Wednesday, 21 December 2011

Leiden Global University Ranking

There are four well-known global university rankings around already (linked here) but my, not unbiased (more later…), view would be that the new Leiden Ranking (available here) seems to be preferable. The two main differences are: (a) the Leiden ranking uses only objective data, such as impact of research – not dubious reputation ranks which only tend to confirm stereotypes about good/not-so-good universities. (b) in the default option the rank normalises for university size, thus giving medium-sized, small or even very specialised universities a fair chance.
So, how does the Leiden ranking compare with the other rankings? First, of all, similar to the latter, the US is dominant in the top 30: here 27 universities are from the US plus two from Switzerland and one from Israel. The top UK universities are Cambridge (31), London School of Hygiene and Tropical Medicine (33), Oxford (36), Durham (42), Imperial (54), UCL (59), Edinburgh (66), St Andrews (71), Bristol (73), Dundee (78), East Anglia (80) and York (97). 
This does not look too wrong to me (noting that my current university, Durham, performs quite well, as do my previous ones, Edinburgh and East Anglia). In particular, as mentioned earlier, I find it positive that top specialised institutions such as the London School of Hygiene and Tropical Medicine and top smaller but general universities such as Durham or St Andrews are rewarded, i.e. ranked better than big but medium-quality universities.

Sunday, 18 December 2011

Counting words in US Supreme Court decisions

There is an interesting new tool available, called Legal Language Explorer. It is similar to the Google Books Ngram Viewer (see previous post), but here the results show the number of hits in the US Supreme Court decisions between 1800 and 2005. In my own research I have done something similar, though based on a hand count, for the highest German and English courts (papers here and here).

I started playing a bit with the Legal Language Explorer - always using the “normalize” option, ie the results show the “average count per case”:

(1) blue: English law - orange: House of Lords
Note: as expected, more important in early 19th century but recently English law a bit up (why?).

(2) blue: efficiency - orange: fairness
Note: both fairly modern terms in judicial language; perhaps a bit unexpected that “efficiency” has become slightly less popular in last few decades

(3) blue: economic - orange: social – brown: moral
Note: a different way of presenting the data. Initially “moral” more popular; since 1930s “economic” and “social” on the rise.

(4) blue: economics - orange: philosophy
Note: just 20th century since 19th century data seem fairly unreliable; today "economics" more popular than "philosophy".

(5) blue: rights - orange: duties
Note: again, only 20th century; “rights-talk” increasing, which presumably makes sense.

Monday, 12 December 2011

Why are the British so Eurosceptic?

I have been wondering about this question for a while - and given the so-called "veto" to the European Fiscal Union now again. Of course, as a caveat it is worth indicating that the "Brussels bureaucracy" is not popular in other EU Member States either, and that here too national identities seem to be stronger than the EU one (see only the French and Dutch rejections of the EU constitution). Still, even listening to mainstream media such as the BBC one cannot help noticing that there is something special about the aversion against the EU in Britain. So, really, "what's wrong with them" (no offence)?

To get a proper answer, one may use country-level survey data such as Eurobarometer as dependent variable, and then contemplate a number of explanatory variables relating to a country’s geography, history, culture, politics etc in order to explain anti-EU feelings, for instance:
  • geography (e.g., dummy for island; distance from centre of EU)
  • age of country; stability of domestic form of government (or last dictatorial government)
  • colonial power (e.g., maximum extent of former colonial empire)
  • death rate in Second World War (or both world wars)
  • cultural attitudes (e.g., mean difference from other EU countries)
  • socio-economic attitudes (ditto)
  • distinctiveness of legal tradition (ditto)
  • proportion of Eurosceptic media (but reverse causality?)
  • language skills (e.g., mean number languages spoken); perhaps also ethnic diversity
  • relatives in other EU countries (or proportion to relatives in non-EU countries)
  • prevalence of EU institutions in country
  • net transfer payments from/to EU
  • corruption score of domestic political institutions (or other data on popularity of domestic politicians etc)
  • political variables such as voting system (encouraging small parties), party funding etc.
  • general variables such as GDP per capita or population (relative to EU mean) 
In political science there are a few studies on similar questions (e.g., here, here, here), but not exactly what I'm interested in. One problem may be that the number of observations is fairly small (27 countries), whereas all of these reasons (and more?) may really have a cumulative and interdependent effect. In any case, assuming that these factors are relevant, the most interesting question is presumably another one, namely, whether in the UK one would expect these factors to become weaker, or even stronger,... see here.

Monday, 5 December 2011

REF 2014: what did the law sub-panel members submit in 2001?

A week ago I posted on the publications that the REF 2014 members of the law sub-panel submitted for the RAE 2008 (post here). I talked to a friend about it and he suggested that the REF-panel members are “old professors who already made it” and therefore do not feel the pressure to publish in the top journals any more. Well, this can be checked by going back further in time. Thus, below the data from the RAE 2001 (from here) and how they compare with the 2008 data of the REF law sub-panel members. We have 54 pieces (for 2008 it was 55):
  • 2001: 46% journal articles with 52% of these in general journals (including the socio-legal ones) and 48% in specialised journals. Most “hits”: six for the Modern Law Review and two each for Journal of Law and Society and Socio & Legal Studies (no hits for LQR, LS and CLJ; one for OJLS). - 2008: 55% journal articles with 77% of these articles in specialised journals, and 23% in general ones. Overall, no journal had more than two “hits” (these were Modern Law Review, Legal Studies, and Journal of Law and Society; no hits for LQR, OJLS and CLJ).
  • 2001: 26% book chapters with 36% of these in OUP, CUP and Hart books (2 OUP, 1 CUP, 2 Hart). - 2008: 25% book chapters with 88% of these in OUP, CUP and Hart books (8 OUP, 2 CUP, 2 Hart).
  • 2001: 24% books with 23% of these OUP books (no Hart or CUP books). - 2008: 16% books with 55% of these OUP, CUP and Hart books (2 OUP, 2 Hart, 1 CUP).
  • 2001 and 2008: 4% government reports (ie 2 pieces).
The interpretation has to take into account that not only the publication choices of the REF members may have changed but, what I would regard as crucial, also the book and journal market. To explain:
  • Books and book-chapters: the overall share (about 50%) is relatively unchanged, yet, in 2008 more OUP-CUP-Hart books. This is likely to reflect that CUP only got more involved in law publishing in the last ten years. Moreover, in 2001 it may have been more common to submit teaching or practical books, whereas now the clear preference is for academic publications.
  • Journal articles: the reduced share of the general journals can best be explained by the emergence of many new specialised journals in the last ten years (e.g, in my field all major journals are fairly new: JCLS, EBOR, ECFR, CMLJ, LFMR). Overall, the submissions are also fairly dispersed in both years (with the only exception the MLR in 2001 but not in 2008), thus there does not seem to be a preference for a small number of selected journals – which may be most interesting point for everyone involved in the REF 2014.