A REACTION TO JOHN DOWNING'S REPORT ON THE I/T/A READING EXPERIMENT IS PRESENTED. DOWNING ASSUMED THAT THERE WAS NO SIGNIFICANT DIFFERENCE BETWEEN THE EXPERIMENTAL AND CONTROL GROUPS ON IMPORTANT VARIABLES, BASED ON THE RESULTS OF THE KOLMOGOROV-SMIRNOV TEST. IT IS CLAIMED, HOWEVER, THAT THE TEST USED IS APPROPRIATE FOR SMALLER SAMPLES. A MORE RELIABLE METHOD FOR LARGER SAMPLES IS THE TRADITIONAL STANDARD ERROR TECHNIQUE WHICH WOULD INDICATE SIGNIFICANT DIFFERENCES BETWEEN THE TWO GROUPS. THE KOLMOGOROV-SMIRNOV TEST HAS 95 PERCENT OF THE POWER EFFICIENCY OF THE STANDARD ERROR TECHNIQUE, AND THIS LEVEL DECLINES AS THE SAMPLE INCREASES. TO GET THE MAXIMUM POWER EFFICIENCY, THE KOLMOGOROV-SMIRNOV TEST REQUIRES THAT DATA BE GROUPED OVER AS MANY INTERVALS AS POSSIBLE. IT IS POINTED OUT THAT DOWNING FAILED TO DO THIS. THUS, INAPPROPRIATE USE OF THE TEST FAILED TO SHOW ACTUAL DIFFERENCES IN AGE AND INTELLIGENCE BETWEEN THE TWO GROUPS. IF THE APPROPRIATE METHOD WERE USED, THE DATA WOULD NOT SUPPORT CLAIMS REGARDING THE SUPERIORITY AND USEFULNESS OF I/T/A. DATA WOULD INDICATE THAT THE EXPERIMENTAL GROUP IS MORE READY TO LEARN TO READ REGARDLESS OF THE METHOD USED. THIS ARTICLE APPEARED IN "READING RESEARCH QUARTERLY," VOLUME 2, NUMBER 1, FALL 1966. (NS)