Thursday, July 7, 2011

h(5,2) - the Best Citation Metric for Economics?

Glenn Ellison has written a paper titled: "How Does the Market Use Citation Data? The Hirsch Index in Economics" on the use of Hirsch type citation metrics in economics. An author's Hirsch index is equal to h if they have published at least h papers cited at least h times.

Ellison argues that generally economists write fewer papers than natural scientists such as physicists so that some top economists have relatively low h-indices not because they aren't cited a lot but simply because they don't have a lot of papers that could be cited. So he thinks we should focus on a narrower set of highly cited papers. Ellison introduces h(a,b) indices where an author's index is h if they have at least h papers with a*(h^b) citations.

Ellison uses a sample of 166 relatively young economists from the top 25 U.S. departments using Google Scholar data. The average h-index in the sample is 18, the average h(5,2) index is 4.5, and the average h(10,1) index is 7. Ellison finds that the h(5,2) index predicts best what NRC ranked department an economist will be at though really there don't appear to be big differences in how well the indices fit the data and the h(10,1) index (Wu index) is about as good . This is in effect estimating the utility function of hiring decisions in top U.S. economics departments.

Using Google Scholar, as Ellison does, my h(5,2) index is 5 - 5 papers with at least 125 citations each. My Wu index is 10 and my regular h-index is 27. But I'm not a very typical economist...

No comments:

Post a Comment