Open Science In Acoustics: Foundations & Best Practices.pdf

The scientific method is a systematic approach to research that has been characterized by natural science since the 17th century. It involves observation, measurement, and experiment, as well as the formulation, testing, and modification of hypotheses. Reproducibility of results is a cornerstone of the scientific method, defined as the ability to duplicate an entire analysis of an experiment or study. However, a wide range of scientific results have been found to be irreproducible, leading to a classification of the application areas of the scientific method into three branches: deductive, empirical, and computational. The deductive and empirical branches are traditional, while the computational branch is considered a potentially novel branch, covering results derived from large-scale simulations and data-driven computational science.

Measures to ensure reproducibility in the traditional branches are well known, but not in the computational branch, where intellectual contributions may be encoded in software only. Results are often not reproducible due to lack of availability of supplementary material such as protocols, data, and implementations. Open science focuses on the ease of access to scientific data underlying research, supporting the ease of reproducibility.

An alternative view on the problem of reproducibility is to ask who should benefit from research, with potential answers ranging from the researcher themselves to all people in the world. Research conducted for the sake of science itself should be as open as possible to support scientific advancement.

Open science aims to make the research process transparent on all levels, building on the general demand for socialization of knowledge. The advent of scientific journals is seen as a first step towards openness, but the exact meaning of the term is still not fully settled. Measures to achieve transparency in the research process include open source, open data, open access, open methodology, open notebook science, open educational resources, and open peer review.

A listening experiment serves as an example for the usage of the different elements of open science. The workflow of a typical perceptual experiment can be decomposed into stages, including idea, design of experiment, computation, experiment, analysis, manuscript, pre-publication peer review, publication, and aftermath. Each stage involves the application of open science elements, such as open notebook science, open data, open source, open access, open methodology, and open peer review.