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. - I00 - Section 2 Recall Devices The tests made on different recall devices on the single term index languages are based on the 42 questions, {subset 2), searched on the 200 document collection. Tables 4.200T to 4.206T give the performance results, and the recall]precision plots given in Figs. 4.200P to 4.206P show the curve for the particular language being tested, together with the curve, (shown by a broken line) obtained by the device-less search in natural language, which was presented in the previous section as Fig. 4.140P. A combined plot of all the seven devices or their aggregates, together with natural language, is given in Fig. 4.207P. In presenting these results, all other possible variables of search rules, indexing exhaustivity and document relevance are held constant, and are fixed to the one particular variation indicated at the head of the tables. LIST OF FIGURES 4. 200TP 4.201TP 4. 202TP 4. 203TP 4. 204TP 4. 205TP 4. 206TP 4.207P Index No. of Question Document Language Questions Subset Collection Plot s 1.2.a 42 2 200 Plot + 4.140T 1.3.a 42 2 200 " " " 1.5.a 42 2 200 " " " 1.6.a 42 2 200 " " " 1.7.a 42 2 200 " " " 1.8.a 42 2 200 " " " 1.9.a 42 2 200 Plot + 4.140T I.l.a 42 2 200 Plot 4.140T, i.2.a 4. 200T, 1.3.a 4.201T, 7.5,a 4. 202T, 1.6.a 4.203T, I. 7.a 4. 204T, I. 8.a 4. 205T, I. 9.a 4. 206T.