Related Experiment Videos
Quantitatively expressed ideas in lymphoma: text versus numerical displays
N Yamaguchi1, L Latinwo, R S Horowitz
1University of Occupational and Environmental Health, Kitakyushu, Japan.
Medical Informatics = Medecine Et Informatique
|October 1, 1987
Summary
Quantitative ideas in scientific papers are best found in tables and graphs, not just the text. This study found numerical displays more important and easier to process than text-based results for understanding research findings.
Area of Science:
- Scientific communication
- Information extraction
- Bibliometrics
Background:
- Quantitative ideas in scientific literature are presented in numerical displays (tables, graphs) and in the Results section text.
- Understanding the relative importance and accessibility of these formats is crucial for efficient scientific document processing.
Purpose of the Study:
- To compare the importance and extractability of quantitative ideas presented in numerical displays versus text within scientific articles.
- To evaluate the added value of reading Results sections beyond information conveyed by tables and graphs.
Main Methods:
- A random sample of 50 lymphoma literature articles with numerical relationship displays was selected.
- Quantitative ideas were extracted from both numerical displays and text in the Results section.
- A database was created to categorize ideas based on their presentation format.
Main Results:
- Quantitative ideas expressed in text were more difficult to find and extract compared to those in numerical displays.
- Ideas from numerical displays were found to be more important than those presented solely in the text.
- Extracting quantitative ideas from numerical displays is a validated method for rapid technical processing.
Conclusions:
- Numerical displays (tables, graphs) are a more significant and accessible source of quantitative information in scientific articles.
- The method of extracting quantitative ideas from numerical displays is effective for efficient scientific document processing.