4.5 Editorial Material

Integrating Computer Prediction Methods in Social Science: A Comment on Hofman et al. (2021)

Related references

Note: Only part of the references are listed.
Article Computer Science, Interdisciplinary Applications

Theory-Driven Analysis of Large Corpora: Semisupervised Topic Classification of the UN Speeches

Kohei Watanabe et al.

Summary: There is a growing interest in quantitative analysis of large corpora among international relations scholars. To address the challenge of using unsupervised machine learning models consistently with existing theoretical frameworks, researchers have proposed a set of techniques that utilize a semisupervised model for efficient document classification. This approach involves creating a dictionary and using an entropy-based diagnostic tool to improve classification accuracy. Experimental results demonstrate the superiority of semisupervised models over unsupervised models, particularly when considering contextual information.

SOCIAL SCIENCE COMPUTER REVIEW (2022)

Article Computer Science, Interdisciplinary Applications

Overcoming the Challenges of Collaboratively Adopting Artificial Intelligence in the Public Sector

Averill Campion et al.

Summary: Despite the popularity of AI, there is limited research on its application in public contexts. This study examines the challenges faced by interorganizational collaborations in adopting AI tools and implementing organizational routines through a case study involving a university and two county councils in England. Findings highlight the key challenges, including resistance to data sharing, inadequate understanding of data requirements, misalignment of project interests and data sharing expectations, and lack of engagement across organizational hierarchy.

SOCIAL SCIENCE COMPUTER REVIEW (2022)

Article Computer Science, Interdisciplinary Applications

Will Algorithms Blind People? The Effect of Explainable AI and Decision-Makers' Experience on AI-supported Decision-Making in Government

Marijn Janssen et al.

Summary: This study compares decision making in different scenarios without algorithm support, with business rule support, and with machine learning support. The findings show that algorithms help decision makers make more correct decisions, but experienced individuals are still not able to identify all mistakes. Careful selection of algorithms and training of decision makers are crucial for increasing accountability and transparency.

SOCIAL SCIENCE COMPUTER REVIEW (2022)

Article Computer Science, Interdisciplinary Applications

The Market for Heritage: Evidence From eBay Using Natural Language Processing

Mark Altaweel

Summary: The trade in antiquities and cultural objects is complex and dynamic, with few computational tools available to systematically understand the legal and illegal markets. This study focused on lower-end but high sales volume objects sold on eBay's U.S. site between October 2018 and May 2019, revealing that the United Kingdom and the United States are major sources of ancient cultural objects. The developed machine learning approach and tool demonstrate the potential to monitor and gain insights into the online market for antiquities and cultural objects.

SOCIAL SCIENCE COMPUTER REVIEW (2021)

Article Computer Science, Interdisciplinary Applications

Citations Gone #Social: Examining the Effect of Altmetrics on Citations and Readership in Communication Research

Ben Wasike

Summary: The study shows that altmetrics have a positive impact on the citation of research articles, with postings on academic social networking sites like ResearchGate and Mendeley increasing both readership and citation likelihood. In addition, social media mentions, downloadable articles, coauthorship, and an active online presence among scholars are also factors that improve readership and citation.

SOCIAL SCIENCE COMPUTER REVIEW (2021)

Article Computer Science, Interdisciplinary Applications

Assessing Replicability of Machine Learning Results: An Introduction to Methods on Predictive Accuracy in Social Sciences

Ranjith Vijayakumar et al.

Summary: Machine learning methods have gained popularity for their emphasis on predictive accuracy, but little attention has been given to replicability. By adapting replication methods from psychology, this study introduces three replication aims and demonstrates how to combine replications for better precision in predictive accuracy measures. Regression was comparable to many complex methods, while support vector machine showed the most reliable performance across scenarios.

SOCIAL SCIENCE COMPUTER REVIEW (2021)

Article Multidisciplinary Sciences

Algorithmic and human prediction of success in human collaboration from visual features

Martin Saveski et al.

Summary: This study explores the characteristics of successful groups in an escape room game and examines the ability of humans and machines to identify successful groups from photos. Larger, older, and more gender-diverse groups are found to be more likely to succeed. Machines outperform humans in predicting group success based on visual cues, showing the potentials and limitations of machine learning systems in this context.

SCIENTIFIC REPORTS (2021)

Article Multidisciplinary Sciences

Integrating explanation and prediction in computational social science

Jake M. Hofman et al.

Summary: Computational social science involves the convergence of different fields with different ways of thinking and doing science. It emphasizes combining explanatory and predictive approaches, advocating for integrative modeling to further research goals.

