ERIC Number: ED222032
Record Type: RIE
Publication Date: 1982-Aug
Reference Count: 0
Computer-Guided Diagnosis of Learning Disabilities: A Prototype.
Colbourn, Marlene Jones
A computer based diagnostic system to assist educators in the assessment of learning disabled children aged 8 to 10 years in the area of reading is described and evaluated. The system is intended to guide the diagnosis of reading problems through step by step analysis of available data and requests for additional data. The system provides a diagnostic report of its findings including areas and skills which require remediation or further assessment. The model on which the educational diagnostic process is based is the McLeod Educational Diagnostic Model consisting of four stages: retrospective, definitive, analytic, and prescriptive. Within each stage, three levels are considered--basic educational skills, psychoeducational skills, and noneducational factors. The system is intended to be used by teachers of regular classrooms or of resource rooms who have some familiarity with standard diagnostic procedures. The programing language LISP and a DEC 2060 computer were used to develop the program. The system was evaluated by comparing machine diagnoses to human diagnostic reports. This required encoding of the human diagnostician reports in terms of the expert system's descriptors and descriptor classification scheme. Results indicated that the export system's diagnoses were accurate, often providing more detailed information than human reports. Human reports often contained subjective impressions of child attitude and deportment not in the computerized reports. The system is thought to be appropriate for extension into mathematics and spelling, into other age ranges, and for use on microcomputers. Much of the report consists of four appendixes containing definitions of terms, an example of the system's performance, experimental data, and the encoded control structure. (DB)
Publication Type: Dissertations/Theses - Masters Theses; Reports - Evaluative
Education Level: N/A
Sponsor: Saskatchewan School Trustees Association, Regina.
Authoring Institution: Saskatchewan Univ., Saskatoon.
Note: Master's Thesis; University of Saskatchewan. Publication of the Dept. of Computational Science.