Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Stereotype Content Model02:16

Stereotype Content Model

The Stereotype Content Model (SCM) was first proposed by Susan Fiske and her colleagues (Fiske, Cuddy, Glick & Xu, 2002; see also Fiske, 2012 and Fiske, 2017). The SCM specifies that when someone encounters a new group, they will stereotype them based on two metrics: warmth—or that group’s perceived intent, and how likely they are to provide help or inflict harm—and competence—or their ability to carry out that objective. Depending on the warmth-competence categorization, a person will feel...
Systematic Error: Methodological and Sampling Errors01:15

Systematic Error: Methodological and Sampling Errors

In the case of systematic errors, the sources can be identified, and the errors can be subsequently minimized by addressing these sources. According to the source, systematic errors can be divided into sampling, instrumental, methodological, and personal errors.
Sampling errors originate from improper sampling methods or the wrong sample population. These errors can be minimized by refining the sampling strategy. Defective instruments or faulty calibrations are the sources of instrumental...
Qualitative Analysis03:46

Qualitative Analysis

For solutions containing mixtures of different cations, the identity of each cation can be determined by qualitative analysis. This technique involves a series of selective precipitations with different chemical reagents, each reaction producing a characteristic precipitate for a specific group of cations. Metal ions within a group are further separated by varying the pH, heating the mixture to redissolve a precipitate, or adding other reagents to form complex ions.
For instance, group IV...
Qualitative Analysis01:10

Qualitative Analysis

Qualitative analysis is the process of identifying elements, ions, or compounds in an unknown sample. It is the first and most fundamental type of analysis based on the hierarchy of analytical goals. This hierarchy is significant as it provides a structured approach to scientific research, with qualitative analysis serving as the initial step, providing essential information before moving on to quantitative or other forms of analysis.
There are two main approaches to qualitative analysis:...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Scaling Surface-Guided Radiation Therapy to Larger Lung Cancer Cohorts: Frameless Immobilization and Enhanced Setup Accuracy.

Life (Basel, Switzerland)·2026
Same author

Postmastectomy radiation therapy for patients with pT1-2N1 breast cancer over 2 decades in Korea: Korean Radiation Oncology Group 14-21.

Radiation oncology journal·2026
Same author

Systematic review and meta-analysis of ocular complications following hyaluronic acid filler injection.

Plastic and reconstructive surgery·2025
Same author

Radiation Pneumonitis Risk Assessment Using Fractal Analyses in NSCLC Patients Treated with Curative-Intent Radiotherapy.

Life (Basel, Switzerland)·2025
Same author

Final Report on the Safety and Efficacy of Poly-L-Lactic Acid Filler (Gana V) Injection for the Correction of Nasolabial Fold: A Double-Blind, Non-Inferiority, Randomized, Split-Face Controlled Trial.

Aesthetic plastic surgery·2025
Same author

Translating SGRT from Breast to Lung Cancer: A Study on Frameless Immobilization and Real-Time Monitoring Efficacy, Focusing on Setup Accuracy.

Life (Basel, Switzerland)·2025

Related Experiment Video

Updated: Jul 7, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Qualitative Analysis of Discrepancy Patterns Between Large Language Models and Human Reviewers in Abstract Screening

Kyung Hwa Lee1, Hakyoung Kim1, Dae Sik Yang1

  • 1Department of Radiation Oncology, Korea University Guro Hospital, Korea University College of Medicine.

Journal of Epidemiology
|July 5, 2026
PubMed
Summary
This summary is machine-generated.

Large language models (LLMs) show promise for systematic review abstract screening, with GPT-5.0-mini offering a strong balance of sensitivity and efficiency. Most disagreements stem from screening conventions, not model errors, suggesting safe integration with safeguards.

Keywords:
Abstract Screening AutomationGPT ModelsLarge Language Models (LLMs)Qualitative Error AnalysisSystematic Review

Related Experiment Videos

Last Updated: Jul 7, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Area of Science:

  • Artificial Intelligence in Research
  • Biomedical Informatics
  • Systematic Review Methodology

Background:

  • Evaluating the reliability of large language models (LLMs) for abstract screening in systematic reviews.
  • Assessing model-human discordance to ensure safe integration into real-world review practices.

Purpose of the Study:

  • To assess the reliability of GPT-4.0, GPT-5.0, and GPT-5.0-mini for abstract screening.
  • To characterize model-human disagreements and inform workflow integration.

Main Methods:

  • Tested LLMs on two systematic review datasets with varying topic densities (core vs. peripheral subjects).
  • Defined datasets by target-to-background ratio (TBR) and used full-text inclusion as the reference standard.
  • Developed a taxonomy to classify disagreements (e.g., human leniency, ambiguity, true misses).

Main Results:

  • GPT-5.0-mini demonstrated the best sensitivity-efficiency trade-off, achieving high workload reduction (>92%) with over 83% sensitivity.
  • Negative predictive value exceeded 99% across both datasets.
  • Disagreements were minimal in core-subject datasets, while peripheral-subject datasets showed higher disagreement, primarily due to human leniency and ambiguity, with rare true LLM misses.

Conclusions:

  • Model-human disagreements often reflect screening conventions rather than intrinsic LLM failure.
  • LLM-assisted screening can enhance efficiency without sacrificing reliability when appropriate safeguards are in place for ambiguous records.