A HISTORIAN CRITICIZES AN ARTICLE BY HARVEY GRAFF IN THE 'JOURNAL OF INTERDISCIPLINARY HISTORY,' 1977, WHERE GRAFF ADVOCATED THE USE OF RECORDS ON INDIVIDUAL OFFENDERS, SPECIFICALLY JAIL REGISTERS, IN RESEARCHING 19TH CENTURY CRIMINAL JUSTICE TOPICS. THE GENERAL AVAILABILITY OF SUCH INDIVIDUAL-LEVEL DATA SOURCES IS QUESTIONED, AND SHOULD THEY BE OBTAINED, RESEARCHERS USING THEM FREQUENTLY IGNORE RIGOROUS QUESTIONS THAT MUST BE ASKED BEFORE GENERALIZED CONCLUSIONS CAN BE DRAWN. IN USING INDIVIDUAL DATA SOURCES, CAREFUL CONSIDERATION MUST BE GIVEN TO THE DATA COLLECTION RATIONALE AND METHODOLOGY EMPLOYED BY THE JURISDICTION THAT COMPILED THE RECORDS SO THAT UNWARRANTED GENERALIZATIONS MAY NOT BE MADE ABOUT TRENDS BEYOND THE JURISDICTION FROM WHICH DATA ARE DRAWN. PREPROCESSED AGGREGATE DATA HAVE BEEN LARGELY NEGLECTED AS AN INFORMATION SOURCE FOR RESEARCHING THE HISTORY OF 19TH CENTURY CRIMINAL JUSTICE. ANNUAL REPORTS BY CITY, COUNTY, AND STATE CRIMINAL JUSTICE AGENCIES WERE PUBLISHED IN THE 19TH CENTURY, AND THE U.S. BUREAU OF THE CENSUS SURVEYED LOCAL AND STATE CRIMINAL JUSTICE INSTITUTIONS EVERY 10 YEARS. IN USING SUCH DATA SOURCES, HOWEVER, THE RESEARCHER MUST ALSO EXAMINE THE METHODS AND REASONS FOR ORIGINALLY COLLECTING THE DATA, SO THEY MAY BE CORRECTLY INTERPRETED. VARIOUS STATISTICAL METHODS SUITABLE FOR ANALYZING AGGREGATE DATA ARE BRIEFLY DISCUSSED. SEE ALSO NCJ-55209. (RCB)