Before we can measure a variable, we need to know exactly what it is. For example, if we wish to examine the relationship between grade point average and library use, we must be able to define "grade point average" and "library use."
Conceptual definitions define a concept by means of other concepts, making them difficult to measure. For example, a conceptual definition of a serial publication is that it is issued in successive parts at regular intervals and intended to be continued indefinitely.
An operational definition identifies one or more specific observable conditions or events and then tells the researcher how to measure that event. Typically, there are several operational definition possibilities for variables and values. The operation chosen will often have an immediate impact on the course of the research, especially the findings.
Operational definitions must be valid -- do they measure what they are supposed to measure? It is easy to develop definitions that are easily operational, but which lack meaning and appropriateness. The operational definition may involve the use of proxy variables. Here, the researcher is never completely certain that she is measuring the precise property they intend to measure.
Operational definitions must be reliable -- the results should be the same when done by different people or by one person at different times.
Here is an operational definition of a periodical. An item is a periodical if (1) the citation does not contain a publication place and (2) the citation contains either volume and listing numbers or a chronological designation that includes month or season. Thus, the researcher knows exactly what to look for when determining whether a bibliographic citation is or is not to a periodical publication. If there are difficulties in creating a good operational definition, that indicates a serious problem. For example, it will be difficult to study "negative book reviews" if you can not operationalize "negative" and "review."
By convention, some terms need not be defined. "Before" and "after" are good examples.
In survey research, the responder often must create and apply the operation if one is note included in the question, i.e. " are you satisfied with library service"? This would not be the case if a definition was included [and this may not be a good example]: are you satisfied {continue to return to the library and use collections and/or services} with library service"?
