Beware the 'Ivy League'

Robert May warns against the sloppy thinking which confuses famous names with research capabilities

In the sometimes fraught discussion of the United Kingdom's greatly expanded tertiary education system and how we might provide the infrastructure for research - the earlier "well-found laboratory" - for those who will use it best, the Ivy League metaphor has become misleadingly ubiquitous.

The Ivy League is portrayed as a privileged elite seemingly enjoying generous funding by right. "We don't want an Ivy League" is taken to mean we do not want the infrastructure for research provided at a supra-departmental level, in a defined set of research universities.

I passionately agree that we do not want any such thing. But the Ivy League is not like this. The League consists of eight private universities in the northeast United States, who play sports with each other. All pride themselves on excellent and intimate undergraduate teaching. As research institutions, however, they differ hugely in size and quality.

One measure of quality is to take individual fields of science and engineering, and rank university faculties by their average citations per paper, over 13 years from 1981 to 1993 (Science Watch, October and November/December 1994). For all its imperfections, this allows smaller departments to compete fairly against those at larger universities which turn out many more papers.

A recent study ranked the top ten US universities in each of the Institute of Scientific Information's 21 rather arbitrarily defined fields of science and engineering. It then produced an overall ranking of universities based on the number of top ten appearances. Stanford led, scoring 17 out of a possible 21. Of the eight Ivies, Harvard and Yale tied for second (on 13), Cornell was seventh, Princeton ninth, Columbia 14th, University of Pennsylvania (Penn) 19th, and Brown 42nd. Dartmouth was unplaced with no department in the top ten in any field.

Some of the Ivies are large, having medical, law and business schools. Others are smaller; Princeton has no such professional addenda. So the quality league table can be re-ranked by the average placings of universities, calculated as the sum of the rankings in the 21 fields, divided by the number of such fields in which the university is represented. On this basis, Harvard is first (averaging just better than third place), Princeton third, Yale sixth, Cornell ninth, Penn and Columbia in the teens, Brown low, and Dartmouth again unranked.

The Ivy League universities' ranking by total intake of Federal research and development money in 1995, gives another picture, because size now becomes important. The big public universities top this ranking, with Cornell ninth overall, Columbia tenth, Harvard 11th, Penn 13th, Yale 16th, Princeton 66th, Dartmouth 96th, and Brown 98th.

As a combination of quality and sheer size, ranking by total number of post-doctoral appointees in 1995 is interesting: overall, Harvard ranks first, Yale sixth, Penn seventh, Cornell 14th, Columbia 22nd; Princeton moves up to 31st, while Brown and Dartmouth remain at 93 and 94 respectively.

Nor are these rankings static. Over the past 30 years or so, Princeton has completed the shedding of its Scott Fitzgerald image and moved up the research league tables, while some other Ivies have fallen back a little. Other private universities, such as Stanford, have also moved up, as have some big state universities. In general, I believe there has been a Red Queen effect, with a need to run faster to stay in the same place.

In short, the Ivy League universities have some important features in common, not least their private status, endowed funds (although these also vary greatly), and commitment to teaching. But as research institutions they are very diverse. To characterise them as if they stood for the "R" in the "R,T,X" scheme propounded in the UK in the mid-1980s, is wholly to misunderstand them. None the less, the diversity of aspiration and accomplishment within the Ivy League has interesting messages.

The accompanying table makes plain just how large has been the welcome sea- change in UK higher education over the past 20 years or so. The country whose "before and after" figures are closest to ours is South Korea. Small wonder that university infrastructure is under pressure. At present, higher education funding council money is distributed among the universities on the basis of assessments of the research strengths of individual departments. It costs time, energy and emotion, but I think that it is working reasonably well.

It has been suggested that assessment at department level makes it hard, if not impossible, for an outstanding individual or group to flourish in an otherwise weak department. All this prompts the question about how diversely research strength is spread among UK universities.

Science Watch has recently applied its citations per paper approach to rank quality of research output among UK universities in particular fields. The picture is instructively diverse. Because the UK science base is smaller than in the US, this study (Science Watch January/February 1997) ranks the top three UK university departments in each of the ISI's 21 fields of science, engineering and medicine, themselves a motley mixture varying in size and proportion of applied to basic work.

There are 25 different UK universities represented in 63 placings thus generated. Oxford leads with a score of 11 out of a possible 21, Cambridge is second with seven, followed by Imperial College with five; University College London and University of East Anglia each with four; Durham, Lancaster and Sussex with three; five universities with two; and 12 with one. The overriding impression, in the UK as in the US, is of a complex distribution of strengths among a wide range of institutions. No unambiguously distinct elite emerges to which the unsuitable tag "Ivy League" could be attached.

Sir Robert May is chief scientific adviser to the government and head of the Office of Science and Technology.