Menu
homeInternet

Types of variables in a psychological experiment. Types of Variables in Psychological Research

Verification and falsification of hypotheses

Verification(from the Latin verus - true and facere - to do) is the procedure for establishing the truth of certain judgments, confirming theoretical knowledge by listing the entire class of empirical referents or objects that are covered by a given concept or a given hypothesis. To establish the truth of any statement - for example, that all ravens are black - it is necessary to conduct an observation, a survey, an experiment. In our case, we will have to travel all over Africa, Asia, Europe and Australia. The exception is the Arctic and Antarctica, where crows are not found. If, after examining all the crows, they turn out to be black and not a single white one is found, then your statement will be verified, i.e. its truth will be confirmed by scientific means. You can take the easy way - go around the places that are accessible to you and make sure that there are only black crows around. In this case, it cannot be said that all crows are black. We will have to be content with a more modest judgment - for example, “some crows are black.” The value of such a judgment is almost zero, since even without any inspection it is clear that some crows are black.

The principle of verifiability assumes that a concept or proposition has meaning (meaning) only if it is empirically verifiable. Unverifiable, and therefore meaningless, statements that all women (men) cheat, or that culture is developing progressively.

Falsification - in ordinary life this means distortion of the truth.
If verification is confirmation of the truth, then falsification is its distortion.
Falsification involves a scientist, instead of searching for all the examples that confirm a hypothesis, to find one single case that disproves it. If you claim that all crows are black, all you have to do is find one white crow and prove its exclusivity - then you will have the right to say that With rare exceptions, all ravens are black.
It turns out that disproving a hypothesis is much easier and more economical than confirming it. If you claim that any social group has its own subculture, then you must either count all large social groups (pensioners, youth, national minorities, professional groups, etc.) and prove that each has its own subculture, or find among them is one that does not have its own subculture, and to say that there is a single large social group that can be called an exception to the rule. In the first case we are talking about verification, in the second - about falsification.



Verification arose earlier than falsification - in the first half of the 20th century. -and originated in the depths of the Vienna Circle of Positivists; falsification invented by K. Popper in the second half of the 20th century. Popper warned: do not assume that science gives the final truth - it can only give partial truth. If a theory cannot be refuted, then, according to Popper, it stands outside science. Real science should not be afraid of refutations: rational criticism and constant correction with facts is the essence of scientific knowledge.

Design for controlling experimenter effects in psychological research.

Psychological experiment- this is a joint activity of the subject and the experimenter, which is organized by the experimenter and aimed at studying the characteristics of the psyche of the subjects. In this joint activity, both parties have their own motives and goals for participating in the experiment, and these motives and goals can become the causes of artifacts in the experiment. Under the artifacts It is customary to understand the results of the study of objects, which sometimes arise as a result of the influence of the research conditions themselves on the objects (which the experimenter, for some reason, is not able to take into account, and therefore is inclined to consider the result as the result of the influences he organized, i.e., to accept the artifact as a fact). The root "arte" means "artificial", that is, an artifact is an “artificially created fact”. The researcher may either not be aware of the reason for this artificiality - and then he accepts the artifact as a fact, or he may be aware - and then he looks for the cause of the artifact in order to turn it into a fact. In a classic natural science experiment, the ideal experimenter would be an automaton. But since this is impossible to do in psychology, the psychologist must take into account the fact that any experimenter, including himself, makes mistakes, i.e. involuntary deviations from the ideal experiment. Some of them are correctable, but there are stable tendencies in the experimenter’s behavior, which are a consequence of unconscious mental regulation of behavior. The fact is that the researcher - as a living person - is interested in confirming (or refuting) his hypothesis, so he can unconsciously introduce distortions into the course of the experiment and the interpretation of the data, ensuring that the subject “works according to the hypothesis.” Such actions by the experimenter are a source of artifacts. American psychologist Rosenthal called this phenomenon " Pygmalion effect».

A researcher interested in confirming a theory spontaneously acts so that it is confirmed. You can control this effect. To do this, experimental assistants who do not know its goals and hypotheses should be involved in the research. Full control is re-checking the results by other researchers who are critical of the hypothesis of the author of the experiment. However, even in this case, we are not guaranteed against artifacts - the controllers are the same sinful people as the author of the experiment.

All experimenters dream of the “ideal subject.” The “ideal subject” should have a set of appropriate psychological qualities: to be obedient, smart, eager to cooperate with the experimenter, efficient, friendly, non-aggressive and devoid of negativism. From a socio-psychological point of view, the model of the “ideal subject” fully corresponds to the model of the ideal subordinate or ideal student. An intelligent experimenter understands that this dream is impossible. However, if the behavior of the subject in the experiment deviates from the expectations of the researcher, he may show hostility or irritation towards the subject.

