Dados Bibliográficos

Resumo

Conventional methods of model construction and testing are not well suited to the features of large longitudinal survey data sets. Conventional methods assume (1) equal time intervals between observations, (2) simultaneous observations, and (3) that missing observations are rare. As a result, the analysis of whole multiwave sets of longitudinal surveys becomes virtually impossible as the num ber of waves increases. This paper poses a research question about how a per son's work experience affects his or her earnings and shows how the Cumu lative Experience Method (CEM) can provide an answer to the question using all available information in a longitudinal data set. CEM interpolates a person's experience between obcervation points and weights these inferred observations by the inverse of their expected error. The linear interpolation and weighting pro cedure of CEM accommodates easily to missing observations where these occur between earlier and later observations.

Referências Citadas