TEST-RETEST RELIABILITY OF CONTINGENT VALUATION WITH INDEPENDENT SAMPLE PRETEST AND POSTTEST CONTROL-GROUPS

Test-retest, the standard method used to investigate reliability of co ntingent valuation, is limited because when the time period between th e two surveys is relatively short the study may exhibit testing recall . Conversely, when the time between the two surveys increases there is an increased chance the true value will change. The test-retest desig n cannot isolate these effects. Independent pretest and posttest contr ol groups allow testing of reliability even when recall occurs or the true value changes. Using this design, we found ex post estimates of H icksian surplus are reliable regardless of whether respondents have di rect experience with an activity.