# Appendix 4A Preliminary Tests of Demands for Forms of Religion

Our hypothesis in chapter 4 is that both full-price elements and a set of independent variables help explain the demand curve for a particular form of religion as well as religious forms themselves. In order to support this hypothesis we offer two preliminary empirical tests of whether the form of religion is related to a particular set of variables that proxy full prices and other factors determining the form of religion. Our tests include two types of animistic religions or beliefs—time-intensive, third-world tribal forms of animism and less time-intensive forms of New Age spiritualism.

Our tests use a method called regression analysis, a tool that may not be familiar to many readers. Basically, any problem will contain numerous elements and the "real-world" data collected by the economist or analyst will not easily give up its truths concerning cause and effect. That is why a favorite tool of the economist is regression analysis. Regression analysis is an econometric tool commonly used by economists to gauge a relationship between a variable, called a dependent variable, and one independent variable or a set of independent variables that move "independently'' and are associated with changes in the dependent variable. Consider, for example, the movement of tides. That movement may be associated with phases of the moon. But simply because we observe two variables moving together—correlation is the statistical term for that— the observation does not imply that one causes the other (e.g., that phases of the moon cause the movements of tides at the seashore). Correlation, in other words, does not necessarily imply causation. And also consider that many other variables may affect the course of tides—depth of the water, barometric pressure, wind speed, and so on. Economists use a tool called multiple regression in order to determine the manner (positive or negative) and the magnitude (how much) the dependent variable (e.g., tides) is affected by a set of independent variables (moon, water depth, and so on). Again, while correlation does not necessarily imply causation, the economic analyst can use fairly sophisticated statistical tools in order to gain confidence in the direction and magnitude of the effects of independent variable changes on the dependent variable of interest. In the following tests, we seek to find out how particular independent variables affect the incidence of animistic religions (our first dependent variable) and how another set of independent variables might explain the presence of psychics in forty-six U.S. states (our second dependent variable).