Ideas for statistical education
During the search conference last March, we set the goal to organize a meeting to share (ideas for) statistical education…
Read more
See here the list with interesting references related to our project.
The NHST Myth: The project
Scientific integrity refers to conducting honest and sound research. Although fraud is the most obvious example going against the practice of scientific integrity, sloppy science is a much more frequently observed problem. The abuse of the Null Hypothesis Significance Test is a good example of sloppy science, that may have major consequences. Read on to find out more about our project on null hypothesis testing and scientific integrity.
Relevance
Statistical testing of null hypotheses is a daily routine for many researchers. Researchers often do not sufficiently realize that with null hypothesis significance testing (NHST) they do not draw conclusions about their data directly, but indirectly. As a result, conclusions from scientific research are often incorrect. Methodologists have been pointing this out for years. Nevertheless, the practice of NHST only seems to be increasing, while there are a number of better alternatives available. We wonder which perceptions or beliefs are at the base if this.
The Study
In this study, we have, by means of interviews and focus groups, investigated researchers’ views on the use of NHST or alternative methods and what they see as the advantages and disadvantages of these methods. We also identified researchers who are, or want to be, pioneers in promoting the use of better alternatives to NHST. Together, we are developing strategies to implement these alternatives into scientific research.
Objectives of the project
Our study consists of two parts. In the first part we aim to gain insight into the views of stakeholders on the use of Null Hypothesis Significance Testing. The aim is to gain knowledge about the perceptions of stakeholders of null hypothesis significance testing and to map their motives for using this method. With the second part of the study, we aim to develop strategies that will overcome the default use of Null Hypothesis Significance Testing and that will stimulate use of better alternatives.
In the first part of the study, we held interviews with representatives of several stakeholder groups from within the science system: researchers, editors, lecturers of statistics classes and funding agencies. The results of these interviews served as input for focus groups with the same groups of stakeholders. In the focus groups, alternatives to null hypothesis significance testing were further explored and concretized.
In the second part of the research, a search conferences was held in which researchers discussed strategies to change current ubiquitous use of NHST in science and higher education.
Timeline
Spring 2017: Application approved
August-November 2017: Rewrite protocol and registration
November 2017- June 2018: Interviews
September-December 2018: Focus groups
March 2019: Search conference
April-December: data analysis and reporting
The NHST myth: About us
This research project is being conducted by the Health Sciences department of the VU Amsterdam, the Netherlands and is embedded in the Amsterdam Public Health research institute. The project team consists of researchers from inside and outside the VU.
Publications
Agenda
Search Conference, March 25, 2019
During the search conference we developed, together with representatives of the stakeholder groups (researchers, statistics teachers, editors of scientific journals…
Presentations
Here you can find and download our presentations.
Contact us
Are you interested in our project, or do you have any questions?
Please contact us. You can find more information on this topic via our selected list of interesting sites and organizations.
Email: info@themythofnhst.com
This March 25th, our NHST search conference will take place! During the SC we'll work with researchers, lecturers in statistics, editors of scientific journals & funding agencies to think about implementing alternatives for NHST in scientific research & education