Chapter 5: Regression Analysis Decisions

How did that happen?

Pivot table analysis is the starting point for identifying relationships between data elements (which is discussed in Chapter 8). Items of interest can be analyzed by using a one-by-one matrix, showing one variable in the rows of a spreadsheet and the other in the columns. At the intersection of the rows and columns, relationships (correlations) can be identified, which can then be used to create predictive models. Scatterplots are useful for visualizing the degree to which two data elements (also known as fields in a data table) move together.

Data can be analyzed to quantify the degree to which one variable moves in relation to another. This is done by creating correlation tables and determining the coefficient (more correctly, the “Pearson moment correlation coefficient,” known as the “r”), or relationship, for a number of variables. This correlation coefficient shows the degree to which variables change in relation to each other – i.e., to what extent does an increase/decrease in one result in an increase/decrease to the other. As a heuristic (i.e., rule of thumb), the strength of the relationship between variables can be interpreted as shown in Table 5.1. (Note that when r = 1.0, there is a perfect correlation and one variable moves exactly in harmony with the other. Correlations can be positive, when variables move in the same direction, or negative, when they move in opposite directions, in which case r = −1.0.)