NotesFAQContact Us
Search Tips
ERIC Number: ED445001
Record Type: Non-Journal
Publication Date: 2000-Apr
Pages: 12
Abstractor: N/A
Reference Count: N/A
Signal Detection with Latent Classes: A Perspective on Paper Grading.
DeCarlo, Lawrence T.
Signal detection theory (SDT) has been widely applied in situations where observers attempt to detect or discriminate between two or more events. The usefulness of SDT with latent classes was illustrated in the context of an educational situation that can be readily conceptualized as a signal detection task: grading term papers. The approach assumes that the graders attempt to discriminate between latent classes of papers by using a decision criteria in combination with their perceptions of the quality of each paper. Three graders (a professor and two graduate assistants) graded 85 term papers from a graduate course on measurement. A fit of the latent class signal detection model indicates that the graders discriminate equally between two latent classes, but their response criteria differ. These are similar to results typically found in signal detection experiments with observed events. The findings show that SDT offers a simple summary of the graders' performance in terms of their ability to discriminate between the latent classes and their arbitrary use of grade categories. (Contains 1 figure, 2 tables, and 11 references.) (SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A