"The decomposition approach is suggested by the work of David Schum and Peter Tillers (Schum, 1992; Tillers and Schum, 1992 [P. Tillers & D. Schum, "A Theory of Preliminary Fact Investigation," 24 University of California at Davis Law Review 931 (1991)]), who have used 'cascaded inference' models to describe the judgments required to evaluate hearsay evidence. Their models show how a 'global' judgment regarding the overall value of a piece of hearsay evidence can be decomposed into a series of more 'local' judgments about such matters as the observational sensitivity, objectivity and veracity of the declarant and hearsay witness. These models are potentially useful to psychologists because they provide a means for testing the logical consistency of 'global' and 'local' judgments about hearsay (Thompson and Pathak, 1999). A psychologist who is interested in whether information about the reliability of a declarant is correctly integrated into people's global judgments about the value of hearsay, for example, can test to see whether global and local judgments are equally sensitive to that variable.
"The global-local comparison can also yield helpful information about the reasons for people's insensitivity to important factors. If people are insensitive to a particular factor because they fail to appreciate its importance, then both local and global judgments should be insensitive. On the other hand, if people appreciate the factor but are insensitive due to a failure to correctly integrate information about it into their judgments, then the local judgments should be more sensitive to the factor than global judgments."
The dynamic evidence page
Evidence marshaling software MarshalPlan
It's here: the law of evidence on Spindle Law. See also this post and this post.