PhD Level Course: Transparent, Open, and Reproducible Policy Research Catalyst Training ProjectPublic Policy

Sean Grant

This course aims to train graduate students at the Pardee RAND Graduate School in Policy Analysis and researchers at the RAND Corporation in the essential theories and methods of transparent, open, and reproducible research. This elective can be taken by graduate students with an analytic concentration in “Social and Behavioral Science” as well as RAND staff eligible to audit Pardee courses. After introductory sessions on the research transparency and open science movement, enrolled students will discuss and learn the essentials of the open science toolkit, namely: pre-registration of study methods and analysis plans; reproducible workflows; open research materials, code, data, and access; reporting guidelines for preparing research manuscripts; and replication.

Reproducible Research Computational Tools for the Next Generation of Social, Behavioral and Policy Scholars Catalyst Training ProjectInterdisciplinary

Jose M. Magallanes

In developing countries, accepted applicants to pursue PhD studies in highly-developed countries in the Social, Behavioral and Policy (SBP) fields will face several challenges. It will be particularly challenging to develop abilities to carry out research in a more transparent way, considering that this topic may not be present in the curriculum. There is enough time for them to learn the basic tools of reproducibility before they leave their countries, so as to encourage their adoption of transparent practices.

To master these tools, this workshop motivates and trains accepted applicants who are about to start their SBP PhD studies. The content will consider the background of SBP students in Latin America, and offer several workshops in Peru. The workshop will guide the students into the steps to ensure their research in shareable and replicable via the use of tools such as LaTeX, R, RStudio, Zotero, GitHub/AWS, Sharelatex/Overleaf, and Docker. This project will also provide curricular material in Spanish, which can be used for future activities in the region. It is also a great opportunity to involve universities and government agencies funding international higher education; which may increase the interest in the introduction of these topics in the higher education SBP curriculum.

Southeast Asian Social Sciences Workshops for Transparency and Reproducibility Catalyst Training ProjectInterdisciplinary

Ben Marwick

Reproducibility, transparency and openness of research has received a lot of attention and resources in Western countries, but the global South and East have had fewer opportunities to participate in these important developments in many research areas. This project will bring to five mainland Southeast Asian institutions with whom the Catalyst has collaborated a two-day workshop to motivate and update social scientists in the global East about current issues relating to the reproducibility, transparency and openness of research, and provide them with skills to improve their research along these lines.

The Catalyst will draw on BITSS materials and his own publications and presentations using the R programming language to teach general social science skills to enhance the reproducibility, transparency and openness of research. The workshops will be optimized for low-bandwidth internet, drawing on previous experience teaching computational methods in these locations. Workshop summaries will be generated in local languages and freely distributed online and in hard copy to more easily transmit the contents to students and other scholars.

Knowledge Sharing on Best Practices in Research Transparency and Reproducibility to Address the File Drawer Problem and Publication Bias in Rwanda Catalyst Training ProjectInternational Development

Mercyline Kamande

This project seeks to share knowledge and skills to motivate open, transparent, and reproducible research in order to address the file drawer problem and publication bias in Rwanda, as well as to improve the usability of research findings for policy-making. The Catalyst will first conduct exploratory research on the extent of the file drawer and publication bias problems in Rwanda. The project will then  introduce to scholars in the areas of economics, business studies, social science and public health the tools that exist to address the problem, as well as the benefits of using best practices in open, transparent, and reproducible research. These practices will include pre-registration and pre-analysis plans; sharing materials, data, and code using platforms like the Open Science Framework; and replication studies and meta-analysis.

Research Transparency and Reproducibility workshops will be held at two of the leading universities in Rwanda – the College of Business and Economics at University of Rwanda and Mount Kenya University Rwanda. Independent researchers from the Economic Policy Research Network will also be included. To ensure learning, weekly peer coaching meetings will be conducted by IERD to extend the peer learning experience realized by those enrolled in an online course currently in progress. The project will also include policy dialogue sessions with other stakeholders including Rwandan policymakers.

Research Transparency and Reproducibility in the Social Sciences and Impact Evaluation Workshop Catalyst Training ProjectImpact Evaluation

Elise Wang Sonne

In order to raise the quality of academic manuscripts written by Niger’s graduate students to allow them to compete in the demanding arena of research, the Catalyst has partnered with the UC Berkeley OASIS Initiative to organize a three-day intensive bootcamp on research transparency and reproducibility in social science research and impact evaluation.

The aim of this activity is to (i) raise the awareness of Masters and PhD students in public health, sociology, psychology, political sciences, economics, and statistics of the Abdou Moumouni University in Niamey (Niger) of common academic research misconduct issues and present solutions to avoid and address them; and (ii) train researchers working for the UC Berkeley-OASIS Initiative in Niamey on data management and organization, developing pre-analysis plans, pre-registering studies, and dynamic documents, which will be extremely useful for their day-to-day tasks pertaining to ongoing impact evaluation projects.

