CRANV2 Main test results chapter Cyril Cleverdon Michael Keen Cranfield An investigation supported by a grant to Aslib by the National Science Foundation. Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. 82 - In the first section the results are given for the single term index languages with the largest set of documents and questions. Because of the doubts regardin[OCRerr] the totalling method used, the 35 seven-term questions set is given for comparison. This is followed by the 42 aerodynamic questions on the 1400 document collection, with a final table for this set of questions with the 200 document collection. The purpose of this group of tables is first to justify the totalling procedure, then the reduced set of questions and finally the reduced set of documents. Section 2 gives the results for the 42-question and 200- document sets for the eight single term languages. The tests using precision devices of interfixing and partitioning are given in Section 3, and are followed in Section 4 by comparative resuits at different levels of exhaustivity. The effect of search rules is shown in Section 5 and the different levels of relevance are tested in Section 6. Results for fifteen simple concepts index languages are presented in Section 7, while Section 8, which deals with the six controlled term index languages, includes results for recall devices, precision devices and search rules. In Section 9 are presented the results of searches on the titles and abstracts, The tables and plots are numbered according to the above nine sections, [OCRerr]Tw indicating a table and wp[OCRerr] indicating a plot. In all cases unless otherwise stated, the plots are of recall and precision ratios.