A. Sachs

Studies on the spatial phenomenon of population and employment lie within the broad scope of urban and regional science. Regional science was conceptualised as a major field of study in the 1950s. It is concerned with the study of social issues with regional or spatial dimensions. In his book, Isard (1975) defined regional science as:

• studies of systems of places, locations, cities, urban regions, patterns of human settlement, industry and economic activity, jobs income generation and receipts and resource use

• the study of diverse organisational and institutional structures of society as they govern the behaviour and spatial distribution of population and economic activities

• the systematic study of the time-space patterns of systems and the ways in which social problems associated with these patterns can be effectively attacked and resolved

• the study of joint interactions of social, political and economic behavioural units and the physical environment with meaningful regions (Isard, 1975).

Regional science extends the traditional social science. It achieves this by using a geographical or spatial perspective to examine human behaviour and activities (for example, population and employment). Human activities play out over geographic space,

space is an important property and context to study the patterns and processes involved in understanding human activities in regional science.

2.1.1 Defining space in geography

The idea of space has been of interest to philosophers and scientists for much of human history (Harvey, 1969). From a scientific perspective, space can be defined as an absolute term that exists for locating objects, but is independent of any matter. In the context of regional science, space is defined from a relative framework that exists only with reference to the spatial entities and processes under consideration (Meentemeyer, 1989). The relative view of space focuses on objects as the subject matter with space being measured as relationships between objects. It is a conceptual framework in which the objects are spatially referenced; and it can be used to compare and quantify the distance between objects, their sizes and shapes (Marceau 1999). Most work in human geography involves a relative view of space, because much of this work involves spatial processes, such as, migration and commuting patterns, or the dispersion of population and economic activities. (Harvey 1969; Abler et al., 1971).

2.1.2 Spatial scale in regional science

Scale is a fundamental concept in geography. The use of spatial scales in geographical studies was initially concerned with measuring the large numbers of individuals over large spatial regions. Geographers usually use a generalised measurement of the individual.

Spatial scale provides a link between an object distributed in space and its simplified representation. Scientific inquiries can incorporate scale to obtain the generalised representation and capture the pattern of objects at a level of detail for particular investigation purposes (Gibson et al., 2000).

In a general sense, spatial scale refers to the degree of spatial detail (spatial variation) at which entities, patterns and processes in space can be observed and analysed (Goodchild, 2007). In absolute terms, scale can be defined as operational system used to partition geographical space into operational spatial units. Examples are provided in the use of census and administrative units, and any zoning system defined for a particular study.

When we focus on studying geographical phenomenon, scale becomes a variable intrinsically linked to spatial entities and the processes to be investigated (Marceau, 1999;

Atkinson and Tate, 2000). By imposing an appropriate scale, the specific spatial

characteristic of a geographical pattern or processes can be more effectively and precisely studied and understood.

Gibson et al. (2000) gives four fundamental areas where scale is important to the regional science:

• the identification of patterns and problems

• the explanation of observed patterns

• the generalisation of propositions made at level of a scale to another level of the scale

• the optimisation of some process or function.

The importance of selecting an appropriate spatial scale for regional analysis has been widely acknowledged (Goodchild and Quattrochi, 1997; Marceau, 1999). The conceptual development of the spatial scale in social science dates from the 1930s. Social scientists first underlined the importance of the scale problem in the regional studies (Gehlke and Biehl, 1934). This was followed by human geographers who were concerned with the patterns and relationships of the geographical phenomenon that varied substantially according to the spatial scales used (Yule and Kendall, 1950). Robinson (1950) defined the term known as ‘ecological fallacy’. It means that the results obtained from data in the aggregate scales lead to false inferences in the relationships at the disaggregate scales. An ecological fallacy occurs when it is inferred that the results based on aggregate data are applied at the more micro level that collectively form the aggregated group. McCarthy et al.

(1956) and other geographical researchers, such as Blalock (1964) and Clark and Avery (1976), also demonstrated that measurements to describe relationships and processes are very much scale-specific. In addition, spatial analysis results are most dependent on the spatial scale of the observation chosen. The scale issue was extensively and explicitly identified by Openshaw and Taylor (1981) who coined the phrase for the modifiable areal unit problem (MAUP). The MAUP describes the fact that the way a geographical region is partitioned into area units determines the result of the analysis. The analysis result based