Dataset Information

Observing many researchers using the same data and hypothesis reveals a hidden universe of uncertainty.

ABSTRACT: This study explores how researchers' analytical choices affect the reliability of scientific findings. Most discussions of reliability problems in science focus on systematic biases. We broaden the lens to emphasize the idiosyncrasy of conscious and unconscious decisions that researchers make during data analysis. We coordinated 161 researchers in 73 research teams and observed their research decisions as they used the same data to independently test the same prominent social science hypothesis: that greater immigration reduces support for social policies among the public. In this typical case of social science research, research teams reported both widely diverging numerical findings and substantive conclusions despite identical start conditions. Researchers' expertise, prior beliefs, and expectations barely predict the wide variation in research outcomes. More than 95% of the total variance in numerical results remains unexplained even after qualitative coding of all identifiable decisions in each team's workflow. This reveals a universe of uncertainty that remains hidden when considering a single study in isolation. The idiosyncratic nature of how researchers' results and conclusions varied is a previously underappreciated explanation for why many scientific hypotheses remain contested. These results call for greater epistemic humility and clarity in reporting scientific findings.

SUBMITTER: Breznau N

PROVIDER: S-EPMC9636921 | biostudies-literature | 2022 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Observing many researchers using the same data and hypothesis reveals a hidden universe of uncertainty.

Breznau Nate N Rinke Eike Mark EM Wuttke Alexander A Nguyen Hung H V HHV Adem Muna M Adriaans Jule J Alvarez-Benjumea Amalia A Andersen Henrik K HK Auer Daniel D Azevedo Flavio F Bahnsen Oke O Balzer Dave D Bauer Gerrit G Bauer Paul C PC Baumann Markus M Baute Sharon S Benoit Verena V Bernauer Julian J Berning Carl C Berthold Anna A Bethke Felix S FS Biegert Thomas T Blinzler Katharina K Blumenberg Johannes N JN Bobzien Licia L Bohman Andrea A Bol Thijs T Bostic Amie A Brzozowska Zuzanna Z Burgdorf Katharina K Burger Kaspar K Busch Kathrin B KB Carlos-Castillo Juan J Chan Nathan N Christmann Pablo P Connelly Roxanne R Czymara Christian S CS Damian Elena E Ecker Alejandro A Edelmann Achim A Eger Maureen A MA Ellerbrock Simon S Forke Anna A Forster Andrea A Gaasendam Chris C Gavras Konstantin K Gayle Vernon V Gessler Theresa T Gnambs Timo T Godefroidt Amélie A Grömping Max M Groß Martin M Gruber Stefan S Gummer Tobias T Hadjar Andreas A Heisig Jan Paul JP Hellmeier Sebastian S Heyne Stefanie S Hirsch Magdalena M Hjerm Mikael M Hochman Oshrat O Hövermann Andreas A Hunger Sophia S Hunkler Christian C Huth Nora N Ignácz Zsófia S ZS Jacobs Laura L Jacobsen Jannes J Jaeger Bastian B Jungkunz Sebastian S Jungmann Nils N Kauff Mathias M Kleinert Manuel M Klinger Julia J Kolb Jan-Philipp JP Kołczyńska Marta M Kuk John J Kunißen Katharina K Kurti Sinatra Dafina D Langenkamp Alexander A Lersch Philipp M PM Löbel Lea-Maria LM Lutscher Philipp P Mader Matthias M Madia Joan E JE Malancu Natalia N Maldonado Luis L Marahrens Helge H Martin Nicole N Martinez Paul P Mayerl Jochen J Mayorga Oscar J OJ McManus Patricia P McWagner Kyle K Meeusen Cecil C Meierrieks Daniel D Mellon Jonathan J Merhout Friedolin F Merk Samuel S Meyer Daniel D Micheli Leticia L Mijs Jonathan J Moya Cristóbal C Neunhoeffer Marcel M Nüst Daniel D Nygård Olav O Ochsenfeld Fabian F Otte Gunnar G Pechenkina Anna O AO Prosser Christopher C Raes Louis L Ralston Kevin K Ramos Miguel R MR Roets Arne A Rogers Jonathan J Ropers Guido G Samuel Robin R Sand Gregor G Schachter Ariela A Schaeffer Merlin M Schieferdecker David D Schlueter Elmar E Schmidt Regine R Schmidt Katja M KM Schmidt-Catran Alexander A Schmiedeberg Claudia C Schneider Jürgen J Schoonvelde Martijn M Schulte-Cloos Julia J Schumann Sandy S Schunck Reinhard R Schupp Jürgen J Seuring Julian J Silber Henning H Sleegers Willem W Sonntag Nico N Staudt Alexander A Steiber Nadia N Steiner Nils N Sternberg Sebastian S Stiers Dieter D Stojmenovska Dragana D Storz Nora N Striessnig Erich E Stroppe Anne-Kathrin AK Teltemann Janna J Tibajev Andrey A Tung Brian B Vagni Giacomo G Van Assche Jasper J van der Linden Meta M van der Noll Jolanda J Van Hootegem Arno A Vogtenhuber Stefan S Voicu Bogdan B Wagemans Fieke F Wehl Nadja N Werner Hannah H Wiernik Brenton M BM Winter Fabian F Wolf Christof C Yamada Yuki Y Zhang Nan N Ziller Conrad C Zins Stefan S Żółtak Tomasz T

Proceedings of the National Academy of Sciences of the United States of America 20221028 44

This study explores how researchers' analytical choices affect the reliability of scientific findings. Most discussions of reliability problems in science focus on systematic biases. We broaden the lens to emphasize the idiosyncrasy of conscious and unconscious decisions that researchers make during data analysis. We coordinated 161 researchers in 73 research teams and observed their research decisions as they used the same data to independently test the same prominent social science hypothesis: ...[more]

PMID: 36306328

Similar Datasets

Project description:BACKGROUND:E-science technologies have significantly increased the availability of data. Research grant providers such as the European Union increasingly require open access publishing of research results and data. However, despite its significance to research, the adoption rate of open data technology remains low across all disciplines, especially in Europe where research has primarily focused on technical solutions (such as Zenodo or the Open Science Framework) or considered only parts of the issue. METHODS AND FINDINGS:In this study, we emphasized the non-technical factors perceived value and uncertainty factors in the context of academia, which impact researchers' acceptance of open data-the idea that researchers should not only publish their findings in the form of articles or reports, but also share the corresponding raw data sets. We present the results of a broad quantitative analysis including N = 995 researchers from 13 large to medium-sized universities in Germany. In order to test 11 hypotheses regarding researchers' intentions to share their data, as well as detect any hierarchical or disciplinary differences, we employed a structured equation model (SEM) following the partial least squares (PLS) modeling approach. CONCLUSIONS:Grounded in the value-based theory, this article proclaims that most individuals in academia embrace open data when the perceived advantages outweigh the disadvantages. Furthermore, uncertainty factors impact the perceived value (consisting of the perceived advantages and disadvantages) of sharing research data. We found that researchers' assumptions about effort required during the data preparation process were diminished by awareness of e-science technologies (such as Zenodo or the Open Science Framework), which also increased their tendency to perceive personal benefits via data exchange. Uncertainty factors seem to influence the intention to share data. Effects differ between disciplines and hierarchical levels.

Dataset Information

Observing many researchers using the same data and hypothesis reveals a hidden universe of uncertainty.

Publications

Observing many researchers using the same data and hypothesis reveals a hidden universe of uncertainty.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets