Lecture
1. Control by means of experimental designs.
Planning an experiment is organizing its conduct in accordance with the anticipated experimental model, while mentally comparing it with a flawless ideal of its realization, in order to obtain reliable data.
Planning presupposes that in the experimental hypothesis the independent and dependent variables have been made concrete and, at the methodological level, operationalized.
A distinction is drawn between formal and substantive planning.
Formal planning is possible in those studies that use the principle traditional for the natural sciences: the variables (the independent variable and additional variables) are represented and can be manipulated independently of one another. If there is more than one manipulated variable, we adopt the principle of isolated conditions.
The experimental plan (design) includes a specification of the scheme for presenting the conditions of the independent variable to different groups of subjects, or of the sequence of levels of the independent variable presented to a single subject, as well as the number of trials or subjects (n). It is also a plan for recording the dependent variable.
To achieve validity, various means of experimental control are used.
This involves ways of setting the levels of the independent variable (functional control of the independent variable, management of the experimental factors) and control of all other variables as sources of competing explanations, i.e. factors that may threaten a valid inference about the effect of the independent variable on the dependent variable. The principal (but not the only) component of experimental control is the development of experimental schemes or plans.
The basic experimental plans as a means of functional control
Experimental plans make it possible to control the influence of extraneous variables and to ensure the possibility of a reliable inference about the causal relationship between the independent and dependent variables.
The classic one is the plan with an experimental and a control group, in which the influence of the independent variable is exerted only in the experimental group, and the results are compared with the control group. Such a plan makes it possible to assess the effect of the experimental intervention and to rule out some alternative explanations.
Between-group plans are widely used, in which different levels of the independent variable are presented to different groups of subjects. Their advantage is the absence of the influence of the sequence of presenting the conditions; however, a need arises to control individual differences between the participants.
In within-group plans, the same subject goes through all the experimental conditions. This increases the sensitivity of the study, since each participant serves as his own control. At the same time, control of the effects of order, practice, fatigue, and transfer is required, for which randomization and counterbalancing are used.
To increase internal validity, factorial plans are also used, which make it possible to investigate simultaneously the influence of several independent variables and their interaction. They provide a more complete picture of the mechanisms of the phenomenon being studied but require a more complex organization of the experiment and statistical processing of the data.
An important element of functional control is randomization — the random assignment of subjects to groups or the random presentation of experimental conditions. It helps to equalize the influence of uncontrolled factors and reduces the likelihood of systematic errors.
Thus, the experimental plan serves as the principal instrument of functional control in a psychological experiment. The choice of a particular research design is determined by the objectives of the experiment, the features of the phenomenon being studied, and the requirements for the internal and external validity of the results obtained.
Comments