Chapter 2. BASICS OF EXPERIMENT PLANNING
If you want to experimentally test whether music radio programs help memorize French words, you can easily do this by repeating one of the experiments described in the previous chapter. Most likely, you will design your experiment along the lines of Jack Mozart. You will pre-determine both conditions of the independent variable, practice at the same time of day, and document each step of the experiment. Instead of four piano pieces, you could memorize four lists of words like this: listening to the radio, no radio, no radio, with radio. In other words, you can apply the same experimental design, which is Jack.
It is possible that you will understand some of the reasons for your own actions. But something will surely remain unclear, and above all - the sequence of conditions of the independent variable, i.e., the experimental scheme itself. This is not your fault, because you have not yet passed the experimental schemes. In this chapter, this shortcoming will be eliminated. Of course, you can conduct an experiment by simply imitating a model, but it is much better to understand what you are doing. There are no two identical experiments, and blind copying of an experimental scheme often leads to difficulties. For example, Yoko could use regular alternation of two conditions (juice varieties) in her experiment, as she did in the experiment with weavers (use or not use of headphones). But then she would probably know the name of the juice being tested, and "this is exactly what she tried to avoid by using a random sequence. Also, if you do not know the reasons for the various plans and schemes, it will be difficult for you to assess the quality of the experiments that you will read about And, as you remember, to teach you this is one of the main goals of our book.
In this chapter, we will compare those plans that
The experiments in Chapter 1 were built, with less successful plans for doing the same experiments. The model for their comparison will be a "flawless" experiment (which is practically impossible). An analysis of this here will allow us to consider the main ideas that guide us in creating and evaluating experiments. In the course of this analysis, we will introduce several new terms into our vocabulary. Finally, we will determine what is perfect and what is not in the three experimental schemes that were used in Chapter 1. And these schemes represent three ways of ordering, or three kinds of sequences of presentation of various conditions of the independent variable used in the experiment with one subject.
After studying the material in this chapter, you will be able to competently and not imitate someone else's experiment to design your own. At the end of the chapter, we will be asked questions on the following topics:
1. The degree of approximation of a real experiment to an impeccable one.
2. Factors violating the internal validity of the experiment.
3. Systematic and non-systematic sources of violation of internal validity.
4. Methods for increasing internal validity, methods of primary control and experimental schemes.
5. Some new terms from the dictionary of the experimenter.
JUST PLANS AND MORE SUCCESSFUL PLANS
Undoubtedly, the first condition for conducting an experiment is its organization, the existence of a plan. But not every plan can be considered successful. Suppose that the experiments described in chapter 1, carried out differently, according to the following plans.
1. In the first experiment, let the weaver wear headphones for 13 weeks, and then work without them for 13 weeks.
2. Suppose Yoko decided to use only two cans of each type of juice in her experiment, and the whole experiment took four days instead of 36.
3. Jack decided to apply the partial method of memorization to the first two pieces, and the holistic method to the next two.
4. Or, keeping the same sequence of methods, Jack chose short waltzes for the experiment, rather than the longer pieces he usually learned.
We feel quite clearly that, compared with the experiments previously described, all these plans are unsuccessful. And if we had comparison sample, then we could definitely say why exactly the original plans were better. The flawless experiment serves as such a model. In the next section, we will discuss it in detail and then see how it applies to evaluate our experiments.
PERFECT EXPERIMENT
We now have examples of successful and unsuccessfully designed experiments. Can a well-designed experiment be further improved? And is it possible to make the experiment absolutely flawless? The answer is: any experiment can be improved indefinitely, or - which is the same thing - a flawless experiment cannot be carried out. Real experiments improve yourself as you get closer to perfection.
If you want to experimentally test a hypothesis, you can do the experiment by simply imitating the model, but it is much better to understand what you are doing. There are no two identical experiments, and blind copying of the experimental design often leads to difficulties.
Undoubtedly, the first condition for conducting an experiment is its organization, the existence of a plan. But not every plan can be considered successful. It is quite clear that in comparison there are plans that are more successful and there are plans that are less successful or completely unsuccessful. Deciding to conduct an experiment, we are faced with the concept experimental scheme . Experimental design of the first sample of the study are three ways of ordering, or three kinds of sequences of presentation of different conditions of the independent variable, used in the experiment with one subject. The model for their comparison will be a "flawless" experiment as reference(practically impossible).
3.1. The concept of a "flawless" experiment
Any experiment can be improved indefinitely, but a flawless experiment cannot be carried out. Real experiments are being improved as we approach the perfect experiment, which can be presented in three forms: as an ideal experiment, an infinite experiment, an experiment of full compliance.
The Perfect Experiment
In an ideal experiment, only the independent variable is allowed to change (and, of course, the dependent variable, which takes on different values under different conditions). Everything else stays the same, and so only the independent variable affects the dependent variable. A perfect experiment is impossible. However, the idea itself is useful; it is this idea that guides the improvement of real experiments.
For example, in an ideal (impossible) experiment, the weaver would have to work with and without headphones at the same time! In this case, the difference in the values of the dependent variable would be due to only independent variable, the difference in its conditions. In other words, all incidental circumstances, all other potential variables would remain at the same constant level.
Endless Experiment
In order to average not only the variability of each of the states of the independent variable, but also possible fluctuations in the states of the subject himself, it is necessary to continue the experiment indefinitely. This is an endless experiment. It is not only impossible, but also meaningless. After all, the general meaning of the experiment is that, on the basis of limited amount of data to draw conclusions that have a wider application. However, this experiment also serves as a guiding idea.
The Infinite Experiment has drawbacks. The very fact of presenting one of the experimental conditions to the subjects may affect (during the study period) their work under another condition. Therefore, neither ideal nor endless experiments are completely flawless. Fortunately, they have not only various disadvantages, but also different benefits and can serve to evaluate real experiments, which are very far from a flawless experiment.
Full match experiment
If, in an unsuccessful version of the study, Jack Mozart memorized waltzes instead of sonatas, an experiment is needed to eliminate such shortcomings. full compliance. This experiment is also meaningless, since Jack would have to memorize same pieces that he will learn after him. But, once having learned the pieces, it is impossible to teach them after the end of the experiment.
All three kinds of flawless (almost) experiments are unrealistic. They are useful as "thought" experiments. They suggest what needs to be done to create an effective experiment. Perfect and endless experiments show how to avoid extraneous influences and thereby gain greater confidence that the experimental results really reflect the relationship between the independent and dependent variables. Experiment full compliance recalls the need to control other important experimental variables that we keep unchanged.
The concept of flawless experiment was introduced into psychology by Robert Gottsdanker. Psychologists view a flawless experiment as a model in which all requirements are embodied and nothing threatens a reliable conclusion. Such an impeccable research model is unattainable in reality. However, this concept contributes to the development and improvement of experimental methods, avoiding possible errors in the experiment.
R. Gottsdanker defined the criteria for an ideal experiment: it must be ideal, infinite and an experiment of full compliance.
In an ideal experiment, only the independent and dependent variables change, there is no influence of external or additional variables on them. A variation of his pure experiment, in which the researcher operates with only one independent variable and its fully refined conditions.
In an infinite experiment, the number of trials and subjects make it possible to cover all sources of variable variability. Such an experiment can continue indefinitely, since there is the possibility of an unknown factor acting. In order to find out all the side effects that distort the effect of the independent variable, the researcher must continue research indefinitely both in time and in number of attempts, since there is always the possibility that in the next trial something can violate the ideality of the experiment.
In a full fit experiment, the additional variables must fully match the reality counterparts of those variables. The experimental situation is completely identical to the actual real situation.
The model of an ideal experiment is an unattainable ideal to which one must strive. The closer the real experiment of this model, the better it is.
D. Campbell offers the following criteria for a good experiment:
1. A good experiment determines a clear temporal sequence: the cause has in time to precede the effect.
2. The influence or influences must be statistically related to each other. If a possible reason and the effect is unrelated (no covariance), then one phenomenon cannot be the cause of the other.
3. There should be no alternative plausible explanations for the causes of the effect, that is, it is necessary to exclude the influence of side variables or at least control them.
