AN EXPERIMENT WAS CONDUCTED, OVER A 2-YEAR PERIOD, AT THE OHIO STATE UNIVERSITY SCHOOL TO TEST THE POSSIBLE EFFECTS OF A STUDY OF A GENERATIVE GRAMMAR ON STUDENT COMPOSITIONS. IN ADDITION TO THE REGULAR CURRICULUM, AN EXPERIMENTAL CLASS OF 50 NINTH-GRADERS (AND 10TH-GRADERS THE FOLLOWING YEAR) WAS REQUIRED TO LEARN TRANSFORMATIONAL-GENERATIVE GRAMMATICAL MATERIALS PREPARED BY THE INVESTIGATORS. SAMPLES OF WRITING BY BOTH THE EXPERIMENTAL AND CONTROL CLASSES WERE COLLECTED DURING THE FIRST 3 MONTHS OF THE PROJECT, AND IN THE LAST 3 MONTHS OF THE 2-YEAR PERIOD. THESE WERE ANALYZED FOR (1) STRUCTURAL COMPLEXITY, (2) THE PROPORTION OF WELL-FORMED TO MALFORMED SENTENCES, AND (3) THE TREND IN THE FREQUENCY AND KINDS OF MISOPERATIONS THAT OCCURRED. RESULTS OF THIS ANALYSIS INDICATE THAT KNOWLEDGE OF GENERATIVE GRAMMAR (1) ENABLES STUDENTS TO INCREASE SIGNIFICANTLY THE PROPORTION OF WELL-FORMED SENTENCES IN THEIR WRITING, (2) SEEMS TO ENABLE STUDENTS TO INCREASE THE COMPLEXITY WITHOUT SACRIFICING THE GRAMMATICALITY OF THEIR SENTENCES, AND (3) CAN ENABLE STUDENTS TO REDUCE THE NUMBER OF ERRORS IN THEIR WRITING. BECAUSE THE LOGIC OF GENERATIVE GRAMMAR SEEMS TO REPRESENT THE PSYCHOLOGICAL PROCESS OF SENTENCE FORMATION, ITS STUDY HELPS THE STUDENT OVERCOME THE DIFFICULTIES OF WRITING WITH GRAMMATICAL EFFECTIVENESS. (THIS REPORT--A CONDENSATION OF THE FINAL REPORT FOR COOPERATIVE RESEARCH PROJECT NO. 1746 (SEE ED 001 241.)--IS ALSO AVAILABLE FOR $1.00 FROM THE NATIONAL COUNCIL OF TEACHERS OF ENGLISH, 508 SOUTH SIXTH STREET, CHAMPAIGN, ILLINOIS 61820, ORDER NO. 14709.) (DL)