Fostering Transparency in Government Institutions and Higher Education: A Research and Teaching Initiative Catalyst Training ProjectInterdisciplinary

Nicole Janz Dalson Figueiredo

Locations: University of Nottingham, UK; Recife, Brazil; Brasilia, Brazil

We find research findings resulting from data that is not publicly accessible to not be credible. Similarly, governments withholding administrative information should not be trusted. We argue that the lack of government and research transparency are connected, and can be tackled in by offering training on reproducibility. This project aims to foster transparency in scholarly research and in government institutions. In particular, we will conduct educational workshops that will leverage insights that have been used to increase governmental and research transparency in the UK to improve transparency in Brazil. Our target groups are 100 undergraduate and graduate students, 20 scholars, and 20 bureaucrats. The project will strengthen research skills and transparency norms that can contribute to scientific innovation, development, and social welfare.

The first workshop was part of a day-long conference: “The Gold Standard of Reproducible Research” at the University of Nottingham on March 9, 2017.

Best Practices of Openness for African Researchers and Research Transparency Workshops at Three Social Science Conferences Catalyst Training ProjectInterdisciplinary

Dief Reagen Nochi Faha Elise Wang Sonne

Locations: LSE-Africa Summit, London School of Economics, London, UK; Population Association of America, Chicago, IL; University of Dschang, Cameroon; UNU-MERIT, Maastricht, Netherlands

This project will communicate best practices for openness and reproducibility in research. We will hold a workshop for African researchers at the University of Dschang in Cameroon, focusing on sensitizing researchers to the necessity of avoiding academic research misconduct such as p-hacking, publication bias, and failure to replicate, but also on data management practices in Stata. A series of workshops will also be held for public policy graduate students, demographers, sociologists, economists, and public health professionals in the Netherlands, USA, and the UK. In addition to sensitizing researchers to the necessity of avoiding academic research misconduct and data management practices, these workshops will also include trainings on Github, the Open Science Framework, Project TIER, and Dynamic Documents using StatTag and Markdoc.

The first workshop “Research Transparency and Reproducibility in the Social Sciences” will be held on March 31, 2017 at the 2017 LSE Africa Summit Research Conference in London.

Improving transparency of complex interventions through the facilitation of process evaluation training Behavior ChangeCatalyst Training Project

Elaine Toomey

Locations: National University of Ireland, Galway (NUIG), Health Behaviour Change Research Group (HBCRG)

Process evaluation is a way of investigating how well an intervention, programme or treatment was implemented as intended. It is crucial for facilitating transparency in the development, conduct and reporting of interventions in numerous research fields, including psychology, social science and public health, as it helps to increase confidence that changes in study outcomes are due to the influence of the intervention being investigated, and not due to differences or variability in the implementation of the intervention. It is particularly important within complex interventions (interventions with several interacting components such as multiple providers or intervention sites) due to the number of components that can be implemented variably and thus influence outcomes separately. This increases scientific confidence in the results of the intervention, enhances the replication and implementation of effective interventions, facilitates the refinement or de-implementation of ineffective interventions and overall serves to reduce research waste. However, at present specific process evaluation training is not available in Ireland which represents a significant barrier to the transparency of Irish intervention research and the implementation of best quality evidence into practice.

At present, the gold-standard for process evaluation training in Europe is run by the Centre for Development and Evaluation of Complex Interventions for Public Health Improvement (DECIPHer) in the UK, an International Centre of Excellence. This project aims to facilitate world-class DECIPHer process evaluation training in Ireland to improve the transparency of complex interventions in psychology, public health or social science settings and overall enhance the quality and impact of this research. Subsequent dissemination of the workshop proceedings and materials will also promote further understanding of this work amongst the wider BITSS community.

This project is supported in part by the John Templeton Foundation.

Research Transparency in the Social Sciences Workshop, Second Edition Catalyst Training ProjectInterdisciplinary

Zacharie Dimbuene

Location: University of Kinshasa (The Democratic Republic of the Congo)

Research Transparency is gaining attention in the scientific community around the world, including the United States, European countries, and Anglophone countries in sub-Saharan Africa; yet the concept is quite a “new world” in Francophone Africa. In my efforts to advance the movement in Francophone Africa, I successfully delivered the first Research Transparency Workshop at the University of Kinshasa (Democratic Republic of the Congo). This project is intended to sustain previous efforts to set up “Research Transparency in the Social Sciences” as a culture in the next generation of social scientists in Francophone Africa.

I will offer a training workshop for 60 graduate students, research staff, and professors at the University of Kinshasa to promote best practices for reproducible research with concrete guidance about how to make materials understandable for publication. The activities to be addressed during the workshop will include (1) organizing file structure; (2) creating understandable variable labels and value codes, as well as connecting variables to survey instruments through consistent labels and codebook creation; (3) version control of code and data; and (4) creating and maintaining documentation files about the survey and data, as well as data cleaning steps.

This workshop is supported in part by the John Templeton Foundation.

