I just completed an interesting project with a researcher in Philadelphia. We were looking at political cultures and used correlation statistics to find certain links within the data. To be specific we used Ordinal data and Spearman's Rho.

There are generally four types of data: Nominal, which are usually categorical. These include male or female: red car or green car: things that can be neatly categorized.

Next we have ordinal data. Here, objects are ranked. First in class, second in line and so on. This is the type of data we chose to use. We took political cultures and ranked them under a pre-defined system.

The third type of data is interval data. This allows the ordering of data and have a definite space between each data point. One example is the Farenheit measurement for temperatures. Although moving for twenty to thirty Farenheit is the same amount of change as moving from fifty to sixty Farenheit, there is no definite zero point.

The last type of data, ratio data is distinct from interval because it has an exact zero point.

These data types can be studied in a number of ways. Spearman's Rho, the process we used, looks at the relationship between two categories. For example, if we wanted to look at how students were ranked in two classes and whether there was a relationship between those rankings, we could use Spearman's Rho. Generally, the numbers range from -1 to +1. A Spearman's Rho of 0.80 means that Y tends to increase as X tends to increase. A Spearman's Rho of -0.80 on the other hand means that Y tends to decrease as X increases. For example, if Y is the number of hours you have left at work and X is the number of hours that passed since you started working, you'll find a negative relationship: If Y=6 and X=2, you work an 8 hour day. If Y=4 and X=4, you still have an eight hour day.

Does this add up to anything? Well that's what significance tests are for!