The experimenter's expectations can lead him to unconscious actions that modify the behavior of the subject. Rosenthal, the most famous specialist on the problem of the influence of the researcher’s personality on the course of the study, found that a significant influence of the experimenter on the result of the experiment was revealed: in experiments with learning, in diagnosing abilities, in psychophysical experiments, in determining reaction time, and conducting projective tests (Rorschach test) , in laboratory studies of labor activity, in the study of social perception.

How are the experimenter's expectations conveyed to the subject?

Since the source of influence is unconscious attitudes, they manifest themselves in the parameters of the experimenter’s behavior, which are regulated unconsciously. This is primarily facial expressions and pantomimes (head nods, smiles, etc.). Secondly, “paralinguistic” speech methods of influencing the subject play an important role, namely: intonation when reading instructions, emotional tone, expression, etc. In experiments on animals, the experimenter may unconsciously change the way they handle them.

The influence of the experimenter is especially strong before the experiment: during the recruitment of subjects, the first conversation, and reading the instructions. During the experiment, the attention shown by the experimenter to the actions of the subject is of great importance. According to experimental studies, this attention increases the productivity of the subject. Thus, the researcher creates the subject’s primary attitude towards the experiment and forms an attitude towards himself.

Planning to control “subject effects” in a psychological experiment.

Errors related to subject motivation: Some of the most well-known causes of this type of artifact are the placebo effect, the Hawthorne effect, the audience effect, the façade effect, and the desire for positive self-representation.

Placebo effect: “A placebo is a neutral substance used in experiments as a control.” The effect is based on the mechanisms of suggestion and self-hypnosis.

Hawthorne effect was first described during socio-psychological studies in factories in the town of Hawthorne. Involvement in participation in the experiment, which was conducted by psychologists, was regarded by the subjects as a manifestation of attention to them personally. As a result, the study participants behaved as the experimenters expected them to.

Audience effect(social facilitation = reinforcement): the presence of any external observer, in particular the experimenter and assistant, changes the behavior of the subjects. The following dependencies were installed: 1. The more competent and significant the observer, the more significant the effect of his presence. 2. The more difficult the task, the greater the impact. 3 . Poorly mastered skills deteriorate in the presence of a significant observer, while the productivity of well-mastered activities increases. 4. Competition, joint activity, and an increase in the number of observers enhance the effect (both positive and negative trends). 5 . “Anxious” subjects experience greater difficulties when performing complex and new intellectual tasks than emotionally stable individuals. 6. The effect of the “audience effect” is well described by law Yerkes – Dodson (dependence of the best results on the average intensity of motivation. There is a certain limit beyond which a further increase in motivation leads to worse results.)

Façade effect(desire for social approval) - desire to look better (the subject gives answers that, in his opinion, are more highly rated by the experimenter, cooperates with the experimenter).

Striving for positive self-representation those. the desire of the subject to look as good as possible in his own eyes, to “not lose face” in front of himself in an unusual and unnatural experimental situation. The motive of self-representation is very difficult to control, since the conditions in which it manifests itself and the direction of its influence on the experimental results are not defined. The motive of self-representation interacts with the motive of social approval: subjects especially strive to prove themselves “in the best way” when the experimenter cannot directly catch them in a lie. For example, if subjects are asked to rate their intelligence, it is particularly inflated when the experimenter does not intend to “test” their intelligence. If subjects know that after subjectively assessing their intelligence they will have to take an intelligence test, they rate it significantly lower.

To control the influence of the subject’s motivation and the effects of communication on the results of the experiment, there are a number of special methodological techniques. Let's list them and give a description of each.

1. " Double-blind or placebo-blind. It is a combination of two “blind” experiments: in one of them the experimenter is “blind”, in the other the subject is “blind”. In both cases, identical control and experimental groups are selected, the experimental group receives the experimental effect, the control group receives “zero” (placebo).

But in the first case, the experimenter does not know which group receives “zero” influence and which is subject to real manipulation, and this eliminates the influence of his motivations on the subjects. There are modifications to this design, for example, the experiment is carried out not by the experimenter himself, but by a invited assistant, who is not told which of the groups is actually being exposed.

An experiment in which the subject remains “blind” is an implementation of the placebo idea: the subject does not know which group - experimental or control - he is a member of, and therefore does not know what effect is being exerted on him, real or "zero" and, therefore, cannot behave “correctly.” Combining these two experimental designs gives “ double blind experience”, which allows you to eliminate both the effect of the subject’s expectations, that is, the Hawthorne effect, and the experimenter’s motivation, that is, it controls the Pygmalion effect, which will be discussed below.

Psychopharmacologist Beecher used this experimental design to study the effect of morphine on pain sensitivity. Using a placebo-blind design, he was unable to distinguish the data from the control group from the data from the experimental group. When he carried out the experiment in the traditional way, he obtained the classic different curves.

2. " Method of deception." The experimenter comes up with a false goal and hypothesis of the study, independent (orthogonal) from the main ones. A fictitious goal and hypothesis—a “cognitive placebo”—are presented to the subjects. A variant of the “deception method” is simply hiding the true goals and hypothesis of the experiment, but then this variable becomes uncontrollable, so “cognitive placebo” is preferable.

