Effects of several feedback methods for correcting reading errors by computer-assisted instruction

As modern technology facilitates the presentation of various forms of feedback in instructional systems, it is important to investigate their relative effects. An experiment was performed to investigate the learning effects of three forms of feedback. Sixty novice readers participated in the experiment. Their task was to select a word from three others on a computer screen after the word was spoken by the computer. Under the first condition ('control condition') the subject did not receive any information at all as to the correctness of the answers while, under the second condition ('partial-feedback condition') information as to the correctness of the answers was given. The third condition ('complete-feedback condition') was similar to the second condition but the correct answer was also given when a subject made an error. Each child participated in the experiment during four learning sessions. On a pre and post-test, reading times and errors were registered, while subjects had to.read the words aloud. Learning was measured as a function of the results before, during and after the learning sessions. The results showed that children learned more under the two feedback conditions than under the control condition. Although pre and post-test results showed no differences in learning effects between the two feedback conditions, fewer errors were made under the 'complete-feedback condition' during the learning sessions.