Decision theorists make assumptions about the decision maker or his surroundings. In this section VDM models are examined in terms of five of these assumptions.
1. Assumptions about the amount of information available to decision makers. Specifically, how much does the decision maker know about possible alternatives, how much is known about a given alternative, what is known about the projected outcomes and the probabilities connecting alternatives to outcomes?
The most apparent contrast is between the Hilton and Kaldor Zytowski models. On the basis of his research on the careers of teachers (Hilton, 1960; Hilton, Levin, and Leiderman, 1957) and business administration graduate students (Hilton, Baenninger, and Korn, 1962), Hilton assumed decision makers have limited knowledge about alternatives. Kaldor and Zytowski, on the other hand, assume that the decision maker has relatively unlimited knowledge about alternatives, outcomes, resources, and values. They recognize the difficulties in accepting this assumption but report that hypotheses derived from their model are consistent with the occupational plans of farm boys (Kaldor, Eldridge, Burchinal, and Arthur, 1962).
The other models apparently assume that more moderate but unspecified amounts of information are available to the decision maker. For example, the Gelatt model encourages a thorough information search, but in the absence of "stop" rules, we can only guess at how much information is considered sufficient. As long as only one model (Kaldor Zytowski) suggests that values be assigned to each new "bit" of information received, this assumption is argued at a very general level.
At least two models have been implemented through computer assisted information systems. The Tiedeman O'Hara model was applied through the Information System for Vocational Decisions (ISVD) (Tiedeman et al., 1967, 1968), and the Katz model has been translated to the System of Interactive Guidance Information (SIGI) (Katz, 1969a). Both systems supply considerable information to the decision maker.
2. Assumption of conditions of risk or uncertainty in VDM processes.
The models can be roughly divided into two groups on this issue. The risk group sees vocational decisions as among those where probabilities about future events are assigned. The uncertainty group sees vocational decisions as among those to which no generally accepted probabilities of future events can be attached. Put another way, the difference refers to the amount of so called "objective" data directly applied to the VDM process.
The risk models utilize "objective" probability statements based on other persons' experiences (for example, expectancy tables or regression equations), where available, in the VDM process. This procedure was suggested in the Katz and Kaldor Zytowski models and implied in the Gelatt model. Perhaps the extreme risk advocates are Gelatt and Clarke (1967) who assume that "objective data" is the baseline against which subjective probabilities are compared for "distortions."
The uncertainty models suggest that the probability statements about future events are filtered through the decision maker's subjective judgments before fitting into the array of information used to make a final commitment. Tiedeman's (1967) distinction between data (facts) and information (interpreted facts) seems to capture the essence of the different assumptions. Hilton, Vroom, and Hsu appear to incorporate information rather than facts into their VDM models.
The arguments probably boil down to choosing between beliefs about the differences between the contemporary experiences of the decision maker and the experiences of several past decision makers. Those who prefer the risk assumption find this difference insignificant for decision purposes, whereas those preferring the uncertainty assumption attach considerable importance to these differences.
3. Assumptions about the decision strategy being implemented by the decision maker.
Decision strategies in VDM models can be considered of two general sorts: the classical models that attempt to select the alternatives with maximum subjective expected utilities, and the satisficing models that attempt to minimize the difference between an alternative and some preconceived standard (for example, level of aspiration). The former group includes Vroom, Gelatt, and Kaldor Zytowski. All see the decision maker as comparing several alternative actions and selecting the one that is "best" usually the one with the greatest multiplicative products of values and subject probabilities summed over all outcomes. Kaldor Zytowski add the concept of net value which is the aggregate value less the aggregate costs for a given occupation. From the satisficing group, Hilton and Hsu assume that the decision maker has a standard in mind that must be met. This standard is usually not fixed and fluctuates as a result of the decision maker's experiences. The alternative selected is the one that first meets the standard (Hilton) or the one that comes closest to meeting the standard (Hsu).
Katz uses strategies from both classical and satisficing models. The latter approach is used when the decision maker is asked to estimate the "threshold level" of a value dimension. The former applies to comparisons of expected value sums across options.
Fletcher does not develop a complete strategy but rather suggests that people decide upon the alternative with the greatest "affect change" at the time a commitment is required.
4. Assumptions about the level of precision in combining information to make a commitment.
The pivotal concern here usually has to do with assumed performance by the decision maker in ordering his value preferences. Hilton assumed a simply binary grouping of "yes" or "no" on all outcomes. Others discussed means for ordering value preferences that increase in complexity from Katz and Hsu's constant sum method to Kaldor Zytowski's indifference curves.
Perhaps the models break down into those that include mathematical computations as necessary to descriptions of the process of combining information and those that do not. Vroom, Kaldor Zytowski, Hsu, and Katz use algebraic formulas to define concepts. Gelatt strongly suggests it would be included in a more specific statement of his model.
5. Assumptions regarding the relationship between two conceptualizations about anticipated future states (outcomes).
The relationship between the subjective probability that the state will obtain and the value attached to it constitutes another issue upon which there are basic differences among VDM models. Two aspects of the relationships are involved: (a) how much distinction a model assumes between the two concepts; and (b) the direction of influence, if any, between them. Kaldor Zytowski, Vroom, Hsu, Fletcher, Gelatt, and Katz assume that the two concepts are distinctive, whereas Tiedeman O'Hara and Hilton blend the two together into one nearly inseparable concept. Gelatt and Clarke (1967) gather evidence to show interrelated effects between the concepts, but most of the models avoided the knotty problems of isolating values and subjective probabilities.
Concepts
VDM concepts appear to be quite similar although their labels are different. For example, alternative actions are labeled as "goals," "plans," "actions," "options," "career concepts," and "occupational alternatives." Generally speaking, the definitions describe cognitions of anticipated behaviors relevant to vocational goals. A notable exception occurs where the function of alternatives is applied to roles rather than actions, for example, occupational roles. If we assume that the roles designation applies to role entry behavior, the differences fade into insignificance. If, on the other hand, the roles designate relatively stable future situations rather than behavior, then they serve different functions: roles as more distant ends, and actions as more immediate means. Indeed roles would be more properly considered as outcomes.
Not all VDM models fit comfortably into the decision theory array of concepts. Models proposed by Hilton, Fletcher, and Katz are particularly difficult. None have a concept that clearly fulfills the "outcomes" function. Hilton's "plans," though not explicitly defined, apparently refer to rather complex actions lasting over a long period of time. Katz starts with personal values and works towards actions a reversal of the order used to explain behavior in other models. Consequently Katz' "options" may fulfill the functions of both alternative actions and outcomes. Fletcher's "career concepts" may be comfortably viewed as spanning the decision theory function for both alternative actions and outcomes.
Key differences among models appear to be related to notions about time. Some models combine outcomes (thoughts about somewhat distant events) with actions (thoughts about more immediate events), and others separate the two at some point on the conceptualized timeline. This difference (which may constitute a sixth set of separate assumptions) may account for the variation in conceptualizations of "alternative actions." "Situations" or roles may be more temporally distant descriptions of actions.
There are concepts important to some VDM models that cannot be neatly placed into Tabular format. For example, Hilton's "dissonance test" and "premises," Kaldor Zytowski's "costs," and Katz' "expected values" are not included. In addition, varied emphasis is placed on concepts from model to model. Emphases on values by Katz and plans by Hilton have already been reviewed. Tiedeman O'Hara emphasize the psychological field, Fletcher the career concept, Kaldor Zytowski the occupational utilities, and Gelatt the information about probabilities.
