IDEAS home Printed from https://ideas.repec.org/t/ranking.html
 

FAQ for RePEc rankings

For authors, we compute 36 different rankings accordings to 36 different criteria (all those from this list except NbWorks and Wu-Index), and then aggregate them up (after removing the two best and the two worst) using a harmonic mean of the rank orders. For institutions, we use three fewer criteria (Close, Betweenn abd NEPCites do not qualify). To see the different criteria and the rankings for everyone of them, see the top 5% authors or the top 5% institutions. Click on each column header for a different method. You can also compute alternative rankings by aggregating the rankings differently or choosing a different set of criteria here. Finally, a document describes in detail the ranking procedures.
The link expires once the next mailing is being sent. Watch your email. Note that you will only get the new link if the email in your profile is current. Please keep in mind that the mailing can take a whole week to prevent an overload of our servers as people check their ranking analysis or their profile. You can also use the old link to log in and get to it without the email.
  1. Write more and/or better papers. But you knew that.
  2. Make sure your papers are listed on RePEc. If your department is not yet participating with its working paper series, here are the instructions. Also encourage journals that do not yet participate to do the same.
  3. Make sure all your works are listed in your profile. That maximizes your chances of having works cited and contributes to your abstract views and download totals. In particular, do not remove from your profile old versions of your works. This would not remove them from the database, just from your profile. In addition, working paper series often have a higher impact factor than journals.
  4. Make sure your name variations in your profile are accurate, that is they cover all possible ways a publisher may call you.
  5. We only add citations when we have a high degree in confidence that they correspond to the proper work. For lower degrees of confidence, you can help in determining whether they cite the right work. Log into your profile and click on citations and see what is offered.
  6. If you have a homepage, link to your profile on IDEAS or EconPapers, and/or put such links in your email signature.
  7. Maintain a RePEc Genealogy entry.
See also this RePEc Blog post, or this one.
Reference extraction and citation linking is done by the CitEc project. See there for details. This data is then used, discarding self-citations and citations by different version of the same paper, to compute totals. There is also a RePEc Blog post that can be of interest.
Not from us. Ask this person directly, we consider rankings to be private, except for the top ones who have every reason to be proud of being listed.
This can happen when you remove items from your profile, when a publisher removes some items from the database (either refering or citing), or when we notice that two citing or cited papers are different versions of the same work. It can also happen that we remove wrongly attributed citations (something you can also do: log into your profile, then click on "citations" and "identified"). This RePEc Blog post has more.
References are only extracted from papers that are available online freely and could be parsed. We have no budget, and thus do not purchase such data or subscriptions. Some commercial publishers explicitely prohibit us from displaying references, as they are considered by them as part of the copyrighted work. Others see the value of getting links to their works and provide us directly with references. As you see, things are very uneven, and unless publishers can be convinced to release more of their records to the public, we are bound to analyze proportionally more working papers than journal articles, which is mostly recent material (and thus added value compared to services that only look at articles). If we were not publishing the reference lists, just making counts, numbers would be higher, but our service would be much degraded by the lack of crosslinking of references. For more details about the citation analysis, see CitEc or this RePEc Blog post.
Note that we only include citations for which we have a high confidence that they match the appropriate paper. For those with lower levels of confidence, you can help us: log into your profile, click on "citations", and approve or reject our suggestions. You can also help us by adding missing reference lists using this form.
Finally keep in mind that for ranking purposes, we do not count self-citations (but we list them), and we count only once when different versions of the same work cite you. Most other services do not perform these distinctions, unfortunately.
Each affiliation gets a weight, with all weights summing to 100%. Authors set those weights themselves. If they have not done so, a formula is applied: if possible, one half is reserved for the affiliation whose URL domain is the closest to the author's email address or web page URL. The rest is allocated according to a formula that gives less weight to affiliations with many authors. For details, see this blog post.
