An experimental analysis of two error correction procedures used to improve the textual behavior of a student with autism

Behavior Analyst Today, The, Summer, 2007 by Michael A. Fabrizio, Sara Pahl

Instructional programs often require that teachers correct student errors. The current project experimentally analyzed the effectiveness and efficiency of two error correction procedures in improving the oral reading of a girl with autism. The project assessed the effectiveness of a Word Supply and a Discrimination error correction procedure at enhancing accurate reading by measuring the number of error corrections required to teach the child to read segments of text from the same book with 100% accuracy. While both error correction procedures proved effective, the Word Supply corrections were more efficient than Discrimination Corrections in terms of both the number of error corrections required per segment of text and the number of instructional sessions required to reach the pre-set accuracy criterion. We discuss these findings and their implications for instructional planning within the context of earlier experimental work related to error correction procedures.

Keywords: error correction, reading, autism, discrimination.

**********

Wherever teaching and learning happen, erring also probably happens. Accordingly, teachers should invest time, energy, and effort in designing, implementing, and evaluating error correction procedures they use with their students. How teachers correct errors their students make should constitute an important part of any instructional system (Heward, 1997) and may play an even more important for students with special needs such as autism, who often present quite a challenge to even the most skilled teachers. (Scott et al., 2000)

One may consider error corrections as one type of feedback that teachers give students. This feedback seeks to increase the probability that students will respond accurately in the future. Hogin (1996) distinguishes between three types of feedback that teachers might give students: (a) feedback given based on correct performance, (b) feedback given based on incorrect performance, and (c) feedback given based on both correct and incorrect performance. Kulvahy (1977) reviewed published literature on feedback in writing instruction and determined that feedback based on incorrect responses (error correction) proved most effective at increasing the accuracy of student performance.

Falvey, et al. (1980) as cited in Wolery, et al. (1988) defined error correction procedures as teacher-delivered feedback that occurs contingent upon a student responding incorrectly to a stimulus. Wolery, et al. (1988) suggested that teachers should group error corrections based on the suspected cause of the error. For example, Wolery et al. (1988) would classify error corrections that occur because a student lacks a prerequisite skill differently than corrections for errors caused by student inattention.

In instruction of students' textual behavior (1), behavioral researchers have evaluated two specific methods of correcting student errors: Word Supply error corrections and Discrimination error corrections. Word Supply error corrections involve providing the student with the full correct answer when the student makes a mistake. Discrimination corrections involve the teacher establishing a conditional discrimination within a student's repertoire by asking the student to respond differentially based on some feature of the textual cue that should have occasioned a correct response but did not. An example of a Discrimination correction would include asking a student to respond to the words "Cat" and "Cats" because the student said "Cat" when presented with the word "Cats." In this example, only a portion of the stimulus (the letters "c", "a", and "t", but not the letter "s") controlled the student's textual behavior.

When evaluating these two methods of correcting student errors with young typically developing learners who had difficulty with their reading skills, Carnine (1980) found that error correction procedures that involved highlighting phonemes were more effective than Word Supply error corrections. In contrast, while studying five developmentally disabled students' ability to read words from lists, Barbetta, et al. (1993) compared Word Supply error corrections to Discrimination corrections and found that students performed better when given a complete model than when using a phonemic prompting strategy similar to a Discrimination error correction.

The current project investigated which of the two error correction procedures described above (Word Supply or Discrimination error corrections) was most effective at teaching a student with autism to read stories correctly. The work undertaken by Barbetta, et al. (1993) is similar to the work done in the current project in that the Word Supply error correction procedures were almost identical to the Word Supply error correction procedures used here. Heward's (1997) phonetic prompting Discrimination correction was similar to the Discrimination correction routine used here.

We also hope that the dissemination of the methods we used to analyze the two error correction procedures and the results garnered from that analysis might encourage other scientist-practitioners to conduct further investigation into the effectiveness of error correction procedures they employ within instructional programs they either design or supervise.