The Panel Study of Income Dynamics.

The Panel Study of Income Dynamics (PSID) offers much useful information for life course analysis because it follows the same people for a long period of time and it also follows individuals (called splitoffs) who leave their original families to form new families of their own. It includes an increasingly large sample of children for whose parents the PSID has both past and current information. Cumulative records of illnesses or unemployment in the same individuals or families allow new ways of studying their incidence. Since 1968 the PSID has interviewed annually a representative sample of about 5000 US families. A least 1 member of each family in the sample was in 1 of the original families interviewed in 1968. The study has followed both the original intact families and all members who have left the family since 1968. The 13th wave of data included about 6500 families. The PSID originally oversampled low-income and minority households. PSID data allows analysis on many different levels--households individuals divorced spouses and male heads for example. The PSID differs from the census in that it codes long-term cohabitors as families. Splitoffs who return to their original families continue to be treated as independent units. Sample members who die are replaced by their own children. The cumulative response rate after 14 years is 47% but if this rate is based only on the families interviewed in 1968 it becomes 67%. The PSID asks questions about income sources; family composition; employment; earnings; time spent working commuting and doing housework; food expenditures; housing; and migration. Its findings contradict many of the stereotypes built up from years of cross-sectional analysis. Peoples economic environment is not stable but volatile. Large numbers of workers or families are occasionally poor on welfare or in certain sectors of the labor market; fairly small numbers are persistently in these states. Frequent changes in family composition play a role in much of this volatility; these changes also create havoc in cross-sectional analyses of groups based on their family relationships.