The study tests the issue stream model of organizational decision making to examine how organizational decisions follow one another or influence each other. The interaction between decisions become important aspect of organizational behavior. Decision linkages can be purely sequential and concerned with the same basic issue at different times. Sequential decisions can be nesting, snowballing or recurring. Nesting occurs when a major decision leads to a number of smaller sub-decisions; snowballing is when a number of small decisions lead to a major decision; and recurring decisions repeat themselves within the issue stream. Linkages can also occur laterally between different issues being considered at the same time. Lateral decisions share resources (in which case they are called pooled), or share people, culture, ideology or strategies (in which case they are called contextual). Finally, issues can be pre-cursive to each other. Precursive decisions can be enabling (one decision removes blocks to other decisions), evoking (creating new problems or opportunities), preempting (rendering other decisions irrelevant or obsolete), cascading (one decision may set off a series of wide ranging issues), merging (a number of unrelated issues came to be seen as one, or learning (early decisions generate learning that may influence later decisions). This study will examine issues that emerged during a re-engineering meeting at an AHC. One data source will be the minutes of 204 meetings of the committee in charge of re-engineering. In addition, data will be collected from internal and external media. A third source of data will be the actual documents produced by the committee. The last source of data will be interviews with members of the committee. Two (or more) reviewers will classify these data into strings of words that capture the basic elements of information expressed about a discrete incidence at a particular time and place. The linkages between these issues will be classified using the issue stream model. The dependent variable examined will be the complexity of issues faced at any time period. Issue complexity is defined as the sum of sequential decision streams plus the total number of linkages present at that day. The independent variables are a number of predictors of issue complexity including discussions of the following: 1) federal health policy, 2) state health policy, 3) health industry events, 4) university events, 5) number of decision makers, 6) presence of organizational sub-unit affected by the decision at the meeting, 7) time pressure as defined by the number of days before deadline, and 8) issue complexity in the previous time period. The relationship between issue complexity at time (t) and the above list of explanatory variables at time (t) and at time (t minus one) will be analyzed. An auto-regressive moving average model will be used to estimate the relationship between the dependent and independent variables. Because ARMAX (autoregressive moving average with explanatory variables) models do not fit the assumption of the lagged independent variable, a maximum likelihood estimation procedure will be used in conjunction with the ARMAX. The model that best describes the data will be selected. A number of relationships between the independent variables and the dependent variables are proposed, some of which are key to the issue stream model of decision making. In particular, it is expected that as the number of decision makers increase, as the time pressure increases, and as external events create new context for the decision making, decision complexity will increase. The project will also examine whether re-engineering (a linear step by step process) creates a series of new issues that are non-linear, in the sense that they are not linked sequentially and that they come to a resolution through seemingly random processes. Recognition of how a structured process can lead to new unstructured problems, helps organizations anticipate and plan for these eventualities. Mathematical chaos theory will be used to test whether the progression of issues over time has a random component, and is sensitive to initial starting point.
