British American Tobacco ensures that all its scientific work incorporates up-to-date and best practice in statistics and data handling.
Statisticians may perform statistical analysis, conduct research on statistical methods, advise on experimental design, and assist in interpreting and reporting of data. They also support and train scientists, and offer statistical services for quality control purposes across the British American Tobacco group.
Our statisticians are involved in most areas of our research, including the design of experiments. In this instance they may help define experimental protocol, including appropriate sample sizes; design questionnaires in such a way that measurements are accurate, sensitive and repeatable; and identify new hypotheses for further study.
As Ronald Fisher, the recognised founder of experimental design, says: “To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of.”
Good data management enables reliable statistics. Implementing data standards allows data sets from various sources to be combined and analysed, as well as allowing data mining. British American Tobacco has adopted a proactive strategy in applying data management standards and procedures that are compliant with good data management practices This is essential for demonstrating reproducible research with robust standards and quality, and meeting regulatory requirements.