distribution. One of the great virtues of structural equation models is that they permit the quantification of causal and noncausal sources of statistical relationship. The present article discusses efficient matrix methods of computation for effect decomposition and extends these methods to models with unstandardized variables and to nonrecursive models. An appendix includes a computer program, written in APL, which implements the techniques described in the article.

On behalf of:

In one of our SIG sessions at New Orleans, an interesting discussion took place about the notion of learning as something taking place between participants in interaction within the constraints present. To me, those constraints, when taken as constraints through other people we connect with, suggested a rather limiting view on learning and development. In my view, constraints imply something static (cf. Valsiner, 1998). Such a description, without involving persons as active, engaged participants, embodies in my opinion a Œspectator theory ‚ (Dewey, 1960, see Doll, 1993; Phillips, 1995). Wanting a different description of social learning, as an active constructive process of dynamic complexity, one had better describe it in terms of reciprocal influences by reciprocal adjustments between interacting actors (Kinderman & Skinner, 1992; Rommetveit, 1992). Interaction described as a jointly organized activity. I call this kind of learning in interaction reciprocal learning (cf. Salomon & Perkins, 1998). To show the complexity of that kind of learning and of the effects, which may potentially be non-linear, one needs a generative theory of learning. With this generative theory, one may focus on interaction as a process of producing knowledge,

Solo asi he de irme? Como las flores que perecieron? Nada quedara en mi nombre? Nada de mi fama aqui en la tierra? Al menos flores, al menos cantos!