RESUMO
PIP: The misinterpretations which can occur when cross sectional data rather than longitudinal data is used to assess patterns of migrant adjustment are demonstrated in a cross sectional, comparative study of housing quality among natives, recent migrants, and longterm migrants in Bogota, Colombia; Seoul, South Korea; and Surabaya, Indonesia. In cross sectional studies, duration of residence is used as a substitute for change over time, and observed differences in adjustment between recent and longterm migrants are generally attributed to length of residence. In actuality, these differences may be the result of changing patterns of selective migration (i.e., earlier migrants may have more education or higher skill levels on arrival or less education or lower skills levels on arrival than more recent migrants) or of selective remigration (i.e., the less successful migrants may leave the community and only the most succcessful go on to become longterm migrants or the most successful leave and only the least successful become longterm migrants). In other words, observed improvements or deteriorations in the socioeconomic status of recent and longterm migrants may reflect changes in migration patterns rather than changes brought about by the length of exposure to the urban environment. In the present investigation, information on a number of control and background variables was collected in order to determine the degree to which differences in housing quality among recent and longterm migrants were influenced by changing patterns of selective migration and by selective remigration. The cities represented diverse cultural and socioeconomic conditions. A multistaged cluster probability sample of male and female, recent and longterm migrants and natives were interviewed in each city. The respondents were 20-44 years of age, and the total sample size was 730 for Bogota, 978 for Seoul, and 606 for Surabaya. A distinctive, but comparable, housing quality index was constructed for each city. Control variables included marital status, age, household size, sex and a number of background variables, such as, education and farm/nonfarm origin. Information on housing investments and residential changes was also collected. For all 3 cities the mean differences in housing quality were slightly, but significantly, better for natives than for migrants. Housing quality for recent and longterm migrants was then compared using the control variables and multiple classification analysis. In Bogota and Seoul housing quality improved with duration of residence, but the relationship was not significant. Further analysis and information about remigration patterns in these cities suggested that, for Bogota, the relationship was influenced slightly by the fact that recent migrants tended to be more educated than earlier migrants and by the fact that there was a tendency for successful migrants to remigrate. However, some improvements appeared to be due to duration of residence. In Seoul the factors influencing the relationship were more difficult to assess, but the conclusion was reached that only a slight improvement in housing occurred over time. In Surabaya housing quality was significantly better for longterm migrants than for recent migrants among migrants with a farm background. The reverse pattern was observed among migrants with a nonfarm background, but this relationship was not significant. Further analysis and a knowledge of local migration patterns suggested that the relationship observed for those with a farm background was influenced by remigration of migrants with low socioeconomic status and that housing quality for migrants in Surabaya had not improved over time.^ieng