NATURE (2021)

Article Multidisciplinary Sciences

New relevance and significance measures to replace p-values

Werner A. Stahel

Summary: The concept of p-value has been widely debated in recent decades, with criticism over its interpretation of testing unrealistic null hypotheses. The focus should shift towards determining the relevance of an effect, leading to the proposal of a relevant threshold and quantitative measure. Statistical inference should be based on confidence intervals for relevance, which could potentially lead to a more comprehensive classification of results beyond simply significant or non-significant findings. If desired, a single number called the secured relevance may provide a scientifically meaningful interpretation of the results.

PLOS ONE (2021)

Article Sociology

Does Sociology Need Open Science?

Nate Breznau

Summary: The paper discusses the need for sociology to change in response to reliability, transparency, and ethical crises that have affected other social science disciplines. It reviews sociology as a discipline by examining current practices, definitions, positions of sociological associations, and arguments from influential sociologists like Weber, Merton, and Habermas. The overview aims to help students and sociologists quickly familiarize themselves with the state of sociology and consider the relevance of open science to their discipline.

SOCIETIES (2021)

Article Political Science

I Saw You in the Crowd: Credibility, Reproducibility, and Meta-Utility

Nate Breznau

PS-POLITICAL SCIENCE & POLITICS (2021)

Article Psychology, Multidisciplinary

Practical Methodological Reform Needs Good Theory

Will M. Gervais

Summary: In the face of challenges in psychological sciences, reforms and metascience movements are calling for more theoretical insights to address issues. Looking at modeling methods from the philosophy of science and cultural evolution can help in changing norms and institutions in the field.

PERSPECTIVES ON PSYCHOLOGICAL SCIENCE (2021)

Article Psychology, Multidisciplinary

The Practical Alternative to the p Value Is the Correctly Used p Value

Daniel Lakens

Summary: Due to the strong reliance on p values in scientific literature, some researchers argue for moving beyond p values and embracing practical alternatives. It is suggested that statisticians should teach researchers what questions they can ask rather than telling them what they want to know, and including minimum-effect tests and equivalence tests in statistical toolbox may greatly improve the questions researchers ask. Developing better evidence-based education and user-centered statistical software to prevent misinterpretation of p values is seen as a top priority.

PERSPECTIVES ON PSYCHOLOGICAL SCIENCE (2021)

Article Psychology, Multidisciplinary

Why Hypothesis Testers Should Spend Less Time Testing Hypotheses

Anne M. Scheel et al.

Summary: Before testing hypotheses, researchers should focus on developing concepts, establishing valid measures, identifying causal relationships and functional forms, as well as determining boundary conditions and auxiliary assumptions. Providing these inputs should be recognized as a crucial goal.

PERSPECTIVES ON PSYCHOLOGICAL SCIENCE (2021)

Article Computer Science, Interdisciplinary Applications

When Digital Trace Data Meet Traditional Communication Theory: Theoretical/Methodological Directions

Sujin Choi

SOCIAL SCIENCE COMPUTER REVIEW (2020)

Letter Infectious Diseases

An interactive web-based dashboard to track COVID-19 in real time

Ensheng Dong et al.

LANCET INFECTIOUS DISEASES (2020)

Article Multidisciplinary Sciences

Measuring the predictability of life outcomes with a scientific mass collaboration

Matthew J. Salganik et al.

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA (2020)

Article Multidisciplinary Sciences

Scaling up psychology via Scientific Regret Minimization

Mayank Agrawal et al.

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA (2020)

Article Multidisciplinary Sciences

Variability in the analysis of a single neuroimaging dataset by many teams

Rotem Botvinik-Nezer et al.

NATURE (2020)

Review Health Care Sciences & Services

A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models

Evangelia Christodoulou et al.

JOURNAL OF CLINICAL EPIDEMIOLOGY (2019)

Article Multidisciplinary Sciences

The preregistration revolution

Brian A. Nosek et al.

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA (2018)

Article Multidisciplinary Sciences

The accuracy, fairness, and limits of predicting recidivism

Julia Dressel et al.

SCIENCE ADVANCES (2018)

Article Psychology

Many Analysts, One Data Set: Making Transparent How Variations in Analytic Choices Affect Results

R. Silberzahn et al.

ADVANCES IN METHODS AND PRACTICES IN PSYCHOLOGICAL SCIENCE (2018)

Article Economics

Machine Learning: An Applied Econometric Approach

Sendhil Mullainathan et al.

JOURNAL OF ECONOMIC PERSPECTIVES (2017)

Article Multidisciplinary Sciences

Estimating the reproducibility of psychological science

Alexander A. Aarts et al.

SCIENCE (2015)

Article Economics

COMPARING OPEN AND SEALED BID AUCTIONS: EVIDENCE FROM TIMBER AUCTIONS

Susan Athey et al.

QUARTERLY JOURNAL OF ECONOMICS (2011)

Editorial Material Multidisciplinary Sciences

SOCIAL SCIENCE Computational Social Science

David Lazer et al.

SCIENCE (2009)