Spatial scale, return and onward migration, and the Long‐Boertlein index of repeat migration

Long and Boertlein's (1990) measure of repeat migration can be used to infer the effects of repeat migration with data sets that did not specifically ask questions about such migration. While recent work by Rogers et al. (2003) evaluated how well this synthetic measure matched empirical fixed‐interval measures, the role of spatial scale remains unclear, particularly since scale influences migration levels, spatial structure, and the representation of the primary, return, and onward migration components within the overall flow. Motivated by the implementation of the American Community Survey and the concurrent need to adjust data from one to five‐year formats, and using the 1996 Canadian Public Use Microdata File and a custom tabulation from the 1996 Canadian census, the accuracy of the index is evaluated. In particular, the effects of spatial scale upon measures of primary, return and onward migration are described and evaluated, with the Long‐Boertlein measure contrasted with fixed‐interval measures of return and onward migration. Results show that the measure is sensitive to spatial scale.