Putting the Pieces Together—
Structural Equation Models
In this chapter we attempt to show how a number of the relationships examined separately in previous chapters can fit together. “Putting the pieces together” is a bit of an overstatement, however; a more appropriate title for this chapter would be “Putting some of the pieces together. ” We begin this chapter by spelling out which of the pieces we put together, why we chose these particular ones, and what rationale we offer for the particular causal ordering we selected for our structural equation models.
In this book and the previous one, we examined a considerable range of post-high'School experiences, and we showed some of the ways in which they overlap. We found, for example, that increases in occasional heavy drinking among full-time college students could be explained primarily by their marital status and living arrangements. We were at pains, in our previous book, to make clear our interpretation that student status does indeed matter, albeit indirectly: “Our interpretation of these findings is that student status does indeed contribute to increased alcohol use, but only indirectly via the marital status and other living arrangements that generally go along with being a college student” (Bachman et aL, 1997, p. 169).
It is one thing to look simultaneously at a number of potential causal factors that may be overlapping in their impacts, document some of those overlaps, and offer interpretations of underlying causal sequences. That is what we did in chapters 4–7. It is quite another matter to incorporate into analy-