Journal on Policy & Complex Systems Volume 5, Number 1, Spring 2019 | Page 12

Meta-Decision Modeling of Wicked Environmental Policy Design Problems
scriptive , and meta-decision-making ( Cleveland , 1973 ; Corner , Buchanan , & Henig , 2001 ; Gal , Stewart , & Hanne , 1999 ; Hwang & Yoon , 1981 ; Kahneman & Tversky , 1979 ; Raiffa , 1968 ; Weiss & Bucuvalas , 1980 ; Winterfeldt & Edwards , 1986 ; Zeleny , 1982 ). These classifications are based on Hume ’ s thesis that descriptive facts are separate from normative values ( Hume , 1955 ). Most decision theorists share the ontological commitment that they can potentially discover an optimizing decision algorithm that can either describe how people make decisions ( Armstrong , 2001 ; Becker , 1976 ; Becker & Sarin , 1987 ; Chew , 1983 ; Dekel , 1986 ; Herriges & Kling , 1999 ; Kahneman & Tversky , 1979 ; Porter , Roper , Mason , Rossini , & Banks , 1991 ; Quiggin , 1982 ; Ringland , 1998 ; Wong , 1978 ), or prescribe how people should make decisions ( Beres & Targ , 1977 ; Farmer , 1987 ; Fishburn , 1976 ; Gregory & Keeney , 1994 ; Hwang & Yoon , 1981 ; Luce , 1956 ; Mulder & Biesiot , 1998 ; Saaty , 1980 ; Tversky , 1972 ; Yoon , 1989 ; Yoon & Hwang , 1985 ; Zeleny , 1982 ). Due to this ontological commitment , decision scientists have created many descriptive and normative decision algorithms , as shown in Table 1 . Table 1 also shows specific situations in these descriptive and normative algorithms that require meta-deci-
Table 1 . cont ’ d .
τ N
Normative Decision Algorithm
1 . Backcasting ( Mulder & Biesiot , 1998 )
2 . Normative scenario analysis ( Beres & Targ , 1977 )
3 . Dominance ( Gregory & Keeney , 1994 ; Yoon & Hwang , 1985 )
4 . Elimination by aspect ( Tversky , 1972 )
5 . Lexicographic ( Luce , 1956 )
6 . Simple Additive Weighting ( SAW ) ( Farmer , 1987 ; Fishburn , 1976 )
7 . Weighted product ( Yoon , 1989 )
8 . Analytical hierarchy process ( Saaty , 1980 )
9 . TOPSIS ( Hwang & Yoon , 1981 ; Zeleny , 1982 )
Specific Situations Requiring Meta-Decision Choices
Choice of values desirable at the end of decision horizon
Choice of preferable scenarios over nonpreferable scenarios ; incomplete information ; uncertainty
No solution with multiple nondominant alternatives
Exogenous rank ordering from most important to the least important values required
Exogenous rank ordering from most important to the least important values required
Decision maker exogenously assigns the weights , which ought to be additive for multiple values
Decision maker exogenously assigns the weights , which ought to be multiplicative for multiple values
One over-arching objective should be selected prior to determining weights through binary comparisons
Positive and negative ideal solutions shall be exogenously determined
Note : It can be formally shown that benefit – cost analysis ( BCA ) is a special case of SAW . 9