Optimal Matching Analysis: A Methodological Note on Studying Career Mobility.

OCCUPATIONAL mobility; LABOR market; CLUSTER analysis (Statistics); MULTIVARIATE analysis

HONG Kong (China); CHINA

The article examines the use of optimal matching analysis (OMA) to trace mobility paths people take to achieve upward career mobility in Hong Kong, China. OMA compares and classifies data in the form of whole sequences of events. It investigates whether there are systematic differences between the users of various mobility paths. OMA is essentially a technique of cluster analysis for sequence data. Its result is sensitive to, among other things, the coding of sequences-the job categories used, and the definitions of substitution, insertion, and deletion costs. By applying OMA to career history data, this article shows that there are four typical mobility paths in Hong Kong. Like all techniques that involve pairwise comparison procedures, computation time required by OMA is roughly proportional to the square of the number of cases. Although this article employs cluster analysis, one should note that clustering is not the core or most important part of OMA. Other multivariate techniques, such as multidimensional scaling, can also be used for modeling the similarity scores obtained in pairwise comparison.