Development and implementation of a short course in Open Science Catalyst Training ProjectInterdisciplinary

Arnaud Vaganay

Locations: Utrecht, Netherlands; Essex, UK; London School of Economics, UK

I will coordinate a multi-site, university-accredited, and extra-curricular summer course in Open Science to be held at the Utrecht Summer School, Essex Summer School, and LSE Summer School. This course will build on existing materials and programs (including those of BITSS and COS), has been endorsed by the Center for Open Science, and will provide ECTS credits to students taking an optional exam. The exam is being developed with Thomas Leeper (Assistant Professor in Political Behaviour at LSE) who will also assist in teaching the course at LSE.

The course aims to develop the perspectives, knowledge, and skills researchers need to make their research more open, i.e. more transparent and reproducible. Upon completion of the course, students will be able to (1) define open science and evaluate the openness of current research; (2) discuss the main drivers and obstacles to openness and critically assess solutions; (3) implement fundamental open science practices in their own workflows; and (4) apply these skills through the use of open science software and apps. Each course will target 20 graduate students and early career researchers.

Disseminating Research Transparency in Perú, Bolivia, and Chile Catalyst Training ProjectInterdisciplinary

F. Hoces de la Guardia

Locations: Peru; Bolivia; Chile (final locations TBA)

The goal of this project is to bring to the attention of the academic communities in Peru, Bolivia, and Chile the recent developments in science regarding transparency and openness. This will be done in a two-fold format. First, a seminar-style talk will present the key issues (the reproducibility crisis, specification searching, and publication bias) and its solutions (pre-registrations, the TOP guidelines, and other tools for reproducible research). Second, a day-long workshop aimed at students will present the main tools for reproducible research (including R, Dynamic Documents, Git, and OSF). Increasing the scope of the research transparency community to this region can have additional benefits as it would bring highly talented researchers and students to elements of frontier research that are usually undisclosed in published papers.

Introducing the Transparent and Reproducible Research Paradigm in Ugandan Higher Institutions of Learning Catalyst Training ProjectInterdisciplinary

Jayne Byakika-Tusiime Saint Kizito Omala

Locations: Universities across Uganda (final locations TBA)

The concept of transparent and reproducible research is not known, nor appreciated, by researchers in Uganda and many other developing countries. There is usually delayed adoption of new knowledge and technologies in developing countries because of the slower flow of information in these regions. The concept of transparent and reproducible research is still relatively new even in the developed world and almost unknown in the developing world. As Catalysts of this paradigm shift, we wish to introduce this concept in Uganda. Groundbreaking research, especially in health has been done in Uganda and much more research continues to be done. However, the practice of transparent and reproducible is non-existent. We thus propose to start in Uganda a project to train 500 established and upcoming researchers in conducting transparent and reproducible research.

Our target population will be graduate and undergraduate students from 30 universities in Uganda, as well as faculty of both public and private universities that train students in research disciplines where theses, dissertations or journal article publications are required for either degree award or promotion. The objective of the project will be to sensitize and create awareness about conducting transparent and reproducible research. Specifically, we shall conduct ten regional workshops across Uganda. Following this introduction, we plan to design course modules to incorporate into existing academic programs at the participating universities.

Development of a Graduate Public Health Online Course in Research Integrity, Transparency, and Reproducibility Catalyst Training ProjectEpidemiology

Dennis M. Gorman

Location: Texas A&M University, USA

There is now a growing recognition within the scientific community that flexibility in study design, data analysis, and the reporting of research findings is increasingly leading to the publication of misleading results that capitalize on chance and cannot be replicated. It has been suggested that the use of such practices, if not made apparent in a manuscript describing the results of a study, is a form of research misconduct. There is little doubt that the widespread use of such practices undermines the integrity of a scientific field as they produce a body of non-reproducible, random findings. Both epidemiology and general public health are among the fields of research in which questions have been raised about research integrity, transparency, and reproducibility.

This course will examine various threats to the integrity of research, the professional and organizational factors that produce these threats, and the solutions that have been suggested to improve research quality (such as registered reports, open data, and team of rivals). Upon completion of the course, students should have the ability to differentiate research that is conducted with integrity and capable of producing valid and reproducible findings from research that is conducted without integrity and produces chance results that are trivial and non-reproducible. Students should also have the ability to incorporate practices into their own research that will increase its transparency and ensure it is conducted with integrity.

Creating Pedagogical Materials to Enhance Research Transparency at UCSD Catalyst Training ProjectInterdisciplinary

Scott Desposato Craig McIntosh

Location: UC San Diego (UCSD), USA

We will develop a core of teaching material around transparency and reproducibility that can be incorporated into graduate courses across the social sciences at UCSD. This project will draw on the library of materials from BITSS as well as from faculty at UCSD and the tools developed through the Policy Design and Evaluation Lab (PDEL)’s Data Replication service to create a one-week short course that can be deployed across courses. Our goal is to educate every new social science PhD student at UCSD about the importance of transparency and replicable research and empower them to incorporate transparency practices in their research from their first quarter at UCSD. Curricula will be made available on the BITSS library of educational materials as an alternative to the full semester course, and encourage the development of a set of discipline-specific add-on modules.

After completing the module, students will understand the importance of transparency for the scientific enterprise, they will recognize the institutional and incentive challenges to replicable research, they will be empowered with appropriate tools to adopt replicable practices, and they will understand the career costs and benefits of transparency.