Plenary talk: “Significance Testing is the Reason that Scientific Results have Poor Reproducibility”
Speaker: Professor Timothy L. Lash, DSc, MPH, Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia, USA
About the speaker:
Timothy L. Lash, DSc, MPH, is Professor in the Department of Epidemiology at the Rollins School of Public Health. Dr. Lash is Honorary Professor of Cancer Epidemiology in the Department of Clinical Epidemiology at Aarhus University in Aarhus, Denmark. He was also President of the Society for Epidemiologic Research (SER) for the 2014-2015 term.
Dr. Lash is leader and member of the Cancer Prevention and Control Research Program at Winship Cancer Institute. Dr. Lash's research focuses on molecular biomarkers that predict cancer recurrence. He and his colleagues have developed large biobanks of tumor tissues from cancer patients, and connected these to clinical outcomes including cancer recurrence. He also has an interest in disparities in the quality of cancer care, especially age-related disparities in the quality of care. Older cancer patients who receive less than guideline care sometimes die of cancer that could have been successfully treated, which his research over the last decade has documented.
Dr. Lash is Editor-in-Chief of the journal Epidemiology. He is also on the editorial boards of Epidemiologic Methods and Epidemiologic Perspectives & Innovations. He has been published in a number of notable journals including The Journal of the National Cancer Institute, Epidemiology and International Journal of Epidemiology. He has coauthored two textbooks on epidemiologic methods.
In the last few years, stakeholders in the scientific community have raised alarms about a perceived lack of reproducibility of scientific results. In reaction, guidelines for journals have been promulgated and grant applicants have been asked to address the rigor and reproducibility of their proposed projects. Neither solution addresses a primary culprit, which is the culture of null hypothesis significance testing that dominates statistical analysis and inference. In an innovative research enterprise, selection of results for further evaluation based on null hypothesis significance testing is doomed to yield a low proportion of reproducible results and a high proportion of effects that are initially over-estimated. In addition, the culture of null hypothesis significance testing discourages quantitative adjustments to account for systematic errors and quantitative incorporation of prior information. These strategies would otherwise improve reproducibility, and have not been previously proposed in the widely-cited literature on this topic. Without discarding the culture of null hypothesis significance testing and implementing these alternative methods for statistical analysis and inference, all other strategies to improve reproducibility will yield marginal gains at best.
Networking, coffee and tea served after the talk.
The plenary talk is organized with financial support from the Strategic Research Program in Epidemiology (SFO Epi) at Karolinska Institutet, the Doctoral Program in Epidemiology at Karolinska Institutet and SINGS (the Swedish Interdisciplinary Graduate School in Register-based Research within SIMSAM).Contact person: Anita Berglund