12.6 Chapter summary

Collectively, we want to find as many true positives as we can, whilst avoiding belief in false positives. We want to submit ideas to severe tests, whilst retaining the capacity to learn new things about the world. There are a few simple good practice guidelines to help you contribute constructively to this collective mission. Also pre-register your studies. Write everything you do up and pre-print it, reporting all measures, regardless of outcome. Perform and report sensitivity analysis. Replicate key results in an additional sample. Make your data and code publicly available and usable by others. Design an analysis strategy suitable for the type of study and goals you have. Make it as simple as possible and as complex as necessary. Try to avoid testing the same hypothesis with multiple statistical models. If you find yourself doing this, think about reducing your number of outcome variables, reducing your number of models, or at least correcting for multiple testing. Try to write elegant, comprehensible, well-structured code. Aim for intelligibility rather than minimal length. When you cut mere stones, try to imagine cathedrals.