SimSeg

A program simulating the dynamics of residential segregation by ethnic and socioeconomic status

The SimSeg program simulates social processes that create and maintain residential segregation between ethnic and socioeconomic groups in urban areas.

The simulation is "agent-based". The behavioral agents are "virtual households" that reside in a "virtual city". At the beginning of each simulation "experiment", the SimSeg program creates a city and populates it with households. It assigns the households ethnic backgrounds and socioeconomic status scores to yield user-specified city-wide ethnic mixes and group-specific status distributions.

The SimSeg program also assigns preferences for various residential outcomes (e.g., housing quality, neighborhood status, neighborhood ethnic composition, etc.). User choices determine which preferences are active, what specific "targets" (i.e., goals) are associated with each preference, and how much "weight" is given to each preference.

During the course of the simulation, the SimSeg program randomly selects households and gives them the opportunity to search for, and possibly move to, new residential locations in the city that better satisfy their housing and neighborhood preferences. However, a given household's chances of improving their situation may be constrained by the household's ability to afford higher quality housing or by other factors such as housing discrimination.

As households move about, the SimSeg program presents and continuously updates a graphical display of the city "landscape". The color and shading of the housing units in this landscape indicates where households with different status and ethnic characteristics reside. As the example below shows, landscape images give the user an immediate and intuitive sense of the patterns of residential patterns that emerge over the course of a simulation experiment.

SimSeg also records the values of summary measures of city-wide segregation and neighborhood characteristics. The program allows the user to examine these outcomes using graphical and/or tabular presentation options.

For example, the figure presented below tracks the values of the index of dissimilarity (a widely used summary measure of segregation) over the course of an example simulation experiment.

I first developed SimSeg to use in the classroom to help illustrate theoretical principles relating to theories of residential segregation. Over time I have extended and refined the program and now also use it to explore theoretical questions about how residential segregation may be affected by different factors ranging from micro-level behavioral processes of residential choice to constraints on choice imposed by processes of discrimination to conditioning factors of demographic and urban structure.

I developed this web site to make it convenient for people to learn more about the program (after they see it used in a presentation or read a paper that presents results obtained using the program). In addition, I post a version of the program (it can be obtained via links provided at the end of this document). Thus, people who are interested (and who are willing to go to the trouble to do so) can download the program and learn more about it by running it themselves.

For those who might consider trying to use the program, I note that SimSeg allows users to conduct model-based experiments to explore the impact that different factors have on residential segregation. SimSeg allows the user to design simulation "scenarios" by using menus to specify the settings of factors such as residential preferences, institutional constraints on housing search and choice, and the city's demographic and urban structure. Then the user can "run" a simulation experiment based on these settings and observe the segregation patterns that emerge (or do not emerge as the case may be).

Furthermore, after a simulation experiment has progressed for a period of time under a given "scenario", the user can modify the scenario and continue the experiment to observe how changing parameter settings in the model affects segregation patterns.

In the interest of brevity, I provide only a cursory description of the features and characteristics of the SimSeg program here. Briefly, the SimSeg program allows the user to explore how three broad classes of factors affect segregation: (a) aspects of the city's demographic and urban structure, (b) the preferences that guide the residential choices that households make, and (c) institutional forces that constrain or inhibit households' efforts to identify and move to preferred residential locations. In each of these broad areas, there are literally dozens of simulation parameters that the user can manipulate to explore its effects. Thus, the SimSeg simulation permits users to explore a wide variety of simulation scenarios and investigate the kinds of segregation outcomes that emerge under each one.

This concludes this introductory description of the SimSeg program. Follow the links provided below to learn more about the program and/or to explore results that it generates.

	1. Landscape Sequences. A time sequence of images depicting the city "landscape" and the changes in residential segregation that occur over the course of an example simulation experiment.
	2. Trends in Segregation Measures. Figures tracking the values of selected measures of city-level residential segregation over the course of an example simulation experiment.
	3. Trends in Neighborhood Characteristics. Figures tracking the values of selected neighborhood characteristics over the course of an example simulation experiment.
	4. Residential Dissatisfaction. Figures depicting how group preferences for housing quality, neighborhood status, and neighborhood ethnic composition translate into dissatisfaction scores in an example simulation experiment. These dissatisfaction scores play a central role in determining the residential choices that households make.
	5. Status Distributions and Inequality. Figures depicting the status distributions for the ethnic groups in an example simulation experiment.
	6. Urban Structure. Figures depicting selected aspects of urban structure for an example simulation experiment.

Return to Mark Fossett's Home Page