{"abstract":"Inference is the process of using facts we know to learn about facts we do\nnot know. A theory of inference gives assumptions necessary to get from the\nformer to the latter, along with a definition for and summary of the resulting\nuncertainty. Any one theory of inference is neither right nor wrong, but merely\nan axiom that may or may not be useful. Each of the many diverse theories of\ninference can be valuable for certain applications. However, no existing theory\nof inference addresses the tendency to choose, from the range of plausible data\nanalysis specifications consistent with prior evidence, those that\ninadvertently favor one's own hypotheses. Since the biases from these choices\nare a growing concern across scientific fields, and in a sense the reason the\nscientific community was invented in the first place, we introduce a new theory\nof inference designed to address this critical problem. We introduce hacking\nintervals, which are the range of a summary statistic one may obtain given a\nclass of possible endogenous manipulations of the data. Hacking intervals\nrequire no appeal to hypothetical data sets drawn from imaginary\nsuperpopulations. A scientific result with a small hacking interval is more\nrobust to researcher manipulation than one with a larger interval, and is often\neasier to interpret than a classical confidence interval. Some versions of\nhacking intervals turn out to be equivalent to classical confidence intervals,\nwhich means they may also provide a more intuitive and potentially more useful\ninterpretation of classical confidence intervals.","author":[{"family":"Coker"},{"family":"Rudin"},{"family":"King"}],"id":"unknown","issued":{"date-parts":[[2020,10,13]]},"language":"en","title":"A Theory of Statistical Inference for Ensuring the Robustness of Scientific Results","type":"article"}