3. Hidden experiment method (another version of the “deception method”): the subject is unaware of his participation in the study as a subject. The use of this method raises ethical problems due to the complete control of the subject, who is the object of manipulation. This model is most often used in child psychology, developmental psychology, and educational psychology. In these cases, the problem of manipulation is less acute, since children are controlled by adults. However, it is necessary to obtain the consent of the parents or persons caring for the child for such a study. The main difficulty in carrying out such an experiment is taking into account uncontrolled variables, since this experiment can only be natural: in fact, it is a modification of the “natural experiment” of A.F. Lazursky.

4. Method of independent measurement of dependent parameters . Difficult to implement in practice, because... the experiment is carried out with the subject according to the usual plan, but the effect of the influence is measured not during the experiment, but outside it, for example, when monitoring the results of the educational or work activities of the former subject.

5. Monitoring the subject’s perception of the situation using a post-experimental interview. Unfortunately, the data obtained during the post-experimental survey only allows us to reject unsuccessful samples or take this information into account when interpreting the results of the experiment, when nothing can be corrected.

Types of Variables in Psychological Research

An experiment differs from observation by intervention in the situation on the part of the researcher, who carries out targeted manipulation of factors and registration of corresponding changes in the behavior of the object being studied. The factors involved in an experimental study are called variables.

They are divided into two types: INDEPENDENT variable and DEPENDENT variable. The variable that is manipulated or changed is called an independent variable.

An independent variable is a condition that an experimenter systematically changes to evaluate its effect on another variable. A variable that is expected to change in response to changes in the independent variable is called a dependent variable.

In other words, an experiment is a research method in which the researcher studies the effect of one class of variables (independent variables) on another class of variables (dependent variables). This assumes that the dependent variable should change as a function of changes in the independent variable. Measured changes in the dependent variable are considered to be "dependent" on the manipulation of the independent variable.

The experimenter's goal is to keep all conditions constant except one - the independent variable.

In a simplified example, the independent variable can be considered as a certain relevant stimulus (St(r)), the strength of which is varied by the experimenter, while the dependent variable is the reaction (R) of the subject, his psyche (P) to the influence of this relevant stimulus.

However, as a rule, the desired stability of all conditions, except for the independent variable, is unattainable in a psychological experiment, since almost always, in addition to these two variables, there are also additional variables, systematic irrelevant stimuli (St(1)) and random stimuli (St(2) ), leading respectively to systematic and random

Therefore, in an experiment, three types of variables can be distinguished:

1. Independent variable

2. Dependent variable

3. Additional variables (or external variables)

The main sources of spillover variability include: 1) variations in different aspects of the situation in which the experiment is conducted;

variations associated with the subjects (subjects and investigators) who take part in the experiment;

variations in the spatiotemporal sequence of conditions, events, responses that are part of the experimental procedure. Although these classes of possible confounding variables are not as clearly defined as they appear upon first reading the proposed classification criteria, they may be useful for framing the subsequent discussion of the kinds of factors that typically complicate the interpretation of experimental results. In table 4.1, constructed by analogy with the tables most often offered in textbooks on “Experimental Psychology,” presents the three above-mentioned groups of sources of secondary variability and the nature of the problems that may arise when selecting methods for their control.

Table 4.1. SIDE VARIABLES: SOURCES OF VARIABILITY AND THEIR CHARACTERISTICS

Sources

Nature of variability

Situational

side

variables

In parallel with different levels of the independent variable, the following can be said:

Environmental conditions: temperature, noise level, etc.

Aspects of experimental tasks: materials, nature of their presentation.

Formulation of instructions offered to the subject.

Experimenters' attitudes toward the results of each experimental condition.

Incidental variables associated with subjects who participate in the experiment as research subjects

The characteristics of the subjects serving as test subjects may vary.

Extraneous variables associated with the spatiotemporal sequence of experimental events

The spatial and temporal sequence of presentation of the experimental intervention may vary.

Final effects from previous experimental exposures may persist.

Effects may occur due to changes in performance on experimental repetitive tasks due to experience, habituation, or fatigue.

Let us consider each of the groups of secondary variables in more detail.

Situational spillover variables

As indicated in table. 4.1, situational variables reflect spatial and temporal aspects of the task, instructions, and the experimenter's attitude toward the expected outcome, which can change in parallel with changes in the independent variable.

Environmental Variables. The environmental conditions in which the experiment is conducted may change at different stages, which leads to undesirable changes in the quality of experimental tasks. These include the lighting of the room where the experiment takes place, air temperature and humidity, noise levels, etc. If these aspects of the environment are not controlled, they can affect the results of experimental tasks due to changes in motivation or as distractions.

Variables associated with experimental tasks. Any aspect of the experimental task, including its design and method of presentation, that is not specifically designed as a component of the independent variable may have an unpredictable effect on the experimental results and become a covariate. For example, undesirable changes in the performance of experimental tasks may arise due to inadequate use of stimulus material or presentation of tasks in a way that increases the likelihood of the emergence of unanticipated performance patterns in the subjects being studied.

