Geoscience Reference
In-Depth Information
4.3.3.1.
Population segregation dynamics in the city of Yaffo between 1955 and
1995
This example was mentioned earlier to discuss the form that the outputs of MAS
models can take. Starting from the distribution of Jewish, Christian Arab and
Muslim Arab populations in 1955, the objective of the authors is to show that it
suffices to formalize the preferences of individuals with respect to residences from
two points of view, in order to reproduce the redistribution of these populations over
the next 40 years. These two points of view are:
- the composition of the neighborhood on the cultural plan, making attractive,
for a given individual, environments in which individuals of the same cultural
profile as him are majority;
- the buildings' architectural style according to two types: the oriental style
preferred by Arab Israelis and the style by “blocks” preferred by the Jewish Israelis.
The model has been tested with different parameter values giving differentiated
weights to these two factors. As previously discussed, the outputs of the model are
apprehended at the level of the city of Yaffo, through two indicators: the share of the
Arab population in the total population, simulated at different dates, and Moran's
spatial autocorrelation index, indicating the segregation degree of the population
(that is the degree of spatial concentration of the population of same cultural type).
If the parameters associated with these two factors characterizing the preferences are
null, the share of the Arab population at the final date is largely underestimated. The
same happens when only the parameter associated with the cultural neighborhoods'
effect is not null. Only when the parameter associated with the architectural style is
not null, is Moran's index not very high, resulting, at some periods, in cohabitation
between Jews and Arab Israelis in the same buildings, a result which contradicts the
empirical observations. The authors conclude that the only simulations correctly
reproducing, at the same time, the developments of the share of the Arab population
in the total population and Moran's index, are those for which the parameters
associated with the two indicators are not null. The two parameters have to be
equivalent and relatively low in order to get simulated indicators close to the
observed indicators [BEN 02]. The authors emphasize particularly that it has not
been necessary to take into account the price of real estate as a constraint in
residential choices to obtain evolutions close to what has been observed.
4.3.3.2.
The “spiral” of school inequalities
Interested in the genesis of the social differences in terms of diplomas, Manzo
[MAN 09] developed a MAS model formalizing the individual decisions of students
and their families about the education level that they have as a goal according to
three families of hypotheses:
