THE PERSONAL EXPERIENCES WHICH ILLUSTRATE THIS DISCUSSION ARE DRAWN FROM AN EXTENSIVE LONG-TERM FOLLOWUP STUDY CONDUCTED IN CONJUNCTION WITH THE PARK HOSPITAL FOR CHILDREN, OXFORD, ENGLAND. THE MAJOR PROBLEMS ENCOUNTERED BY SUCH STUDIES ARE THE IDENTIFICATION OF ABUSED CHILDREN AND THE IDENTIFICATION OF SUITABLE CONTROLS. SOME STUDIES ARE SO RESTRICTIVE THAT A BROAD UNDERSTANDING OF THE PROBLEM IS NOT ACHIEVED, WHILE OTHERS USE A DEFINITION OF ABUSE THAT IS SO INCLUSIVE THAT PROBLEMS EXCLUSIVE TO ABUSE CANNOT BE IDENTIFIED. THE PERSONAL INTERVIEW IS THE MOST EFFECTIVE DATA COLLECTION TECHNIQUE BECAUSE FAMILIES TEND TO THROW QUESTIONNAIRES IN THE WASTEBASKET. THIS STUDY SOUGHT THE HELP OF THE FAMILIES THEMSELVES IN DESIGNING THE QUESTIONNAIRE TO AVOID MIDDLE-CLASS BIASES. THE PROBLEM OF MISSING CHILDREN IS DISCUSSED AT LENGTH. THERE IS NO GUARANTEE THAT 'LOST' CHILDREN HAVE THE SAME CHARACTERISTICS AS THOSE AVAILABLE FOR FOLLOWUP. THEY MAY BELONG TO FAMILIES SO DEBT RIDDEN OR IN TROUBLE THAT THEY MOVE FREQUENTLY AND USE ALIASES. OR THEY MAY BE UPPER-CLASS FAMILES WHO MOVE TO AVOID THE SHAME OF TREATMENT. FOLLOWUP IS MADE DIFFICULT ALSO BY SOCIAL AGENCIES WHICH DROP A CASE WHENEVER THEY GET A NEW ONE (TO KEEP THE CASELOAD CONSTANT) AND WHICH LOSE CHILDREN. IT IS POINTED OUT THAT SOME OF THE 'LOST' CHILDREN MAY BE DEAD. A CHECK OF THE VITAL STATISTICS REGISTRY IS URGED. ADOPTION, INSTITUTIONALIZATION, AND DISTANCE FROM THE RESEARCH CENTER ALL HINDER FOLLOW-UP. AVAILABLE FOLLOWUP FIGURES FROM THE PARK HOSPITAL STUDY ARE INCLUDED. DATA COLLECTION TECHNIQUES ARE REVIEWED. REFERENCES ARE APPENDED. (GLR)