Friday, 27 October 2017

What is Research Replication?

Sage Publications have released a 30 minute video from Social Sciences author, Professor Gary King, who talks about what science is, and what the benefits of science are. He goes on to link these ideas with why scientists need to share results, concerns, methods, results and the data.

Academics don't currently share data at all. Our papers and books are more like "advertisements for the research rather than the research itself" (King, 2.37), because we don't share our "data, code and procedures" (King, 3.19) with the entire audience. I think Professor King makes a very good point. We cannot ensure academic integrity or honesty without seeing the data set and being able to experiment with all the data ourselves, seeking replicability. 

What should result, if someone follows our method accurately in a similar population, is that we should get replicable research. Professor King has, over the years, encouraged his students to do just that: and often, has found that students have neither been able to replicate the data nor the following analysis. Recently there have been some studies which have done this work on a much broader scale. Findings are that most research is not replicable, and that claims made from studies have been over-blown, causal links over-stretched. Professor King talks about why this is, in the following video.

Very interesting!


No comments :

Post a Comment