Authors and institutions are ranked within a region using scores that are weighted for multiregional affiliations (that is, their score is attributed to each region according to the weights) and computed by ranking authors or institutions among those of the same region. The ranks in square brackets are regional ranks if they were simply extracted from the worldwide rankings. Those to rankings can differ.
  1. Aggregate rankings are based on a harmonic mean of rankings for each criterion. Those individual rankings differ between worldwide and regional groups. As an example, an author who is much more cited than his regional peers is ranked far ahead in the worldwide rankings, but no such gap would appear in the regional rankings.
  2. For authors with multiple affiliations in different regions, only part of their score will count towards their regional ranking. See above how those weights are computed, or see this blog post.
In any case, the ranking as it would result from an exerpt from the world ranking is listed in square brackets next to each regional score. Note that their maybe rankings reversals between the regional and worldwide rankings. To see this, take an example of two economists from the same region and seven ranking criteria:
Economist A:  1, 2, 3, 4, 5, 6, 7,
Economist B:  7, 7, 7, 3, 4, 5, 6,
As the top and bottom ranks are dropped, the harmonic mean from (2, 3, 4, 5, 6) is 3.5 for Economist A and 5.5 (from (7, 7, 4, 5, 6) for Economist B. In the regional ranking, assuming these are the two top economists for each criterion, A has (1, 1, 2, 2, 2) for a mean of 1.43 and B has (1, 1, 1, 2, 2) for a mean of 1.25. Thus A is ahead in the world ranking, and B is ahead in the regional ranking. This happens because A is much ahead of B for some criteria and little behind for some others in the world ranking.
As gender is not declared during registration, it is inferred from the first name using a gender likelihood database. For low likelihoods, an exception file is used. If you think you are miss-classified, email Christian Zimmermann.
Sure, here, listed by month.
Nowhere, we compute them ourselves. See the series sections on the rankings pages.
  1. Are you registered with RePEc?
  2. If you registered this month, please wait for the next refresh of the rankings.
  3. Are you sure you would score high enough to be listed. Only the top authors are publicly listed.
  4. If you believe you should be in a regional ranking, do you have an affiliation listed? This should help locate you in a region.
  5. If your affiliation is not listed in the database (EDIRC), its URL is not in a geographically recognizable domain, and the same applies to your email address and your homepage, you cannot be attributed to a region.
There scores for each criterion are simply added. In the case of multiple affiliations, an author's score is distributed across affiliations. The only exception is the h-index, where it corresponds, for an institution, to the number of authors with an individual h-index at least that high.
Yes, they should register as well. As all individual scores are added, and not averaged by the number of authors, everyone contributes positively.
This document should satisfy you.
Author profiles do not contain information about the date of birth or the date of the last graduate degree. However, the date of the first publication, be it a working paper or an article is know. This date is used to determine who a young economist is. Quite obviously, there is a lot of measurement error for the youngest ones, both because of the uncertain dating, and because these economists have relatively few publications. More in the RePEc blog.
Authors do not declare what field(s) they work in, thus we have to impute them. We do that from the proportion of working papers announced through NEP that have been assigned to a particular field report. This proportion is then applied to the various ranking criteria and then aggregated. Authors (and institutions) are then ranking within the field. Note that these proportions may sum to more than 100% if papers have been selected for several NEP reports.
We use various algorithms to weed out anything that does not look like regular human activity. It turns out the vast majority of the traffic gets thrown out this way, in larger part due to the spoidering activity of many search engines (and some screen scrapers who would be better served to use the API. For more details, see the LogEc about page.
As they pertain to still young careers, there is a lot of noise in their measurement. The shorter the time span that is considered, the more likely the results could be spurious.
There can be several reasons:
  1. Some items were removed from your profile, either because you refused them or because the publisher deleted those records.
  2. Some ranking criteria are computed using a rolling window: abstract views and downloads consider the last 12 months, so if you had high numbers for those 13 months ago, your stats drop. Also, discounted citation counts and impact factors reset every January to account for the new year.
  3. You can drop in a country ranking if you added an affiliation in another country. Your score is distributed across country rankings according to the affiliation shares you indicate.
  4. Different versions of the same work have been consolidated.

Comments and requests for additions are welcome. Send them to Christian Zimmermann.
IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.