The textbook “Experimental Psychology” by American authors Anderson and Borkowski examines an example of a hypothetical study, the purpose of which was to establish whether the rate at which children develop the ability to distinguish objects depends on the relative value of the latter as potential objects of the child’s attention. In this study, the experimental conditions manipulated may have been two sets of objects that differed from each other in terms of their likelihood of attracting the child's attention to them. The dependent variable could be the speed with which the child differentiates objects that are more likely to attract her attention from objects that are less likely to attract her attention. The problem immediately arises of determining which objects can claim more or less attention from the child (that is, seem more attractive to her). This is necessary both in order to select appropriate objects that can be used in the experiment as “stimulus material”, and in order to be confident in the correct disclosure of the content of the concept


“object of the child’s attention,” which should correspond to existing ideas.

One possible criterion for identifying a potential object of a child's attention might be “object complexity,” which can be operationalized in terms of the number of angles of 60° or more that can be distinguished in the object's configuration. When selecting stimulus material, the experimenter must pay attention to the fact that the two sets of objects differ only in one criterion - the number of angles of 60°. Having decided on what basis objects should differ, the researcher needs to make sure that they do not differ in other characteristics - color, size, meaningfulness, etc. As we see, the task of the researcher, developing and conducting psychological experiments, is really difficult: for him Many aspects of experimental problems must be taken into account to prevent the emergence of extraneous variables.

There are numerous examples of real-life studies in which secondary variables in a psychological experiment can arise due to unsuccessful selection of experimental tasks. For example, in studies related to the recognition of differences in the frequency characteristics of acoustic signals, it is necessary to maintain a constant level of their intensity.

The choice of method for presenting experimental problems can also become a source of collateral variables. For example, if stimuli A and B are always presented in such a way that A is on the case and B is on the left and stimulus B is such that it is reinforced, then the question arises: the subject reacts differently to these two stimuli because one of them is reinforced , and the other does not, due to the fact that they appear in different positions.

Corollary variables associated with the instructions provided to the subjects.

The main problem in formulating instructions is to do so in a way that somewhat disguises the true purpose of the experiment. Sometimes instructions are a composite independent variable. Consider a hypothetical study in which instructions are used to manipulate perceptual set. Imagine that the stimulus material for the experiment is drawings, the perception of which is based on the distinction between figure and background (such as a face-vase). In such an experiment, the perceptual set can be manipulated by asking subjects in one of the groups to see a face in the drawing, and in the other group not to give indicative instructions. Let us assume that the effect of the various instructions is to orient the subjects of the experimental group to see in the drawing exactly the figure for which they are given an explicit instruction to perceive. However, the question arises: what is actually the independent variable in an experiment? Different instructions, internal installation initiated by instructions, or both? Some researchers believe that the real manipulated variable in this case is the different settings initiated by different instructions. But there is reason to believe that the independent variable in such an experiment includes two components: 1) a specifically oriented instruction and 2) an internal attitude that is assessed using instructions. So, instructions can be considered like any other stimulus materials with the help of which independent variables are formed.

Typically, the same sentence can be perceived differently by different subjects. Even changing the order of one or two words in an instruction can completely change the meaning of an entire phrase. This can be avoided if the following caveats are followed: 1) present the same stimulus material several times; 2) invite the subject to reformulate the instructions “in their own words”; 3) use “trial series” of tasks to make sure that the subject adequately perceived the instructions; 4) change the wording of the instructions to make sure that the subject understands them correctly.

Variables associated with the researcher as a participant in the experiment.

The researcher can subconsciously give certain non-verbal signals that tell the subjects what type of reaction is expected from them (the Pygmalion effect, which Rosenthal described). Such signals can distort the real experimental effect, especially since on the part of the subjects sometimes there is a tendency to “please” the experimenter, to “help” him confirm the hypothesis. So, secondary variables in determining the results of an experiment can be no less influential than the independent variable. Several aspects of the possible influence of the experimenter on the results of the experiment were studied, and the phenomenon was specially called the effect of experimenter bias. Thus, the appearance of certain gestures and even statements by the experimenter, correlated in time with the appearance of stimuli, on which, according to the hypothesis of the experiment, a certain experimental influence was expected. The occurrence of such behavior on the part of the subjects reflects their recognition of the status of the experimenter and their comfort.

Other options for identifying experimenter bias are possible, for example, in the form of a tendency to observe only what he is determined to see. This effect is called pre-experimental expectations. To prevent these undesirable effects, quite effective means have been developed, which will be discussed in more detail in the next section. We will also learn about ways to prevent inappropriate behavior of the experimenter's assistants, which can reveal to the subjects the real goals of the experiment and thus turn into another potential source of the appearance of side variables.

When conducting observations as a methodological method for collecting experimental data, many different effects can arise (gala effect, “first impression” effect, “average tendency” effect, etc.), which also need to be considered as a source of situationally determined side variables.

