Steps in research work, and the place of statistical analysis
Relation of Statistical Analysis to Research. Statistical analysis is only a tool to be used by the investigator. The analyst must be a worker in some field, or in several; he cannot use his statistical training except in analyzing problems any more than a carpenter can use his skill without lumber and something to be made. Now that the routine of statistical analysis has been discussed, and the types of problems to which it may be applied have been surveyed, it is pertinent to ask just what are the steps in research work and just where and how does statistical analysis fit into the picture.
The research worker must have an adequate knowledge of the facts, technical and otherwise, of the field in which he is to work. This knowledge is usually insured by the situation that in most cases the worker is a biologist, an economist, a psychologist, or an agronomist, first, and then a statistician only secondarily or in addition. When his training has been primarily in mathematics or statistics, however, the statistician must acquaint himself thoroughly with the facts and theories of the field involved before he can expect to do significant and substantial work on applied problems.
Stating the Objective. If adequate acquaintance with the field is given, the first step in a particular research problem is setting up the objective of the project. The objective can best be stated in the form of a direct question, such as "Why did sales of European automobiles in the United States increase from 1946 to 1958?" The more exact and specific the question can be made, the more clearly is the field of the investigation defined. Stating the objective as a question has the important effect of clarifying the issue, and so insuring that the worker knows what he is really trying to find out. It has the further effect of instantly challenging the attention and of instinctively calling forth mental answers which aid in the next step of the research.