start, while newer RePEc services do not report traffic data. If they have taken any traffic away from the reporting ones, overall RePEc has been going up all along, we just cannot measure it.

RePEc in April 2021

May 15, 2021

With some delay, here is what we have to report about the month of April: We welcomed a few new archives: Wojskowa Akademia Techniczna, Swiss Federal Institute of Aquatic Science and Technology, International Journal of Cyber Forensics and Advanced Threat Investigations, Journal of İbn Haldun Studies. We counted 674,225 file downloads and 3,522,965 abstract views. And here are the milestones we reached:
3,500,000 indexed works
3,200,000 indexed works available online
120,000 indexed book chapters
50,000 indexed books
1,000 seminars announced on the Economics Virtual Seminar Calendar

Searching on IDEAS: a tutorial

May 5, 2021

With over 3.5 million items of economic research, the RePEc bibliographic database is daunting. Lucky, various RePEc services make is manageable for the users. One such service is IDEAS. It allows the contents to be browsed in various ways. This post is about the search options it offers.

IDEAS has a basic search box on every page. The front page, the advanced search page and any search result page offers more options and features.

Keyword syntax
The search syntax is quite simple. Type a series of keywords and the search engine will look for matches assuming you want all those words present. In other words, it inserts a + (plus) sign between each keyword. If you do not want one keyword, that, is, you want to exclude results mentioning a particular word, prepend a ~ (tilde) sign before that word:

money growth ~inflation

looks for all mentions of money and growth that do not mention inflation. If you only want results that include the literal “money growth”. then use these double quotations marks. Results will then only show when these two words appear together:

“money growth” ~inflation

Sometimes, a keyword can be expressed with various words. Below we discuss synonyms, but in the meanwhile, you can list alternatives with | (pipe). Use parentheses to make the search clear:

(money | currency | monetary) growth ~inflation

You can nest these declarations at will.

Keyword stemming and synonyms
IDEAS search uses algorithmic word stemming. This allows to find on the basis of one word others in the same family. For example, searching for “count” finds also “counts”, “counting”, “counter”, etc. This is algorithmic, and thus not foolproof. In doubt, use more words with the pipe (|). Note that you cannot use wildcards (“count*”).

IDEAS search uses a custom synonym dictionary for economic terms. For example, “ML” will also search for “maximum likelihood”. If you feel some synonyms are missing, there is a form to suggest more (also available from any search result page).

Search options
You can limit the search in various ways:

  • Document types: articles, papers, books, chapters, software, or all
  • Search field:abstract, keywords, title, author, or all
  • Start year
  • End year

Result options
IDEAS offers various sort options for the results. Every search result receives a relevance score that is used for the default sorting but there are other options, such a date published, date added to RePEc, citation counts, alphabetical (title), and combinations thereof.

Search within series or JEL code
There is also a search box on any series, journal or JEL code page. This allows to perform a search within that series, journal or JEL code. The initial search form is simple, the results page allows to refine in the same way as for general searches.

Saving your search results
MyIDEAS is a personalized tool that allows in particular to save searches and search results. It requires to have an account on the Author RePEc Service, which is used for authentication. Once logged into MyIDEAS, you can click on any blue “Save to MyIDEAS” next to a search result and it will be saved to your personal bibliography. The collected results can be sorted into folders and exported into various formats.

You can also save the search itself to MyIDEAS. This allows you to see what has been added to RePEc for since particular search since you last checked in MyIDEAS. You can also get weekly email digests with the additions.

Other tricks

  • Searching for a journal name will likely give a link to the journal page on IDEAS
  • Search for “AuthorOne AuthorTwo (Year)” yields all items these two authors have published together in that year.
  • Search only among software and keywords with the name of a software gives all components using that software.

RePEc in March 2021

April 7, 2021

Traffic on the reporting RePEc sites was stellar, we almost hit record numbers with 689,048 file downloads and 3,796,782 abstract views. This is very encouraging, as traffic has suffered for a few years, most likely from the competitions of sites that do not report their traffic to LogEc. We also has a good crop on newly participating archives: Islamia University of Bahawalpur, Ural Federal University, University of Camerino, Romanian Fiscal Council, Ss. Cyril and Methodius University, European Research Center, International Fellowship Journal of Interdisciplinary Research, Ikatan Sarjana Ekonomi Indonesia. Finally we reached the following milestone:

1,400,000 items with extracted references

RePEc rankings for economists to add new criterion: forecasting prowess

April 1, 2021

RePEc has been providing for many years rankings of economists. These are based on a set of criteria that follows how economists like to rate themselves: where they publish, how they are cited, and how much they are read. But this is not how the rest of the universe evaluates economists. They are supposed to forecast well the path of the economy.

This is exactly what the RePEc rankings will be considering as well. How well can the ranked economists forecast key economic variables? What is their track record in prognostication? For this purpose, RePEc will be leveraging FREDcast, a forecasting game conveniently managed at the St Louis Fed that allows users to forecast every month four key variables for the US economy: inflation, unemployment, GDP, and employment. Scores are computed based on actual data releases following a formula that takes into account forecast errors across the four economic variables weighted by forecasting difficulty. Over time, monthly scores are aggregated using a discount rate. The all-time score is the one that will be used.

For their FREDcast points to be recognized in the RePEc rankings, all economists need to do is to use for their FREDcast account the same email address they used for registering in the RePEc Author Service. FREDcast is available on the web as well as with Android or iOS apps.

Note that in a second phase, forecasting accuracy will also be used for ranking economic institutions. Here the decision is still pending on whether RePEc will aggregate the scores of each institution’s members, or whether each institution with become part of a league within FREDcast, to leverage the “battle league” functionality of FREDcast.

Institutional metadata in RePEc: the EDIRC project

March 30, 2021

RePEc disseminates metadata about publications in economics as well as about their authors. This data is then leveraged by various so-called RePEc services that make that data available to the public in various forms. Any element in this data has a unique and permanent identifier, the RePEc handle, that allows to link an article to its journal and authors, for example, as well as to its citations. Each have a handle.

EDIRC is a project started in 1995 that catalogs institutions that primarily employ economists: these are mostly academic departments, but also includes think tanks, policy institutions (central banks, ministries, regulators), international organization, and research centers. A listing of economics associations and societies is also provided. Each of those entities is assigned a RePEc handle and this handle can thus be used in RePEc to associate items to these institutions. These handles always start with RePEc:edi:, followed by seven letter or digits, the last two reserved for the country.

The major user of this data is the RePEc Author Service, which leverages it to assign affiliations to the registered authors. The author metadata then includes the handles of the relevant institutions, which the RePEc services then can use to display author profiles. The RePEc Genealogy does the same to identify where economists received their terminal degree from. RePEc archives, the providers of the bibliographic data in RePEc, can also include institutional handles in the data about their series and journals.

There may be still other uses of this data, such as using the institutional data for ranking purposes. But the ones described above feed back to EDIRC. Indeed, institutional entries on EDIRC may show a listing of their members registered with the author service, their alumni shown in the genealogy, and the publications indexed in RePEc. All in all, close to 15,000 institutions are indexed in EDIRC, with well over half having some metadata from another RePEc service.