Effects of time-compressed speech signals on children's identification accuracy and latency measures

Abstract

Identification accuracy and latency of responses were studied under various rates of time-compressed speech signals. Ss were 51 elementary school children, 10 kindergarteners, 16 first-graders, 14 second-graders, and 11 third-graders. Each S was instructed to respond to one of a set of three line drawings selected from the Picture Identification Children's Standardized Index when hearing a speech signal. The speech signals, 50 in number for each treatment plus 24 appropriate adaptation trials, were presented under normal speech and the three experimental conditions (30, 50, and 70% time-compression). Each S received only one treatment, exactly 14 days after the normal run. Mean identification accuracy, based on corrected percentage scores, significantly decreased as a function of time-compression. Graphically, very little decline appeared before the 50% condition. Mean latencies for incorrect as well as for incorrect and correct responses combined were significantly lower for the 50% treatment than for the 30 and 70% treatments, and the means for the 30% condition were lower than those for the 70% condition. The incorrect mean latency scores also were significantly different from the correct mean latency scores.