130 Understanding Rational Decision Making
weight rule. The difference is that they do not weight the relative importance of each attribute.
Instead, they produce an overall value for each alternative by adding the unweighted values for each
attribute of that alternative.
Unlike compensatory choice rules, noncompensatory rules do not allow trade-offs among
attributes. Examples of noncompensatory rules include the elimination-by-aspects rule and the lexi-
cographic rule , two rules that rely on attribute-based search and two rules that are suggested by our
model of audience decision making. When audiences use the elimination-by-aspects rule, they
fi rst rank all the attributes or decision criteria according to their importance.^243 Then they com-
pare each alternative’s value for the top-ranked attribute to some minimally satisfactory value and
eliminate any alternative from further consideration whose value is less than the satisfactory value.
For example, if safety were the most important feature to a new car buyer, then the buyer would
immediately eliminate any car with a poor safety record. Next the audience eliminates alterna-
tives that do not have satisfactory values for the second-ranked attribute. The audience continues
this process of elimination until only one alternative remains. Many fi nancial analysts report that
they rely on the elimination-by-aspects rule when making multi-attribute investment decisions.^244
When audiences use the lexicographic rule, they also rank each attribute or decision criterion
in terms of its importance. They then compare the values of the alternatives on the highest ranked
attribute and choose the alternative with the highest value on that attribute, regardless of the
values the alternatives have on other attributes. If, for example, miles-per-gallon were the highest
ranked attribute or decision criterion for a new car buyer, and if the new Toyota Prius had the
best gas mileage of all the cars under consideration, then the buyer would chose the Prius without
examining other attributes. Such noncompensatory choice rules can minimize cognitive effort
without severely decreasing the accuracy of the audience’s decision. For example, in one experi-
ment a simple noncompensatory rule increased the error rate from 8% to only 14% over that of
the compensatory expected-value rule.^245 But the time saving was dramatic. Audience members
took about two minutes to apply the expected-value rule but only 15 seconds to use the non-
compensatory rule.
Constraints on the Use of Compensatory Choice Rules
Can professionals predict which choice rule their audience will use? As we have seen, consumers
show a strong preference to use choice rules that rely on attribute-based search and compari-
sons.^246 Choice rules that rely on attribute-based search include noncompensatory rules such as the
elimination-by-aspects rule and the lexicographic rule, as well as compensatory rules such as the
additive-difference rule and the majority-of-confi rming-dimensions rule.
When audiences choose between just two alternatives, they typically use an attribute-based com-
pensatory choice rule.^247 However, as the number of alternatives they have to consider increases,
their use of compensatory choice rules decreases. For example, physicians asked to choose among
three anti-infective drugs used a compensatory choice rule. But they used a noncompensatory
choice rule when asked to choose among six anti-infective drugs.^248
If from the outset the audience has a large number of alternatives to choose among, they
will typically use a noncompensatory choice rule such as the elimination-by-aspects rule.^249
Venture capitalists use the elimination-by-aspects rule to screen the many business plans that
come across their desks.^250 Investors use the elimination-by-aspects rule to pick stocks.^251
Industrial buyers use similar noncompensatory rules to choose among large numbers of poten-