Data sharing practices and data availability upon request differ across scientific disciplines

Abstract

Data sharing is one of the cornerstones of modern science that enables large-scale analyses and reproducibility. We evaluated data availability in research articles across nine disciplines in Nature and Science magazines and recorded corresponding authors' concerns, requests and reasons for declining data sharing. Although data sharing has improved in the last decade and particularly in recent years, data availability and willingness to share data still differ greatly among disciplines. We observed that statements of data availability upon (reasonable) request are inefficient and should not be allowed by journals. To improve data sharing at the time of manuscript acceptance, researchers should be better motivated to release their data with real benefits such as recognition, or bonus points in grant and job applications. We recommend that data management costs should be covered by funding agencies; publicly available research data ought to be included in the evaluation of applications; and surveillance of data sharing should be enforced by both academic publishers and funders. These cross-discipline survey data are available from the plutoF repository.

Fig. 1

Schematic rationale of the study.

Fig. 2

Differences in partial (grey) and full (black) data availability among disciplines depending on journal and publishing period (P1, 2000–2009; P2, 2010–2019) before contacting the authors (n = 875). Letters above bars indicate statistically significant difference groups among disciplines in full data availability compared to no data availability. Asterisks show significant differences in full data availability between journals and publishing periods.

Fig. 3

Types of critical data (n = 875). (a) Distribution of data types among disciplines (blue, dataset; purple, image; black, model); (b) Partial (light shades) and full (dark shades) data availability among disciplines depending on the type of critical data (DS, dataset; Img, image; Mod, model) before contacting the author(s).

Fig. 4

Differences in partial (grey) and full (black) data availability among disciplines after data requests (n = 672) depending on the type of critical data (DS, dataset; image; model) and publishing period (P1, 2000–2009; P2, 2010–2019). Numbers above bars indicate statistically significant difference groups among disciplines in full data availability.

Fig. 5

Histogram of time for receiving data from authors upon request within the 60-day reasonable time period (blue bars) and beyond (purple bar; data excluded from analyses; n = 199 requests). Note the 2-base logarithmic scale until 60 days.

Fig. 6

Authors’ response to data request (n = 199) depending on discipline (blue, declined; orange, ignored; purple, obtained). Bars indicate 95% CI of Sison and Glaz. Letters above bars indicate statistically significant difference groups in frequency of data availability by each category based on Tukey post-hoc test and Bonferroni correction.

Fig. 7

Decay in critical data availability initially (blue circles; n = 672), at the end of a 60-day contacting period (purple circles; n = 672) and upon request from the authors (black circles; n = 310).

Fig. 8

Frequency distribution of authors’ (a) Concerns and requests (n = 199) and (b) reasons for declining data sharing (n = 67). White bars indicate answers where no concerns or reasons were specified.

Fig. 9

Preferred ways of data storage in articles (n = 875) representing different disciplines (blue, text and supplement; purple, data archive; yellow, authors’ homepage; vermillion, previous publications; grey, museum; black, upon (reasonable) request; white, none declared.