Persuasive Communication - How Audiences Decide. 2nd Edition

28 Understanding Rational Decision Making

To apply the rules of normative decision making, an expert must fi rst create a model, that is,

identify the appropriate attributes or decision criteria, the correct weight for each, as well as

the appropriate alternatives or benchmarks. Once the model has been generated, they must

enter the correct attribute values, or slot values, for each alternative, multiply each attribute

value by the weight for that attribute, calculate the total attribute value for each alternative,

and select the alternative with the highest overall value. A meta-analysis fi nds that experts’ lack

of consistency in applying normative rules accounts for much of the advantage linear models

have over them.^246

Most experts are effective at collecting information about each of their alternatives but are less

effective at weighting that information and integrating it into a fi nal decision.^247 However, some

experts, such as some experienced investors, neither collect all the available information about their

alternatives nor do they fully compare the available choices. For example, experienced investors

are much more likely to tally pros and cons for the handful of investments they decide to consider

than to compute weighted sums from the complete set of options they have to choose among.^248

Moreover, most experts show only moderate levels of self-insight about the cognitive processes they

use to arrive at their decisions.^249

Other anomalies and biases crop up in the decision-making processes of expert audiences

as well. Expert fi nancial analysts often neglect base-rate benchmark information and focus

exclusively on case-specific information when predicting stock prices.^250 Experts may

make different decisions based on the way their choice was elicited.^251 For example, experts

may make different decisions when asked to select instead of reject one of two alternatives.^252

They may also make different decisions when asked to evaluate one alternative at a time as

opposed to all alternatives concurrently.^253

In addition, many experts do not exclusively use compensatory choice rules—rules for decid-

ing that involve trading off a low value on one criterion (e.g., poor gas mileage) for a high value

on another (e.g., excellent safety features)—when making decisions as the normative rules of

decision making dictate. Instead, experts often use noncompensatory choice rules and simply

eliminate from consideration any alternative that has a low value on an important decision cri-

terion prior to computing the remaining alternatives’ overall values.^254 Of course, some experts

do routinely use compensatory choice rules when making important decisions. For example,

corporate recruiters regularly use compensatory choice rules when making hiring decisions.^255

Consumer Reports also uses a compensatory choice rule to compute an overall score for each

product it evaluates.

Even when decision makers fully understand compensatory choice rules, they do not necessarily

use them to make decisions. One study found that MBAs taught to use compensatory choice rules

did not use them to make their real job decisions. However, MBAs did use compensatory choice

rules to justify their already-made decisions.^256 Similarly, undergraduates taught to use compensa-

tory choice rules did not increase their use of them. Nonetheless, they did increase their desire for

their agents, for example, their doctors and advisors, to use compensatory choice rules when mak-

ing decisions on their behalf.^257

Another cause of experts’ normally inferior performance is that some types of decisions are

inherently diffi cult for human beings to make. As indicated in Table 1.2 , expert audiences perform

relatively well in domains that involve decisions about objects or things.^258 These domains have a

high degree of predictability and provide feedback readily. Expert audiences perform less well in

domains that involve decisions about human behavior which is highly unpredictable and for which

feedback is less readily available.^259 In such domains, experts tend to disagree with each other and

have a low degree of consensus.^260