Applied research is based on the following hierarchy of information (with knowledge the goal):

A descriptor is simply a descriptive statistic of one sort or another. This many students or this much money, are both common descriptors and your lowest level of data.

Alternatively an indicator is a calculated statistic with an implied value. An example of this would be the money spent on post-secondary education divided by the number of students. This ratio has the implicit message that how much money spent per student is important.

Descriptors have a face value that is incontrovertible. The square footage devoted to bathrooms in museums may not be widely useful but it is a real statistic.

Indicators on the other hand can be developed that produce meaningless statistics. An example of this could be the school instructional budgets over the square footage of school bathrooms. There is no reason why the two should be connected. However the ratio of janitorial costs to operating budgets or the ratio of sanitary facility expenditures to overall capital budgets may have some uses.

Thus an indicator is a calculated statistic based on two or more descriptive statistics. Additionally, to have meaning the statistic must be created with a purpose in mind.

Indicators should if possible be unambiguous as to what is desired. Is a low pupil/teacher ratio a sign of inefficiency or quality instruction?

A final step is determined by W.E. Deming who defined the difference between information and knowledge as the addition of the element of time. Thus if you can examine the created indicator over time then you can begin to make some inferences how a system is behaving both good and bad.

The extension of information into knowledge is also accomplished with the addition of spatial elements (e.g., is this region doing better than that region).

At the end of the day you know that one region is doing better or that your profits by a certain sector have declined over time. The next question (and step) is "Why?".