Hoi K. Suen, Pui-Wa Lei, and Hongli Li
Data analyses are statistical manipulations of numbers to help discern patterns in the data. As such, they need guidance from substantive theories in order to decide what patterns to look for, and to make sense of these patterns. The core outcomes of these manipulations are summary descriptors used to reduce the amount of information a researcher needs to digest. When such descriptors summarize data from a sample, and a researcher wishes to make inferences from the sample descriptors to the population, various inferential techniques are used. These techniques approach the inferential task from three different angles: significance testing, parameter estimation, and statistical modeling. For decisions regarding a single individual, reliability and validity of information need to be assessed. For the evaluation of the efficacy of intervention on an individual, however, the typical design used is that of an interrupted time-series analysis.