Reference Poster Session at the CEC Conference, April 20, 2001.

Evaluating a Computer-Based Instructional Environment for Autism

Susan Osborne, Chan Evans, and Dorothy Strickland

Children as young as eighteen months old may be diagnosed with Pervasive Developmental Disorder (PDD). This umbrella term is used for any of five syndromes, including autism, in which a child exhibits varying degrees of deficits in the areas of verbal and nonverbal communications, social interaction, and behavior.

Purpose

The objectives of this research were to (1) develop a computer-based instructional system to teach language acquisition to young children with PDD and other developmental delays; (2) determine whether children as young as three-years-old could be taught to use the mouse; (3) determine whether young children could learn new words taught with an interactive computer program; and (4) determine the effectiveness of using an animated Graphics Interchange Format (GIF) to teach generalization from the computer screen to the real world.

Participants and Setting

Students from three elementary school special education classrooms were selected to participate in this project. Each school is located within a countywide district of 94,000 students that includes eleven towns and one midsize metropolitan city. Special education teachers nominated a total of nineteen students as potential candidates for this program. Parental consent was obtained for all the children before each was given a computer-based pretest. Ten children were excluded when it was determined that they would not benefit from this type or level of instruction. Nine children between the ages of three and seven from four public special education classrooms were chosen to participate and were subdivided into two groups according to skill development. Five of the children were diagnosed with Pervasive Developmental Disorder (PDD); two with Developmental Delays (DD); and two with Preschool Delays Atypical (PDA). Four of them had developmentally appropriate verbal skills and five used very little verbal communication (one- or two-word responses with some echolalic utterances).

Equipment and Software

For general availability to the largest audience, the computer implementation was done for a low end PC using web standard JAVA Script. The lessons work on all versions of Netscape and Internet Explorer 4.0 and later and do not require any plug-in software, graphics hardware or special set-up by the users before playing. The exception is that the user may choose to reset his screen to display 600 by 800 pixels. Lessons also play on MAC systems, although at a slightly increased rate. All programs were coded with time lines and customized functions to incorporate specialized screen actions and responses. In-house software was designed to create play arrangements to teach effectively while avoiding a known pattern recognition by the students. Live images of people using the items or doing the actions were processed as minimum sized animated GIFs, with a standard sequence of three flowing images, to simulate live action within the computer-generated images.

Research Design

We employed a multiple probe design across objects to be learned. This variation of the multiple baseline design permitted baseline phase for each participant and avoided requiring the children to repeatedly attempt a task for which they had not yet received instruction.

I. Using the Mouse
Before beginning the object identification lessons, each child was introduced to the mouse and the icons used in the instructional program. The researcher explained and showed the child how to point and click with the mouse. Hand over hand instruction or "point to screen" directions were given if a child continued to have difficulty with this task. This phase consisted of three stages:
(I) Introduction to "Sunny", the home icon; (II) Differentiating between two objects, Sunny & Square, and pointing and clicking alternately on each object; and (III) Introduction to the "Help" icon. The child was shown how to point and click with the mouse with each stage. Then the computer program demonstrated the skill as the computer voice gave instructions (e.g., "Show Sunny"). Next, the child was given a chance to practice using the mouse. He was encouraged to "Try again" by the computer voice if he moved incorrectly. After two mistakes, the computer "Help" feature directed him to the correct icon.

II. Object Identification
In order to generate a list of known and unknown common home and school items, a baseline was established by asking each child to identify photographs of objects such as "book", "doll", "cookie", "potholder", or "headphones". One word from the unknown list was chosen as the target. In the demonstration, the program showed a photograph of the item and the computer voice named it. This was followed by a GIF illustrating the function of the object. A sequence of three images gave the impression of movement as a young person demonstrated its use. For example, for "cookie", the GIF showed a girl taking a bite of a large chocolate chip cookie.

Generalization

A real world test was given by displaying six objects, three knowns and three unknowns on a black poster board. The child was asked to "Show me….", alternating between knowns and the target object. Mastery in the real world was established at 66% accuracy. If this criterion was not met, we scheduled additional computer training sessions until the child again met the teaching phase criterion.

Results

All nine children learned new words using the interactive computer software and they all transferred this knowledge from the computer to the real world. Also, the use of GIFs seemed to help establish the function of the target words. During the real world probe several children picked up the objects and verbalized or demonstrated their use (e.g., grater: "It's for cheese"; blender: "Milkshakes!"; spatula: "Flipping pancakes"). Some of the children learned to use the mouse for object selection, while others pointed to the screen to show recognition of target words.

III. Daily Living Skills: Set the Table

Purpose

The objectives of this research were to (1) develop a computer-based instructional system to teach a daily living skill to children with PDD or autism; (2) determine whether children could learn how to set the table when taught with an interactive computer program; and (3) determine the effectiveness of using computer animation (Graphic Interchange Format - GIF) to teach generalization from the computer screen to the real world.

Participants and Setting

Five students from two public school cross-categorical special education classrooms were nominated by their teachers to participate in the project. They were between the ages of seven and eleven years old and diagnosed with either PDD or autism. All the students, four boys and one girl, were European Americans. Their receptive and expressive language skills were sufficient for comprehension and expression necessary for the tasks presented to them, and they were all familiar with computers and very competent with the mouse. They all were able to sustain attention to the task for at least ten minutes each session.

Teaching Set the Table
In order to establish a baseline for each child's table setting skills, a plate, napkin, cup, and silverware were placed on a real table with the direction, "Set the table." No child was able to perform this task with more than 50% accuracy. The first stage of the instructional computer program demonstrated the correct placement of silverware, cup, and napkin. Next, the computer voice gave the command to show, and then place each item (e.g.," Show napkin"; "Place napkin"). A video animation of a girl putting each item in its correct spot reinforced the placement. The next five phases of the computer lesson were variations of the same task, ranging from left-to-right placement of items to random selection with the computer voice giving a general command, "Set the table". Mastery was set at 90% accuracy for 75% of the five stages.

Generalization
A real world test was given by repeating the pretest with the real tableware. The plate was again placed on the table and the child was asked to "Set the table". Mastery in the real world was established at 90% accuracy. If this criterion were not met, we scheduled additional computer training sessions until the child again met the computer teaching phase criterion.

Results
All five children learned the table setting task using the interactive computer software and they all transferred this knowledge from the computer to the real world. Also, the use of the animation accompanied by the computer voice seemed to help establish the placement of the tableware. During the real world probe, several children talked themselves through the task, repeating the phrases used by the computer.

Maintenance
A one-week real-world follow-up confirmed mastery, with four of five students setting the table with no errors. The fifth student was 80% accurate with the real-world probe. After a one-phase computer review, he set both the computer and real-world tables with no errors.

return to top