Another possible source of side effects is the physical and psychological characteristics of the experimenters, which may have different effects on the subjects' performance of the experimental tasks.

Rosenthal proposed to distinguish at least three categories of characteristics of experimenters: biosocial (age, gender, race, religious beliefs); psychosocial (level of anxiety, need for social approval; hostility, authoritarianism, intelligence, dominance, social behavior related to status) situational factors (the presence or absence of previous social contacts between the experimenter and the subjects).

The presence of a sufficiently large number of characteristics of the experimenter that can influence the results of the experiment does not mean that the situation with a psychological experiment is so complex. As McGuigan rightly notes, even if these factors influence the dependent variable, the influence will be uniform in both the control and experimental groups, largely neutralizing its effect. However, psychological experiments differ from natural science experiments in that the “human factor” plays a large role in them. Consequently, the experimenter's task is to track the influence of this factor and prevent it or control it. attention when interpreting study results.

Possibility of observation and measurement variables there is a condition for using the experimental method. Observation does not mean in this case the applicability of the method of psychological observation. We are talking about the possibility of fixing or registering some indicators as psychological variables. The variable can be represented in terms of the subject's reaction time, measured using a stopwatch. This may be the frequency of occurrence of certain changes in the behavior of the subjects obtained on the basis of psychological observation methods. The definition of a variable in the most general sense of the word may sound like this: a variable is a reality whose changes can be measured in some way. Problem measurement variables for the purpose of building psychological scales is the subject of another branch of experimental psychology - psychological scaling. To introduce the experimental method into the standards, it is enough to indicate the main types of variables as already measured, or rather, methodically presented, and take into account their role from the point of view of their place in the organization of the experiment (the role of controlling variables and recording results from the point of view of establishing causal dependence).

4.2.1. Control of the independent variable and the problem of experimental influences

What can act as an experimental influence, or NP, in a psychological experiment? What changes in reality can be interpreted as experimental influences? Can NP in a psychological experiment be reduced to variable conditions? The specificity of a psychological experiment often results in the fact that none of these questions can be answered unambiguously. However, the list of signs of NP is quite definite, guided by which the psychologist determines whether NP has been identified and whether it has been carried out in relation to it. experimental control and, therefore, whether a truly experimental study (or any other) was conducted.

The first sign of NP is managing the levels of the influencing factor, i.e. implementation of special efforts of the experimenter to functional variable control carried out with the aim of interfering in the process being studied. The second sign is the representation of changes on any scale, qualitative or quantitative. The fact that changes in a variable depend on how the researcher controls it allows us to consider NP as influencing factor, or experimental influence(X-impact).

In subject matter different areas of psychological experimentation, different ideas have emerged about what and how a psychologist can control NP as a causally active variable. In Russian, this controlled variable is called independent(although it depends on the experimenter), and dependent variable is called the measured variable, the indicators of which, when planning an experiment, are considered as responses or consequences influence of NP.

Scientific schools in psychology differ not only in their substantive conceptual foundations, but also in what types of experiments they introduced into the practice of psychological research. In the behaviorist approach to organizing experimental research, it is precisely the control of conditions as stimulus factors set NP levels. In the school of K. Levin, starting positions field theory suggested, on the contrary, the interaction of variables situations And personalities(in concretizing her motivational aspirations at the level of the hypothetical construct of “quasi-need”). If socio-psychological And psychophysical experiments and implemented the general logic of experimental conclusion, they were based on significantly different ideas about the methods of specifying (selecting, operationalizing) controlled variables.
In a psychophysical experiment, NPs appear in stimulus factors. The physical parameters of the stimuli are changed experimentally, in accordance with which the subjective characteristics corresponding to them are established empirically. A subjective series of sensations in a psychophysical experiment is measured by constructing a scale of the subjects’ reactions (verbal or nonverbal). The goal of measurement, realized in a psychophysical experiment (as the construction of a subjective scale), brings to the fore the problem of establishing quantitative laws as representing functional connections between the dimensions of the stimulus series and the dependent variables of the subjective series, but pushes into the background the problem of causal interpretations. Establishing the type of dependence (a metric of psychological space, be it psychophysical or psychosemantic measurement procedures) does not necessarily imply an answer to the question of why a psychophysical function has one form or another. Therefore it is no coincidence that psychological dimension not only appears as a separate set of methods, or research paradigms, but is formalized in the form of separate academic disciplines in special manuals on psychometrics, psychological measurement, and experimental psychosemantics.
In experimental psychology textbooks, psychophysical experiments are discussed either as part of historical excursions or as rigorous diagrams demonstrating the features of design intraindividual experiments. In this textbook, the topic of psychological measurement will be touched upon in only one aspect - the need to distinguish between scales that reflect qualitative and quantitative changes in psychological variables.

Changing instructions is the most widely used method of actual manipulation of stimulus conditions as NP in psychological research. The task that a person performs with this method of experimentation remains the same, but the experimental and control conditions are different changing instructions.

