A study of external subjective and external objective validations for an expert system to diagnose reading deficiencies
George Philip Bluhm, Lehigh University, United States
Lehigh University . Awarded
The dissertation considered the objective (quantitative) and subjective (qualitative) external validation of an expert system called IRX. IRX is an expert system designed to diagnose reading errors made during oral reading. The dissertation also addressed the question of whether the objective and subjective validations were equivalent. The Friedman statistic was used to determine whether or not a rank order was present.
The objective validation clearly showed that IRX outperformed ten reading diagnosticians on three levels of measurement, namely the uncorroborated, binary and majority counting methods. An inspection of these data showed that IRX was the only diagnostician that identified every possible diagnosis in each of the IRI (informal reading inventory) protocols.
The subjective validation showed that the four members of the peer review group were mixed in their evaluation. Two of the peer review group members ranked IRX the best of the diagnosticians. The remaining two members were unable to rank order the diagnosticians.
Bluhm, G.P. A study of external subjective and external objective validations for an expert system to diagnose reading deficiencies. Ph.D. thesis, Lehigh University. Retrieved March 21, 2019 from https://www.learntechlib.org/p/127967/.
Citation reproduced with permission of ProQuest LLC.
For copies of dissertations and theses: (800) 521-0600/(734) 761-4700 or https://dissexpress.umi.com