FLOOD09697 - Laserfiche WebLink

Home Browse Search

Table 5.1. Variables in Equation I: public concern. Possible Range of Variable Definition Schedule Item Unstandardized Scoresa CONCLb Perception of Need (in sample) for Improved Flood (Appendix B) 9.44 to 43.Sc Control in a Local Area (CONCL) XI Number of Types of Sources of Information About (9) o to 6 Flooding (NSORSE) X2 Perceived Likelihood of Flooding at Personally Owned (2-A) o to 100 Property in the Area (KPERFL) X3 Length of Years of Awareness of Neighborhood Flooding (7-A) o to 99 Problems (LNElGH) 14 Closeness of Groups of Persons With Whom They Dis-- (9-P) o to 3 cussed Flooding Problems (KLOSGR) Xs Number of Young Children (NYONGC) (119) o to 3 ~ Willingness to Pay for Government Expenditures (Appendix B) o to 24 (P A yU;)d X7 Attitude Toward Effect of Man-Made Objects Upon the (Appendix B) o to 16 Beauty of Nature (MANL-4) Xs leisure Orientation (LEIL4) (Appendix B) o to 16 ~ Proximity of Flooding Experience (KLOSF) (S) o to 3 XlO Cost of Damage from Flooding to the respondent (2) o to 999 (1COSTF) (X9 XIO = 0 to 2997) aFar a definition of the meaning of the scores, refer to Appendix A. bSca1es used to measure the dependent variable, CONCL and the variables fA YL, MANL, and LEIL are listed in Appendix B. The values of the items in these scales may be the fact in the weights given for each question as shown in Appendix A. orhe range of the dependent variables is the minimum to maximum that could be obtained from extreme values of the in- dependent variables. dJames et a1. (1971 :37) investigated a similar variable for willingness to spend personal money, but in a different context. They were measuring willingness to pay for relief from flood damage by taxes or flood insurance. tion using the same variables was also tried with no improvement. The only interaction term in Equation I (5.4) is ~XIO' and Ibis term has the leasl impor. lance of the terms as can be seen from Ihe coefficients of the standardized form. Perhaps one reason for the low r2 is the lack of lerms representing public flood- ing concern. In any case, underlying causal relalion- ships involved will require better defmiIion to greally improve the level of prediction. The coefficienls of the standardized form (Equa- tion 5.5) should only be used to compare the relalive importance of terms presently in the equalion. They should not be considered as refteclive of the weights of these faclors in the "real world." When Ihe propor- tion of explsined variance is low, the sizes of the co- efficients are very unslable and could, a1moslliterally, disappear with the addilion or change of variables in the equation (Gordon, 1968; Schoenberg, 1971). Equation I: Perception of need (or concern) for flood control in a local area (CONCL)8 Unstandardized form: 8 Also called concern for flooding: hence the acronym, CONCL. n - lUNCL ~ (10.S + .S06XIi + .12SX3i n i=l + .3ooX4i - ,453XSi + .148X6i + .l46X7i +.230XSi+.0012(X9iXlOi) = 10.8 + ,S06XI+ .033X2+.128X3 + .399X4 - .4S3XS + .14SX6 + .146X7 + .230XS 9 +.0012 (X9 XIO) (5.4) Standardized form: n CONCL = n ~ (.163Xli + .142X2i + .l80X3i i=l 9The presentation of Equations 5.4 and 5.5 are techni. cally correct. It should be remembered that in calculation of coefficients by regressjonamlysis, the values ofvariab1es are considered individuaUy and the.best match made Qver allcases. The mean va1\le signs matter in the case of product terms since the product of the means is not generally equal to the mean of the products. Howev~r, it is cumbersome to write the sum- mation and mean value signs continually and henceforth they will be omitted for simplicity of presentation. "n" as used here represents the sample size and the used calibration by the regression progr8.JTl. 63