Data, Estimation, and Statistical Reliability
9.0. Introduction: What Is the Issue?
One of the first things we do with business performance data is to summarize it. We might break it up into groups and look at how the performance measure differs between the groups. This is known as descriptive statistics. Tables of means and standard deviations allow us to extract the main patterns and features in what would otherwise be a meaningless spreadsheet of numbers. So, one function of the descriptive statistics is to convert data into information that can be understood. This facilitates communication of the main patterns to a third party. Another person may not have access to the raw data or the inclination to look at it, but descriptive statistics will convey the main patterns and relationships.
However, there is a much deeper reason that we might calculate summary statistics, such as the average customer satisfaction for different age groups. We do it because we think it tells us something fundamental about how age affects customer reaction. The sample means suggest something about the underlying structure of the business or, more accurately, about the business process that generated the data. Inferring something about the process fundamentals from data is known as inferential statistics.
This chapter is about a coherent conceptual basis for inference and data analysis. You could do data analysis without reading this chapter, but you would just be applying recipes. Specifically, ...