Ideally, these choices are guided by the principles of best practice and prespecified in a publicly available protocol. For example, in nearly any analysis of data there are several “researcher degrees of freedom”- i.e., choices that must be made in the process of analysis. fishing, p-hacking), but each essentially involves probing the data in unplanned ways, finding and reporting an “attractive” result, without accurately conveying the course of analysis. Data-dredging bias is a general category which includes a number of misuses of statistical inference (e.g.
0 kommentar(er)
