A DEMAND FOR REPEATABLE SUCCESSFUL CRIMINAL JUSTICE PROJECTS IS MADE BY THE AUTHORS AFTER HAVING COMPLETED A LONG-TERM SCIENTIFIC STUDY WHICH USED ECONOMIC VARIABLES. SUGGESTIONS ON HOW TO DESIGN EVALUATION TO ADDRESS EXTERNAL VALIDITY ISSUES AND HOW THESE SUGGESTIONS WERE APPLIED TO THE PREDICTIVE SENTENCING EXPERIMENT ARE ENUMERATED. THE APPLICATION OF THESE PARTICULAR RULES INVOLVED: GATHERING DATA OVER A CONSIDERABLE TIMESPAN (IN THIS CASE BRIDGING HIGH AND LOW ECONOMIC CYCLES) UNDER SEVERAL CONDITIONS, FROM SEVERAL GEOGRAPHIC AREAS; TESTING TREATMENT OVER A VARIETY OF PEOPLE; AND LASTLY, COMPARING RESULTS BASED ON DIFFERENT METHODS OF EVALUATION. THE RESULTS OF THE STUDY IMPLIED A CORRELATION BETWEEN ECONOMIC VARIABLES AND CHANGES IN CRIME STATISTICS. THE AUTHORS FURNISH SUGGESTIONS AS TO HOW SUCCESSFUL CRIMINAL JUSTICE PROGRAMS CAN BE REPEATED, AND REASONS WHY EXPERIMENTERS AND PRACTITIONERS WHO IGNORE THE STATE OF THE ECONOMY MIGHT RISK CERTAIN 'POLICY DISASTERS'. IN CONCLUSION, THE AUTHORS PROPOSE TO TEST THE HYPOTHESIS THAT ECONOMIC CONDITIONS CAN ACCOUNT FOR VARIATION IN SUCCESS/FAILURE BETWEEN AND WITHIN CRIMINAL JUSTICE PROJECTS. IN REFINING THIS TEST, THEY SUGGEST REANALYZING PAST PROJECTS IN TERMS OF ECONOMIC VARIABLES. REFERENCES ARE INCLUDED.