V. Knowledge Gained Through Actual Experience (Empirical Research)
The first credible source of information is what a person comes to know by direct personal experience. These empirical sources can be categorized into several general classes. These classes could be seen as lying on a rough continuum which varies from a relatively passive observation to the active manipulation of abstract variables in a completely controlled environment.
It’s important to note at this point that personal experience is not as simple as it seems. What about errors or misunderstandings? How are we to deal with dreams, psychotic episodes, and drug-induced experiences. Additionally what about events which cannot be seen, heard, tasted, or felt. The solution is to demand both integration within a theoretical network and consensual validation as well as some form of empirical support. We must have some mechanism whereby we can resolve disagreements of substance advanced by people of good conscience. These issues are the point of this manuscript.
If we wish to construct truthful systematic knowledge, then we will have to have procedures which will assure that we get exactly that.
Observational techniques take nature the way it comes. If knowledge were "food", observation is getting a meal by going to a restaurant, you get it the way it is served. Sometimes it's useful knowledge that you can use to better understand things, but sometimes it's mixed with confounds which make it impossible to consume.
The task of observational research is one of abstraction of commonalities. The researcher comes up with axis labels which show systematic and reliable functions. Correlational techniques. Identification of functional relationships. This tends to be describing events and measuring the capacity of the organism. Methods vary along a dimension of constraint or structure. Simply watching something in order to see what is going on is relatively unstructured, conceptually demanding, and holds the potential of being the most productive. While at the other extreme, a check sheet is very structured, conceptually easy, but often tells us little more than we already know. As with all research, the focus could be on the behavior and functional relations at face value only, or the intent could be to see the behavior as an instance of a more general process.
i. Naturalistic Observation
This is simply observing nature exactly as it occurs. The observer must hope that the data obtained are relevant and that the observation itself did not affect the behavior.
ii. Restricted Observation
Often we cannot gather information on every conceivable event. This category is the explicit recognition that many things are ignored.
(1) Restricted With Respect to Subject
A case study is an observation limited to a single individual.
(2) Restricted With Respect to Information Collected
A survey is a collection of a set of information (such as TV viewing habits) by gathering a limited subset of the potential data.
A check sheet is a collection of a precisely specified subset of variables in a controlled format. It is the most restricted but has the advantage of speed and that the desired information will always be included.
iii. Contrived Observation
Contrived observations alter the situation. In a sense experiments could be seen as extremely contrived observations. This category is under the heading “types of observation” because if only minor changes are made, it is called an observation. If major changes are made which allow a comparison, then it is called an experiment. Obviously however, it is a name for a position on a continuum.
(1) With Respect to Context or Events
(2) With Respect to Subjects
Experimental techniques do whatever is necessary to reveal what you need to know to better understand the world. Building on the previous metaphor, if knowledge were food, then experimentation is getting a meal by making it yourself. You have the power to optimize it in any way you want. You have complete freedom and are only limited by your own imagination and skill. However, if it doesn't reveal useful knowledge, then you have only yourself to blame. Experimental design gives you the ability to eliminate confounding and to establish causation. It is, therefore, the best way to gain knowledge. Keep in mind that the result may be that you discover that what you initially believed was wrong. Again, the focus can be on the behavior at face value or behavior as an illustration of a broader process. Gathering knowledge directly through the active manipulation of variables can range from simple to complex experiments. On the one hand, they can document the effects of some continuous series of manipulations such as dose effect curves or parameter documentation. At a different extreme, they could be "crucial" experiments which determine which of several theoretical accounts of behavior are correct. Alternatively, they could involve the deliberate manipulation of interesting variables simply to "see what happens."
A number of procedures have evolved which help extract knowledge from the environment while minimizing confounds and confusion.
i. Single Subject Designs
In these designs, the difference in an individual's behavior with and without the treatment is the dependent variable. The difference in what the individual receives is the independent variable.
x
x
x
x
(3) Schedules under Discriminative Control
x
x
In group designs, the difference in the average score of the two groups is the dependent measure. The difference in what the two groups receive is the independent variable. Note that the differences in the individuals contributes to the variability within each group.
(1) Quasi Experimental Designs
x
x
(2) Single Independent Variable
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
(c) Hierarchical (nested)
x
x
x
x
Sometimes knowledge is btained through experience which is not the result of typical “scientific” observation or experimentation procedures.
i. Technological By-Product
ii. Puttering Around / Accident / Serendipity
Date Last Reviewed: November 17, 2002