The "neutral" instruction is usually presented in control condition where the independent variable is depicted by its inactive level. Thus, when conducting experiments in which subjects solve problems with matches, the purpose of the “neutral” instruction is to indicate to the subject the purpose of his actions, to ensure his inclusion in the experiment, but not to emphasize one or another motivational direction of his actions.
Excursion 4.2
According to the “neutral” instructions, the subject is required to find as many solutions to the problem situation as possible. Let's give an example from the class of so-called small creative problems. The instructions for the match problem say: “You are required to remove four matches so that three squares remain. Report all your actions out loud, calling out the numbers of the matches being removed.” The initial and final configurations with such instructions correspond to the figures of the arrangement of matches in Fig. 4.1.

Rice. 4.1. Condition (A) and result (B) of solving the problem with matches.
When subjects independently form goals, or in the process of voluntary goal formation, the instructions continue: “Now you are presented with another configuration of matches. What changes can be made to it, turning it into a new configuration? (Fig. 4.2.)

Rice. 4.2. Condition (A) and completion (B) of the task.
A “neutral” instruction supposedly evokes an attitude on the part of the subject toward the experimental situation, which is called “expertise motivation.” Within the framework of this relationship, acceptance of the goal - to complete the task - simultaneously means a willingness to show oneself well in front of the experimenter.

In another - “motivating” - instruction, special emphasis may be placed on what requires finding as many solutions as possible. So, in one case, this will be the researcher’s focus on updating the subject’s self-esteem (if it says: “This will determine your intellectual level”). In another case, a motivating instruction may involve the motivation of competition with others, an orientation towards the “achievement” orientation of the subject’s goal (for example: “According to your indicators, we will determine those who will go on an excursion to another city”). Any motivating emphasis in the instructions sets, according to the assumption of such a method of managing motivation as instructions, more active level independent variable " motivation" The situation with an active level of NP acts as an experimental condition. The experimental effect is determined by the quantity and quality of solutions demonstrated by subjects in two groups - respectively experimental And control.


In the book “Psychological Mechanisms of Goal Formation” you can read about the effects identified when analyzing the qualitative and quantitative parameters of solving such “ill-defined” problems in which the processes goal setting act as leaders and determine the features of decisions. With significant individual differences in problem solving, when comparing experimental and control conditions, a tendency is revealed such as an increase in efficiency in conditions with special “motivating” instructions. In this case, there are more solutions, and their originality is higher.

4.2.2. Two main schemes for setting NP levels

Comparison of solution indicators under conditions of “neutral” and “motivating” instructions can occur when implementing two main schemes: intra-individual And intergroup. In the first case, the same subject is presented with successively different levels of NP or goes through different conditions of NP. In example 4.2, these changes were expressed only in a change in instructions. It is important that both the experimental material changes (in this example, a mental task; you cannot solve the same one again!), and the subjective state—the subject’s preparedness to accept a certain type of problem (the degree of familiarity with the experimental activity). For now, let’s single out the task factor as the first side variable, which can be mixed with the effect of NP.

In the second case, the intergroup design, different instructions can be given to different groups of subjects simultaneously performing the same activity. It is for the material of small creative problems that group solutions are poorly controlled conditions. The subjects can spy on them, give hints, but it is difficult to keep a record of their reasoning for each individual. In this regard, the “simultaneity” of the presence of subjects in one condition or another is usually relative: it is possible to conduct individual experiments, the results of which are combined into a common group of decisions with the same instructions. The experiment will be called intergroup not by the form of implementation (the decision is made in a group or individually), but due to the method of comparing data, or sample values ​​of the salary. At intergroup The main thing in the scheme is that some subjects went through the control condition, and others went through the experimental condition, and the results of decisions in these two groups are compared.

The first factor, or the first side variable, which obviously influences the result of comparison in these schemes, is the difference between people who find themselves in different groups, or the composition of the groups. The main experimental effect - the result of exposure to NP - in this case adds up (multiplies) with the effect of individual differences. Groups may find themselves unequal composition, and there are many reasons for this. Let us remember the teacher’s “natural” desire to give “advanced”, from his point of view, students a more difficult test task. This may act as a “bias effect” on the part of the experimenter. Another mixing option: students who usually spend time together asked to join one group; often these are people with similar levels of intellectual capabilities. The list of variables that “threaten” the correct decision-making on the action of an NP can be continued. Let us dwell on the fundamental point: the experimenter must somehow disrupt all those non-random shifts in the supposed similarity of groups that can mix the NP effect and the effects of non-equivalence of groups (factor group composition). Selection strategies And selection subjects into groups– the main form of experimental control in intergroup designs.

So, experimental effect for both comparisons – inindividual And intergroup is established in a similar way, namely as a comparison of two series of sample values ​​of PP measured in the subject or subjects under experimental and control conditions. For a number of psychological hypotheses, the researcher’s choice between these two main experimentation schemes is not fundamental; it is determined by considerations of convenience, cost-effectiveness of research, preferences of certain forms control threats to withdrawal about the dependence being studied.