4. It is important to correctly identify the cause and effect in terms and concepts.
The use of experiment in various areas of psychology has its own specifics. Psychological experiments are aimed at investigating certain problems, and in various branches of psychology, their specific problems occupy a central place. The main problems are determined by the researcher's curiosity, his versatility, imagination, and also by the fact that there are opportunities for the implementation of experimental designs. For example, the eminent psychologist John Watson, in his book "Behaviorism" (1924), expressing conviction in his ideas, suggested the following: "Entrust me with a dozen healthy normal children and let me bring them up as I see fit; I guarantee that by choosing each of them randomly, I will make him what I think: a doctor, a lawyer, an artist, a businessman, even a beggar or a thief, regardless of the data, abilities, vocation or race of his ancestors. The design of such an experiment may be astonishing, but of course such a proposal goes too far beyond what most of us consider acceptable.
Professional requirements for psychologists do not clearly define the nature and content of their research activities, teaching and consulting. In any environment, not just at a university or scientific institution, there is an opportunity for research. The experiment is possible in a school, office, state or commercial organization, in everyday life, on vacation.
psychology test experiment
Man and the features of his personality have been the object of interest and study of the great minds of mankind for more than one century. And from the very beginning of the development of psychological science to the present day, people have managed to develop and significantly improve their skills in this difficult but exciting business. Therefore, now, in order to obtain reliable data in the study of the characteristics of the human psyche and his personality, people use a large number of various methods and methods of research in psychology. And one of the methods that have gained the greatest popularity and proven themselves from the most practical side is a psychological experiment.
An experiment in psychology is a certain experience that is carried out under special conditions in order to obtain psychological data through the intervention of a researcher in the process of the subject's activity. Both a specialist scientist and a simple layman can act as a researcher during the experiment.
The main characteristics and features of the experiment are:
- · Ability to change any variable and create new conditions to identify new patterns;
- · Possibility to choose a reference point;
- Possibility of repeated holding;
- · Possibility to include in the experiment other methods of psychological research: test, survey, observation and others.
There are many views on the differentiation of experimental techniques and a significant number of terms denoting them. If we summarize the results in this area, then the totality of the main varieties of the experiment can be represented as follows:
I. According to the validity and completeness of the procedure
- 1. Real (specific). A real (specific) experiment is an experiment carried out in reality under specific experimental conditions. It is real research that provides factual material used both for practical and theoretical purposes. The results of the experiment are valid for specific conditions and populations. Their transfer to broader conditions is probabilistic.
- 2. Thought (abstract): A thought experiment is an imaginary experience that cannot be done in reality. Sometimes this category also includes mental manipulations regarding the organization and conduct of a planned real experiment in the future. But such preliminary "playing" in the mind real experience- in fact, its mandatory attribute, implemented on preparatory stages research (statement of the problem, hypotheses, planning).
- a) ideal;
- b) endless;
- c) perfect.
An ideal experiment is an experiment in which the dependent variable is not influenced by any other than one independent variable. In reality, it is impossible to exclude the additional influences of many attendant factors. Therefore, the ideal experiment is not really feasible. In practice, the approximation of real experience to the ideal is realized by controlling additional variables, which is described in the description of the experimental procedure.
An infinite experiment is an experiment that covers all possible experimental situations for the entire study population (general population). In reality, the set of such situations is limitless due to the huge, and often unknown, size of the general population and the countless number of factors acting on the subject. Accounting for all this infinite set of situations is feasible only in the imagination of the researcher. Due to its infinity (in variety and in time), such an experiment was called the infinite. The practical meaninglessness of an infinite experiment is in contradiction with one of the main ideas of empirical research - the transfer of results obtained on a limited sample to the entire population. It is needed only as a theoretical model.
Flawless is an experiment that combines the features of both ideal and endless experiments. As a standard for an exhaustive experiment, it makes it possible to assess the completeness and, accordingly, the shortcomings of a specific real experience.
II. According to the purpose of the experiment
1. Research.
A research experiment is an experience aimed at obtaining new knowledge about the object and subject of study. It is with this type of experiment that the concept of "scientific experiment" is usually associated, since the main goal of science is the knowledge of the unknown. While the other two types of goal-criterion experiment are predominantly applied in nature, the research experiment mainly performs a search function.
2. Diagnostic (exploratory).
A diagnostic (exploratory) experiment is an experiment-task performed by the subject in order to detect or measure any qualities in him. These experiments do not give new knowledge about the subject of research (personality quality). In fact, this is testing.
3. Demo.
A demonstration experiment is an illustrative experience that accompanies educational or recreational activities. The immediate goal of such experiments is to familiarize the audience with either the appropriate experimental method or the effect obtained in the experiment. Demonstration experiments are most widely used in educational practice. With their help, students master research and diagnostic techniques. Often an additional goal is set - to interest students in the relevant field of knowledge.
