Statistics is the grammar of science…#

-Karl Pearson, British mathematician (1857-1936)

Unless you have engaged in research, you may not realize how important statistical analysis is for the study of science (and many other fields, for that matter).

For scientists, statistical principles underlie how we design our experiments and how we test and support our theoretical hypotheses. As the introductory quote highlights, statistical analysis helps provide structure to scientific exploration and communication.

Statistical analysis also serves as a reality check for us. If the data does not support a particular hypothesis, then we may need to re-evaluate our experimental design, data collection protocol or our physical understanding of the phenomenon we are studying.

Finally, in this new world of Big Data, we are relying more and more on statistical methods to help us extract signal from noise.

Why are statistics so important?#

Although many students dislike taking statistics (I was one of them!), there is no escaping statistics in our everyday lives and even more so in our professional lives as scientists.

Statistics are used to support theoretical ideas with real data (e.g., climate change, clinical trials), to assess risk (e.g. insurance, disease prevention), to make predictions (e.g. politics, financial markets) and to facilitate decision-making more broadly.

Even though statistics are all around us, an appreciation of statistics arises by doing statistics, so this course is all about doing.

A Few Best Practices#

Before we jump in, here are a few best practices to keep in mind:

  • Visualize your data

    • ALWAYS look at your data before, during and after your analysis

  • Hypothesize first, test second

    • statistics can support, but not replace, sound scientific hypotheses

  • Keep it simple

    • the simpler the statistic, the better (assuming it applies)

  • Seek out feedback

    • when it comes to real-world data, things can get messy. If you are unsure about what your (or someone else’s) analysis is telling you, solicit a second opinion.

Remember, statistical tests are never conclusive. Statistics cannot prove hypotheses and, therefore, must be combined with sound scientific interpretation.