ABSTRACT
The construction of explanatory accounts of developmental phenomena rests on both traditional manipulative experimentation and alternative methodological approaches, data sources, and analysis techniques. Because design and analysis needs vary according to the level of theory development and the experimental accessibility of the constructs employed, making the best use of analysis procedures involves some matching of one to the other in relation to a range of exploratory and hypothesis testing objectives. With primary emphasis on multivariate correlational analysis, selected aspects of convergence and divergence of theory and methodology are examined with special attention paid to the development and testing of hypotheses deriving from developmental concerns. Structural modeling techniques offer the promise of a more systematic way for develop-mentalists to organize and exploit multivariate correlational data in building and testing their theoretical conceptions than has been available heretofore. The techniques are described and some examples illustrating both measurement and causal analysis aspects of structural modeling are presented and discussed in relation to developmental research issues.
