Experiment as a special type of research
We have come to more precisely formulate the requirements that an experiment must meet as a special type of research procedure. The experiment is considered to be an ideal scientific method, since it presupposes strict control over all conditions of the experiment. This is most easily done in a specially equipped laboratory. A natural experiment is more difficult to conduct, and its results almost always leave room for alternative interpretations.
The overall structure of the experiment is as follows. That factor, the action of which is studied, is designated as an independent variable. In the example of the independent variable, the teaching method is the method. Most often it has two meanings, as in the case considered, where two methods of learning were compared – old and new. Accordingly, two groups of subjects are used: experimental and control.
The effect that the researcher observes is denoted as a dependent variable. In the example given, the progress of students after training was the dependent variable. It is assumed that for an objective evaluation of the learning effect, there is a sufficiently reliable test. Thus, the test acts as a measuring tool, and the results on it represent a numerical expression of the values of the dependent variable. The researcher is interested in the extent to which the values of the dependent variable differ in the experimental and control groups. To answer it, methods of mathematical statistics are used.
In addition to an independent and dependent variable, there is another group of factors that a researcher cannot pass by. These factors are designated as extraneous variables. From them the experimenter tries to rebuild, for they violate the purity of the experiment, obscure the true nature of the relationship between the independent and dependent variable. So, testing a new curriculum assumes that the experimental group differs from the control group only by the method of instruction being used.
All other moments had to be completely equalized. Otherwise, we cannot argue that the observed differences in the dependent variable (students’ successes) are due solely to the differences in the independent variable (the way of learning). In a natural setting, this is not always easy to do. Suppose a new teaching method is presented as extremely effective. This in itself can already cause a special interest of students, stimulate their attention and diligence. There is an additional factor – motivation, the effect of which is superimposed on the effect of learning as such. We say that the researcher was not able to properly monitor all the conditions for conducting the experiment.