FLOOD09697 - Laserfiche WebLink

Home Browse Search

cases to establish a regression equation. If regression techniques are to be applied to an agency, some treat- ment of multiple estimates as separate cases appears to be required in the absence of data on the same agen- cy for many proposals. The preceding was explained 10 show how suffi- cient cases for regression analysis were obtained. By combining data from two agencies and three proposals for eighl judges, 48 cases were obtained. Separating Ihe two agencies would have resulted in 24 degrees of freedom, an insufficient number. Had the judges' data all been pooled for each proposal and agency, there would have only been a total of six data sets. The besl way to avoid having to do this is to gather data on many proposals fo r an agency in an area. Because of the method by which the data were obtained, it cannol be determined what proportion of comparatively high correlation coefficients of the agency equations results from real association and what proportion is from a tendency toward cognitive con- sistency or internal mental agreement by each judge. Adjustment toward consistency is not the only adap- tive mechanism for matching experience and percep- tions, but it may be common. To the degree that judges achieve consistency at the sacrifice of accurate representation, a misleading result could occur. Hope- fully, judges can provide reasonably accurate data when other sources are not available. The problem discussed here applies only when one person must judge another's values. What matters in application is the ability to correctly predict the evaluation of flood controi proposals by a particular group. Evaluation by the planning agency (PAEV) Equation 1II (5.9) is needed for predicting the evaluation by the planning agency of a flood control proposal idea. If a plan Is evaluated favorably and approved by the planning agency, it is considered as proposed to the decision agency. Since this is an initial evaluation, no Type N in- teraction terms are used to reflect the influence of a previous judgment on the plan, as the last two terms, #8 and #9 in Table 5.3, are Types IV terms. Term #7 reflects the ecoloeical consequence of the proposal if enacted and was not included in Equalion 1II be- cause its significance was too low. With the data used, ecology did not have an appreciable effect other than Ihrough its relationship to aesthetics. Term # 7 did not enter an agency evaluation lhis early in the plan- ning process. The values of the agency variables which would be absorbed into the coefficients of Equation III (5.9) can be substituted for those variables in the non-stan- dardized form of the equation. These variables are X2 through X5 for Equation II (5.9).15 The values of these variables are in Table 5.2. Substituting the val- ues listed there are the planning agency, in the un. standardized form of Equation III (5.9) results in the following: 16 Y = -.241-.305XI-.0695(1.75)X9 + 8.43(1.75)/XlO + .00568(3.75)Xll + .136(.625)X12 + .228(.625)X13 = -.241- .305Xl - .122~ + 14.753/XlO + .021Xll + .085X12 +.143X13 . (5.11) Y in this case is specifically the evaluation by the plan- ning agency of a flood control plan. Initial evaluation by decision agency (DAlE) Equation III is also applied to predict Ihe inilial evaluation by the decision agency under the model as- sumptions that no important evaluative input from another group enters in at this point and that the plan- ning agency is considered as recommending the plans it proposes. The second assumption is justified by the logic that the planning agency would not propose a plan that it did not favor and that the atlitude of the agency would be relatively stable in a favorable direc. tion and thus tend to have a consistent influencing effect. Also, a planning agency and a decision agency oflen work together closely enough to develop con- gruent perspeclives toward types of flood control proposals. If these assumplions are true, separate in- clusion of the planning agency evaluation would tend to dominate the effect of other variables in a regression analysis that accounts for much of the same variation as already considered. Omission of a decision agency acceptance function for the planning agency evalua- tion would consequently allow the effects of the other varaibles to be more closely seen through the equation. Equation III further distinguishes the inilial de- cision agency evaluation from the revised evaluation (Figure 5.1). Even if Equation IV were used both places, the results would stili be different as the out- put from the planning agency is different than that from the decision agency, but the similarity of the agency evaluation values in the model would be great- er. If the plarming agency evaluation has a significanl differenlial effect upon the initial decision agency evaluation from proposal to proposal, then Equation IV could be used in place of Equation III. When the 15X6 and Xl are also used in Equation IV (5.16), Xg is added for Equation IV (5.21). 16See also discussion of Equation III. 68