This study examined the effects of five dichotomous variables on human performance using a computer-based telephone information system. The five variables were: speech rate (120 or 240 words per minute), length of input time-out (two or ten seconds), feedback (available or not available), wallet guide - a graphical representation of the information (available or not available), and the database structure (8x2 or 2x6). The research methodology implemented a one-half fraction of a 2⁵ factorial design, requiring only 16 of the 32 possible treatment combinations.
Two tasks were included in this study: a search task and a transcription task. The search task consisted of each subject accessing an information system through a touch-tone telephone. The subject listened to the computer as it used synthesized speech to list available menu options. The search task continued until the subject found the target item. The transcription task consisted of listening to and typing an information message for each target item. The experiment ended when 16 target items were found.
Four dependent measures were used to evaluate user performance. The search task was evaluated with three measures: user added time (the amount of additional time the subject required to complete the search in excess of the minimum search time imposed by the system design); invalid key presses (the number of times undefined keys on the touch tone telephone were pressed during the search); and user added key presses (the number of additional, valid key presses the subject required to complete the search in excess of the minimum number of key presses required to complete the search). Only one measure was used to evaluate user performance of the transcription task: transcription accuracy score (the number of words that each subject transcribed correctly).
The results show four of the five variables (speech rate, database structure, input time-out, and wallet guide) to have a significant effect on human performance. The following interactions were found to have a significant effect on at least one of the dependent measures: database structure by input timeout, database structure by wallet guide, input timeout by wallet guide, and speech rate by wallet guide. Twelve subjective ratings were also analyzed. The results show at least one of the 12 subjective ratings was significantly affected by speech rate, input time-out, or the database structure.
Perhaps the most important finding of this research is that complicated auditory information structures can be accessed easily if a wallet guide is provided. In addition to decreasing search time, a wallet guide reduces the number of search errors users make. / Master of Science
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/44578 |
| Date | 05 September 2009 |
| Creators | Cary, Michele Marie |
| Contributors | Industrial and Systems Engineering, Williges, Robert C., Price, Dennis L., Williges, Beverly A. |
| Publisher | Virginia Tech |
| Source Sets | Virginia Tech Theses and Dissertation |
| Language | English |
| Detected Language | English |
| Type | Thesis, Text |
| Format | viii, 89 leaves, BTD, application/pdf, application/pdf |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
| Relation | OCLC# 30502867, LD5655.V855_1993.C3725.pdf |
Page generated in 0.0022 seconds