Sampling, Maintenance, and Weighting Schemes for Longitudinal Surveys: a Case Study of the Textile and Clothing Industry

A longitudinal survey of textile and clothing firms in Emilia-Romagna (Italy) was conducted using a ``replenished'' panel survey to producereliable cross-sectional estimates. The firms (sampling units) were stratified according to size and age, and in all phases, dead and/or emigrated firms were considered as having already left the sample, while new units selected from newborn and immigrant firms, were considered as having already entered the sample. Population dynamics change firm composition over time and the methods of cross-sectional estimation must be adapted to the longitudinal features of the samples. This paper reviews some weighting schemes used in pursuing the aims of repeated sample surveys and presents application trials conducted on real and simulated data.