statistical analyses of war and peace
scale, and may extend further back in time; that started by Peter Brecke at the Georgia Institute of Technology extends down to magnitudes of around 1·5 (about 30 deaths) and backwards to the beginning of the 15th Century. Brecke was also an early proponent of the idea that analytic progress can only be made if a taxonomy of confl icts can be developed for their classifi cation. Spatial factors were identifi ed by
A geographic civil war model in the International Confl ict Research group’s GROWlab Gambling on military-economic odds
An analyst at a US policy consultancy, demonstrating the use of her very substantial SAS installation, summed up the role of her own confl ict analysis work with the comment that ‘We, [the US] defeated the Soviet Union by driving up the necessary cost of preparedness for war, and the cost of third-party client war, to the point where we could afford it and they couldn’t... that sort of data analysis, as circumstances change, continues to ensure my funding.’
The data sets she handles bring together a huge variety of economic and military indicators for every country on earth, plus several blocs (such as the EU and the League of Arab States) and subdivisions with daily updated policy and politics transactional fl ows. Her outputs become partial input for case simulation scenarios exploring potential links between politico-military policy and international shifts in economic balance.
the engagements between government and insurgency over the past couple of weeks in an SPSS worksheet, maintaining a live curve fi t as the data is added. Though they are all in the lower reach of Richardson’s scale, between zero and 1.7, the Poisson fi t gets more defi nite with every data point.’ This sort of microconfl ict, spatially
and temporally confi ned as well as low in magnitude, would be described by CoW as extrastate, having ‘at least one major participant... not be a member of the state system’, but would not be included in the CoW database which imposes a lower limit of 1,000 battle-related casualties per year. Other databases go lower down the
10 SCIENTIFIC COMPUTING WORLD On the usefulness of such work, she points out
that, ‘A multi trillion dollar cost, around 20 to 50 times the administrations’ estimates, publicly aired at the time, was one of the outcomes we predicted as a result of intervention in Iraq. The actual cost is greater than we predicted, but only a few per cent above our upper sigma. The rise in the price of oil is another example. There again, we were slightly too conservative – but by the same sort of margin and closer to actual events by several thousand per cent, in retrospect, than the optimistic forecasts upon which decisions were actually based in practice.’ As this article is fi nalised she is engaged on updates of comparative analysis, following a UN resolution, of geo-economic and politico military data relating proposed degrees of aerial interdiction (a so called ‘no fl y zone’ at various operational levels) in Libya to those in Iraq in the 1990s.
Richardson as predictors of confl ict – particularly the length of shared borders between belligerents, a variable included in the CoW databases. Distribution of vital resources (oil being an obvious example, but increasingly rivalled by water) clearly play a major part. Spatiotemporal correlations are now generally accepted as essential to most social interaction modelling and analysis, and peace/confl ict research is no exception. The Geographic Representations of War network (GROWnet) is an affi liation of researchers, primarily in Europe, which pursues this aspect specifi cally. The International Confl ict Research group (ICR) at the Swiss Federal Institute of Technology’s Center for Comparative and International Studies in Zurich is one example of a research centre which embraces both aspects, maintaining databases, tools and groups, most of which have a geospatial dimension. GROWlab is a freely downloadable
simulation tool; the GeoReferencing of Ethnic Groups (GREG) data project is one of several exploring the relationships between confl ict and regional grievance; WarViews is a visualisation tool which runs within a browser or (more revealingly) Google Earth. Within the past few months, Weidmann (of the ICR) and Ward have described10
improved modelling
of Bosnian civil war period confl icts, using spatial correlates and an R based set of progam libraries. Previous work by GROWnet and collaborators has explored the underlying structure of violent group interactions with states removed from consideration.
Superimposition of a Voronoi infl uence tesselation (yellow lines), and socio-ethnic interfaces (blue lines) onto a map of high incidence (red) and low incidence (green) confl ict areas, in a southern Sudanese study area
Discovering a correlation A small group of students within Libya are, as I write this, running an intensive Wolfram Mathematica-based analytic study trying to tie together Richardson’s border lengths with the distribution of nonstate factions. Their hypothesis is that prediction of incidence, magnitude, outcome and subsequent infl uence on other confl icts can be linked to the length and complexity of factional interfaces. They are gathering live data as the civil war progresses and comparing it with archived examples from other theatres. They are also exchanging notes and results with another similar group in Sudan
www.scientific-computing.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48