PEGC Nordic Doctoral Network Course: 'Statistical Analysis of Spatial Data'

Course directors: Roger Bivand (Norwegian School of Economics and Business Administration) & Kristian Skrede Gleditsch (University of California San Diego & CSCW)

Please note: This page refers to an event that has already taken place.

Time: Mon 21 (10:10) - Fri 25 (15:00) June 2004
Place: Dept. of Sociology and Political Science, NTNU, Trondheim

The course is organized by Department of Sociology and Political Science, Norwegian University of Science and Technology (NTNU) for the Nordic Network on the Political Econo­my of Governance and Conflict (PEGC), in collaboration with the Centre for the Study of Civil War, International Peace Research Institute, Oslo (PRIO) and the Standing Group on Political Geography of the European Consortium for Political Research.

This course will consider the problem of spatial dependence in applied research in the social sciences. Spatial dependence occurs whenever observations for different units are related to one another so that the value for one unit cannot be considered independent of the value of other units. In many areas of social science, observations have a clustered structure, where outcomes for one unit seem to be related to outcomes for other units, and where the strength of this dependence is greater for near neighbors than those further apart. For example, conflict often displays tendencies to diffuse to neighboring states, and countries are more likely to experience transitions from autocratic rule or sustain democracy if located among other predominantly democratic countries. The spatial clustering of the data violates the assumption of independence in standard regression models, and can make inferences problematic, as is the case for dependence over time. However, whereas observations gathered over time have a clear ordering, patterns of spatial dependence can be more complex. Assessing spatial dependence will often reflect interesting substan­tive features of the processes generating the observed data, and increasingly researchers are exploring hypotheses on spatial processes.

This five-day course will provide a self-contained introduction to spatial statistical analysis for social scientists. The problems of spatial dependence have been known for a long time, but many of the solutions and contributions to this field have been relatively technical and inaccessible to applied users. Although recent software development has made it much easier for practitioners to apply spatial approaches, this material is rarely covered in general research methods training for PhD Students. The course will cover properties of geographical data, working with data with a spatial representation (including the introduction of data structures typically used in geographical information systems), and how to deal with spatial dependence in regression models. Many of the applications considered in the course will be related to research on governance and conflict. The course is open to interested doctoral students in all disciplines. It does not require background knowledge of spatial statistics or statistics beyond linear regression models. We intend to distribute copies of software and data sets used to participants on CD-ROM and the course home page. The teaching format of the course will be by workshop, lectures, practical lab exercises, and seminar-style presentations.



Course Schedule

  • Monday 21 June, 10:15-12:00 –  “hello, world”

Lunch, 12:15-13:15.

  • Monday 21 June, 13:15-16:00 – GIS and spatial data; Maps, visualization and ESDA

A.  Monday 21 June, 16:15-18:00 Practical lab I


  • Tuesday 22 June, 8:15-10:00 – Antecedents of spatial relationships
  • Tuesday 22 June, 10:15-12:15 – Spatial weights

Lunch 12:15-13:15

  • Tuesday 22 June, 13:15-15:00 – Impact of spatial dependence

B.  Tuesday 21 June, 15:15-18:00 Practical lab II


  • Wednesday 23 June, 8:15-10:00 – Indicators of spatial association
  • Wednesday 23 June, 10:15-12:00 – Spatial regression: models

Lunch 12:15-13:15

  • Wednesday 23 June, 13:15-15:00 – Spatial regression: tests

C.  Wednesday 21 June, 15:15-18:00 Practical lab III


  • Thursday 24June, 8:15-10:00 – Spatial regression: estimation
  • Thursday 24 June, 10:15-12:00 – Spatial regression: advanced topics

Lunch 12:00-13:15

  • Thursday 24 June, 13:15–16:00 – Student presentations and discussion

D.  Thursday 21 June, 16:15-18:00 Practical lab IV


  • Friday 25 June, 09:00-12:00 – Student presentations and discussion

Lunch, 12:00-13:00

  • Friday 25 June, 13:00-15:00 – Roundtable discussion on issues arising from the course



Roger Bivand (Economic Geography Section, Department of Economics, Norwegian School of Economics and Business Administration, Bergen)

Halvard Buhaug (Department of Sociology and Political Science, Norwegian University of Science and Technology & Centre for the Study of Civil War, International Peace Research Institute, Oslo)

Kristian Skrede Gleditsch (Department of Political Science, University of California, San Diego & Centre for the Study of Civil War, International Peace Research Institute, Oslo)

Jan Ketil Rød (Department of Geography, Norwegian University of Science and Technology)