SMPA 2151 Lecture Notes - Lecture 2: Null Hypothesis, Causal Inference
Document Summary
Statistics can be used to obscure or deceive. Change what we are measuring and call it something bigger. Of all problems confronting those of us who use data, human error looms largest. What other ways can our data be problematic, as referenced in the. Lesson for journalism: how to catch statistical lies. Consider the incentives facing the people producing or collecting the data. Little fudges made by well-meaning people can affect cumulative measurement, as in the case with the head start program. Conclusion of the study at first was that head start has no benefit. Head start employees were messing with the protocol. Human error, and human response to incentives, are constant features of human behavior in regards to data. How could we do in a way to make. Some problems are technical, others are conceptual, others are related. A model is a simplified description of something about the world.