For other hypotheses, such a choice is not necessary due to the meaningful interpretation of the cause-and-effect relationship or the characteristics of the processes or phenomena being studied. If the hypothesis is formulated in the direction of individual generalization applied to an individual, then intergroup comparisons will be inadequate.


For example, a working hypothesis is tested that a person has impaired color perception (he is a color anomaly). It is clear that different kinds of color stimuli will change in a series of presentations to the same person. It is this person who will be concerned with subsequent generalizations about the type of established dependence. Similarly, when determining the range of another type of sensitivity - auditory - constructed in a certain way audiogram reflects the curve of the threshold audibility of sounds of different frequencies by a specific person. Another question is that a discussion of the type of this empirically obtained psychophysical dependence can be based on the most general theories that apply as theoretical models to all people.
Another type of hypothesis involves generalization to individual groups of people or populations. In such a case, the content of the hypotheses often suggests that between-group comparisons will be used. Many hypotheses in the field of social psychology focus on comparing individual and group decisions. It is discussed how the quality of decision-making changes, the level of risk accepted by a person, his assessment of the attributes of alternatives, etc. Without giving examples of the hypotheses themselves, we note one commonality: they presuppose a type of group or individual activity, but in the presence of other people. Here, too, the experiment can serve the purposes of individual communication, i.e. answer the question of, for example, whether a particular subject’s (Ivanov, Petrov, Sidorov) decision-making method changes if others are present nearby. Here the “group” of these others can only serve as an experimental condition. At the same time, an experiment can serve the purpose of broader generalizations; its result can apply to all people of student age (if the subject was a student), to all people of similar intellectual level or expected moral development, to the entire surveyed empirical community, etc. Ensuring a representative sample of subjects is a fundamental method of experimental control for such decisions about the breadth of generalization.
Suppose that experiments are planned to test hypotheses about how crowd effects. On the one hand, we are talking about changing the worldview of one person (a person in a crowd), but on the other hand, it is supposed to identify general trends in people’s actions, i.e. the answer to the question of how any person can behave in a crowd. Another thing is that a specific person, a specific personality may, to a certain extent, turn out to be independent of external influences. This assumption will entail a chain of new hypotheses - about what personal properties contribute to resisting the influence of the crowd. However, this will not change the original hypothesis, which applies to any person, i.e. without taking into account the additional personal variable “independence”.
So, management situational factors, including instructions And stimulating factors, possibly with different experimental designs. This cannot be said about other methods of specifying NP.

In modern psychology gender differences the “gender” variable acts either as an analogue of the NP, implying the similarity of groups of people in all other characteristics, or as a key additional variable. Thus, children of the same age and education can perform verbal tasks with different effectiveness: girls are ahead of boys at a certain period.

Excursion 4.3
Consider an example of gender differences presented in a study of the motivational sphere higher education teachers. Motivational profiles were compared in three groups of subjects: undergraduates, graduate students, and university teachers. At student age and during graduate school, similar motivational profiles were observed for men and women (the first places in them were occupied by “achievement motivation” and “aggression”). The group of university teachers was also divided into subgroups based on gender. In the new groups of men and women, the motivational profiles turned out to be significantly different. For male teachers, no significant changes were observed; in all three groups - undergraduates, graduate students and teachers - the indicated types of motives were in first place. On the contrary, in subgroups of women, significant changes were observed in indices of various types of motivation. For female teachers, in the average indicators of motivational profiles, the first places were taken by other types of motivation: “willingness to provide care” and “willingness to accept care.” Our proposed interpretation of the “floor effect” was as follows. The professionalization of teachers in higher education is associated with different lines of personal development of men and women. Social conditions turn out to be different for men and women, leveling out for the latter the importance of focusing on external indicators of success (the achievement motive) and contributing to the change of aggressive tendencies in communication with others to caring tendencies.
Let us leave unanswered the question of why working conditions in higher education allow male teachers to “preserve,” as it were, those prevailing types of motivation that characterized them as students. It is important to emphasize another point of the problem: the interpretative field of hypotheses in relation to the effect of the “gender” variable actually covers other variables and interactions (in the system of social conditions for personal growth and personal preferences of people belonging to different samples).

The formula introduced by K. Levin that behavior is a function of the individual and the environment highlights another class of variables, contrasted with situational ones. These are any variables that reflect dispositional factors tested using psychodiagnostic techniques. individual differences. Variables reconstructed as personal properties (dispositions, traits, motives, self-attitudes, etc.) are sometimes also called latent. This fixes their status as potential factors, determining the behavior of the subject in experimental situations. It has been established that in the same situations people act differently. Psychologists have spent a lot of effort trying to connect actually recorded changes in dependent variables with personal dispositions. However, as shown in the book Man and Situation, the dispute between situationism And dispositionism cannot be resolved empirically: a sufficient number of schools and studies argue in favor of inequality (and greater predictive power) of both variables. Both classes of variables are presented in experimental designs as actually controlled or in quasi-experimental designs as allowing their levels to be distinguished by selecting groups.