III. By level of research
1. Preliminary (reconnaissance)
A preliminary (reconnaissance) experiment is an experiment carried out to clarify the problem and adequately orient it. With its help, little-known situations are probed, hypotheses are refined, questions are identified and formulated for further research. Studies of this reconnaissance nature are often called pilot studies. On the basis of the data obtained in the preliminary experiments, questions are being addressed about the need and possibilities for further research in this area and the organization of the main experiments.
2. Main
The main experiment is a full-scale empirical study carried out in order to obtain new scientific data on the problem of interest to the experimenter. The result obtained as a result is used both for theoretical and applied purposes. The main experiment may be preceded by preliminary ones of both reconnaissance and fact-finding nature.
3. Control.
A control experiment is an experiment whose results are compared with the results of the main experiment. The need for control may arise for various reasons. For example: 1) errors were found in the conduct of the main experiments; 2) doubts about the accuracy of the procedure; 3) doubts about the adequacy of the procedure to the hypothesis; 4) the emergence of new scientific data that contradict those obtained earlier; 5) the desire for additional evidence of the validity of the hypothesis accepted in the main experiment and its transformation into a theory; 6) the desire to refute the existing hypotheses or theories. It is clear that the control experiments should not be inferior to the main ones in terms of accuracy and reliability.
IV. By type of impact on the subject
1. Internal.
An internal experiment is a real experiment, where mental phenomena are caused or changed directly by the volitional effort of the subject, and not by influence from the outside world. Experimentation is carried out in the subjective space of a person, where he plays the role of both the experimenter and the subject. The internal impact always includes an independent variable, and ideally it should be limited to only it. This brings the inner experiment closer to the mental ideal.
2. External.
An external experiment is a common experimental way of studying mental phenomena, when their appearance or change is achieved due to external influences on the sense organs of the subject.
V. According to the degree of intervention of the experimenters, the vital activity of the subject (according to the type of experimental situation)
A. Classic grouping
1. Laboratory (artificial).
A laboratory (artificial) experiment is an experiment carried out under artificially created conditions that allow strictly dosing stimulation (independent variables) and controlling other effects on the subject (additional variables), as well as accurately registering his responses, including dependent variables. The subject is aware of his role in the experiment, but his overall intention is usually not known to him.
2. Natural (field).
Natural (field) experiment - an experiment carried out in the usual conditions for the test subject with a minimum of interference in his life by the experimenter. The presentation of an independent variable is, as it were, "woven" in a natural way into the normal course of his activity. Depending on the type of activity performed and the corresponding situation, types of natural experiment are also distinguished: in conditions of communication, labor, play, educational, military activities, in everyday life and leisure. A specific type of this type of experiment is an investigative experiment, in which the artificiality of the procedure is combined with the naturalness of the conditions for illegal actions.
3. Formative.
A formative experiment is a method of active influence on the subject, contributing to his mental development and personal growth. The main areas of application of this method are pedagogy, age (primarily children's) and educational psychology. The active influence of the experimenter consists mainly in creating special conditions and situations that, firstly, initiate the emergence of certain mental functions and, secondly, allow them to be purposefully changed and formed. The first is characteristic of both laboratory and natural experiments. The second is the specificity of the considered form of experiment. The formation of the psyche and personality traits is a long process. Therefore, the formative experiment is usually carried out for a long time. And in this respect it can be classified as a longitudinal study.
B. Extraordinary grouping:
1. An experiment that duplicates reality.
Experiments that duplicate reality are experiments that simulate specific real life situations, the results of which have a low level of generalization. Their conclusions are applicable to specific people in the conditions of specific activities, which is why they are also called full compliance experiments. These experiments are purely practical purposes. This type of experiment is close to the natural type in terms of classical grouping.
2. An experiment that improves reality.
Reality-enhancing experiments are experiments in which only some of the variables to be studied are changed. The rest of the variables are stable. This type is similar to the laboratory experiment according to the generally accepted classification.
