Data analysis – Issues, challenges and solutions

The first official meeting of the Preclinical Data Forum took place in Berlin in 2014 and one of the central seminars was given by Viswanath Devanarayan, head of Discovery Biostatistics at AbbVie at that time. And the first thing Devan did was to tell us the following story:
A biologist and a statistician are in the death row, and are to be executed around the same time. They are each granted one last request. The statistician: “I want to give one last seminar”. The biologist: “I want to be executed first”.
We invited Devan to explain the famous Ioannidis 2005 paper “Why Most Published Research Findings Are False”. Devan did a great job. Explanations were at our, non-statistician, level, simple and yet professional, convincing and leaving no doubt about the importance of the claims made in that paper.

In the ideal world, biomedical researchers would always go to professional statisticians for help. But we know that this is usually not possible for most preclinical scientists, who nevertheless need help and cannot get such simple answers from attending yet another statistics course or reading yet another book or other resource that is actually not meant for non-statisticians.
We would like to collect typical data analysis problems that scientists often face and for which solutions are not readily available. We will approach professional statisticians and, with their help, try to develop answers and examples that could help our peer scientists.

The first such piece is in the following section.

https://test2.paasp.net/wp-content/plugins/biological-vs-technical-replicates-now-from-a-data-analysis-perspective/
Share this

Leave a Reply

Your email address will not be published. Required fields are marked *