Constructive Proposals
The initial section of Table 8 lists three constructive and sufficiently defined proposals, ready for application and testing. ‗t Hart‘s (1998) article characterized various types of decisions and made a case that groupthink should only be applied to ―problem solving‖ decisions and not other types of decisions (such as those driven by political factors). ‘t Hart, as well as Mohamed and Weibe (1996), advocated for adding accountability to the list of prevention steps. Rosander, Stiwne and Granstrom (1998) developed a tool for assessing groupthink tendencies.
Mohamed and Weibe (1996) advocated that groupthink is a process model. They make the argument that many of the experimental tests have failed because the researchers are assuming a causal order variance model. Other articles also support this process approach; for example, Courtright (1978) stated that Janis specifies ―a probabilistic relationship‖ versus the causal order assumed by many (see Tables 4 & 5).
It appears the assumption that groupthink is a causal ordering variance model resulted from a Janis (1982) figure that implies a causal order. However, Janis stated, ―even when some symptoms are absent, others may be so pronounced that we can expect all the unfortunate consequences‖ (p. 175). This statement supports a process versus variance approach.