VI. If possible, the influence of the experimenter on the independent variable
1. Provoked experiment.
A provoked experiment is an experiment in which the experimenter himself acts on the independent variable. Changes in NP can be both quantitative and qualitative. And then the results observed by the experimenter (in the form of reactions of the subject) are, as it were, provoked by him. Obviously, the vast majority of experimental studies refer to this species. P. Fress, not without reason, calls this type of experiment "classical".
2. Experiment referred to.
A referenced experiment is an experiment in which a change in the independent variable is carried out without the intervention of the experimenter. These include personality changes, brain damage, cultural differences, and so on. According to P. Fress, these cases are very valuable, “since the experimenter cannot introduce variables whose action would be slow (education system), and does not have the right to experiment on a person if his experiment can cause serious and irreversible physiological or psychological disorders » . There may be cases where an experiment on some variables is provoked, but on others it is referred to.
VII. By the number of independent variables
1. One-factor (two-dimensional).
A one-factor (two-dimensional) experiment is an experiment with one independent and one dependent variable. Since there is only one factor influencing the answers of the subject, the experience is called one-factor or one-level experience. And since there are two measured quantities - NP and ZP, the experiment is called two-dimensional or bivalent. The selection of only two variables allows us to study the mental phenomenon in a "pure" form. The implementation of this version of the study is carried out using the procedures described above for controlling additional variables and presenting an independent variable.
2. Multifactorial (multidimensional).
A multivariate (multivariate) experiment is an experiment with several independent and usually one dependent variable. The presence of several dependent variables is not excluded, but this case is still extremely rare in psychological research. Although, apparently, the future belongs to him, since real mental phenomena always represent the most complex system of many interacting factors. The term “bad” that is common in science is applicable to them. organized systems”, which just emphasizes the multiplicity of determination of their manifestation
VIII. By number of test subjects
1. Individual.
An individual experiment is an experiment with one subject.
2. Group.
Experience with several subjects at the same time. Their mutual influences can be both significant and insignificant, they can be taken into account by the experimenter or not taken into account. If the mutual influence of the subjects on each other is due not only to co-presence, but also joint activities, it is possible to speak of a collective experiment.
IX. By the method of identifying relationships between variables (by the procedure for varying the experimental situation)
1. Intraprocedural (inside).
An intraprocedural experiment (lat. intra - inside) is an experiment in which all experimental situations (in fact, all values of the independent variable) are presented to the same contingent of subjects. If the subject is alone, i.e. individual experience is carried out, then one speaks of an intra-individual experiment. Comparison of the responses of this subject received in different situations(for different values of NP), and makes it possible to identify relationships between variables. This option is especially convenient for quantitative changes in NP to determine functional dependencies.
2. Interprocedural (between).
Interprocedural experiment (lat. inter - between) - an experiment in which different contingents of subjects are presented with the same experimental situations. Work with each individual contingent is carried out either in different places, or at different times, or by different experimenters, but according to identical programs. The main goal of such experiments is to clarify individual or intergroup differences. Naturally, the former are revealed in a series of individual experiments, and the latter in group experiments. And then in the first case one speaks of an interindividual experiment, and in the second case one speaks of an intergroup, or more often an intergroup experiment.
3. Cross procedural (intersection).
A cross-procedural experiment (English cross - to cross) is an experiment in which different contingents of subjects are presented with different situations. If the subjects work alone, then we are talking about a cross-individual experiment. If each situation corresponds to a certain group of subjects, then this is a cross-group experiment, which is sometimes called an intergroup experiment, which is a terminological inaccuracy. Intergroup is synonymous with inter-, not cross-group experiment. This inaccuracy stems either from inadequate translation of foreign sources, or from a careless attitude to terminology.
X. By type of change in the independent variable
1. Quantitative.
A quantitative experiment is an experiment in which the independent variable can decrease or increase. The range of its possible values is a continuum, i.e. continuous sequence of values. These values, as a rule, can be expressed numerically, since the NP has units of measurement. Depending on the nature of the NP, its quantitative representation can be carried out in various ways. For example, time interval (duration), dosage, weight, concentration, number of elements. These are physical indicators. The quantitative expression of NP can also be realized through psychological indicators: both psychophysical and psychometric.
2. Quality.
A qualitative experiment is an experiment in which the independent variable has no quantitative variation. Its meanings appear only as various qualitative modifications. Examples: gender differences in populations, modality differences in signals, etc. The limiting case of qualitative representation of NP is its presence or absence. For example: the presence (absence) of interference.