How to avoid seven deadly sins in the study of behavior

Scientific journals have become both thicker and more numerous during the last twenty years. An enormous flood of papers has driven the traditional refereeing system to its limits and scientists to increasing specialization. For someone of my generation who originally analyzed his data with pencil and paper, drew his figures with ink, and wrote his first publications on a mechanical typewriter, producing papers in our software age has become extremely easy and many times faster. Two steps in the paper milling process must still consume much time and effort, namely designing and performing experiments. The consequence of speeding these steps up is a high rate of flaws and mistakes in papers, even those that are published in respected journals; obviously, referees cannot manage (or are not trained) to find all the procedural mistakes when they do their altruistic job in their limited time. I should not be too pessimistic. I am happy to acknowledge that the application of statistical procedures has been improved on average in behavior research and that there are exceptional scientists who do high quality research despite a high publication rate. The following is a list of mistakes that I have often found in published research: