Related Experiment Videos
Linking verbal and non-verbal representations: computer analysis of referential activity
1University of Ulm, Germany. merg@sip.medizin.uni-ulm.de
The British Journal of Medical Psychology
|October 19, 1999
Summary
Researchers developed a computer-assisted method to measure Referential Activity, the link between non-verbal experience and language. This new tool accurately models human scoring of this linguistic function.
Area of Science:
- Psycholinguistics
- Computational Linguistics
- Content Analysis
Background:
- Referential Activity (RA) is crucial for understanding the connection between non-verbal experiences and language.
- Accurate measurement of RA is essential for various psychological and linguistic research.
- Existing methods for scoring RA can be subjective and time-consuming.
Purpose of the Study:
- To develop a novel, computer-assisted procedure for modeling Referential Activity scales.
- To create an objective and efficient measure of Referential Activity based on text analysis.
Main Methods:
- Utilized a large, judge-rated text corpus, selecting extreme samples for Referential Activity (RA).
- Identified Characteristic Vocabularies (frequent words unique to each corpus end).
- Developed a Computerized Referential Activity (CRA) measure using 181 frequent words as dictionaries and content analysis techniques.
Main Results:
- A set of 181 frequent words was identified, accounting for 50% of the text corpora.
- The Computerized Referential Activity (CRA) measure demonstrated a correlation of approximately .50 with human judge scores.
- The CRA measure showed consistent performance across both development and test datasets.
Conclusions:
- The developed computer-assisted procedure provides a reliable method for measuring Referential Activity.
- This computational approach offers an efficient alternative to manual scoring of RA.
- The findings support the utility of content analysis techniques in psycholinguistic research.