4.2.3. Variables “situational” and “personal”

In the experiments of K. Lewin's school, control personality variables was supposed to be possible on the basis of the creation of dynamic tension systems that predetermine the direction of behavior and were called quasi-needs. The methods of creating these quasi-needs were different: interrupting the sequence of actions performed by the subject (the effect of unfinished actions in the study of B.V. Zeigarnik), setting instructions for different requirements for reality-irreality, or more precisely, for the real feasibility of experimental actions (in the study of “magical thinking” by Vera Mahler). Later, it was the sequences of successful and unsuccessful attempts at completing mazes (in Yuknat's study), which influenced the level of aspirations of the subjects. All these experimental models recreated living space, in fact, implementing a certain theoretical idea of ​​the methods of personal self-regulation as the formation of individual intentions.

Other theoretical ideas about personality structures, on the contrary, assumed the status of personality variables as independent of the experimenter and not amenable to control as independent variables. Such latent dispositions began to be included in experimental designs: for example, in the book of X. Heckhausen there is a section representing experimental designs to study motivation .

Measuring indicators indicating the nature and changes in processes influenced by experimental influences, in conjunction with accepted standards for their psychological interpretation (in relation to the psychological constructs and methods used), makes it possible to reconstruct the processes presumably behind changes in the dependent variable.

As for NP, for GP it is important to distinguish between a fixed indicator and hypothetical mechanisms of functioning basic process , standing behind the relation implied in the hypothesis. The independent and dependent variables, and the implied relationship between them, are the constituent parts experimental hypothesis(EG). Some authors introduce the concept as a component of EG hypothetical construct, including unobservable and reconstructable mechanisms of change in variables. Other authors seem to put all those interpretive components, which are associated with the understanding of psychological constructs and function as “bridges” between the empirical content of EG and the theoretical justification of the alleged dependence.

By managing the NP, the researcher is actually trying to influence the basic processes being studied - the unobservable subjective (psychological) reality that he reconstructs. Understanding subject studying this should not be confused with understanding the possibilities of controlling variables in a psychological experiment. Thus, J. Campbell identifies the following types of NP as the basis for the creation experimental And control conditions (irrespective of theoretical ideas about the processes being studied):


  • controlled variables or factors such as teaching method. Other authors more often use the concept of stimulus conditions, or stimulation conditions, which can be either holistic situations (for example, simulated on simulators) or changes in individual characteristics of stimuli;

  • potentially manageable variables that the experimenter could, in principle, change, but for some reason does not. J. Campbell names school subjects. The question of why the experimenter does not include potentially controllable variables within the framework of the experimental design often brings up the discussion of evaluative aspects of experimentation: assessment of the cost-effectiveness of its implementation, the ethics of certain forms of experimental influences, masking of experimental conditions, etc.;

  • relatively constant environmental aspects(socio-economic level, locality, school, etc.). These variables are not under the direct control of the experimenter, but can act as fixed grounds for dividing subjects or conditions into certain classes as levels of NP;

  • "organismic" variables: gender, age and other objectified characteristics. In this case, we are also talking about the possibility of selecting groups that are equivalent in this characteristic or differ in it;

  • test takers or pre-measured variables. It is clear that this refers to the entire arsenal of psychological techniques, according to which classifications or the identification of groups of subjects are possible. These variables can be classified as these types of internal condition variables, which form, qualitatively, perhaps the most diverse class of psychological variables.

In the aforementioned book by another authoritative psychologist, H. Heckhausen, the planning of experiments to study problems of human motivation is presented as dependent on a theoretical understanding of the constructs motive And motivation. Ways to control variables - managing instructions, selecting groups of people differing in latent motivation, etc. - appear not as arbitrary decisions of the experimenter, but as a choice determined by his understanding of the subject being studied. Thus, resolving issues meaningful planning research leaves an imprint on the methods of implementation experimental control of variables. The same methodological techniques can be found in completely different (in terms of the subject content of the hypotheses) experimental studies.


Methodological tools, considered in the context of diagnosing the cognitive sphere or personal characteristics of a person, lead the researcher to at least bivalent experiments, taking into account the difference between two groups of subjects on a specific indicator. For example, the method of median sample splitting can be used to set different levels of such an analogue of NP as the internal conditions variable. However, the impossibility of changing these levels in relation to a specific person requires clarification of the nature of the “control” of the variable. For such variables, reconstructed on the basis psychodiagnostic indicators, functional control comes down to the selection of groups that differ in a given indicator, and not to management as an influence. This means using not strictly experimental, but “quasi-experimental” designs (see Chapter 13).
The given version of the classification of types of NP does not imply taking into account theoretical interpretations or phenomenal characteristics of psychological reality. Such a formal approach is only possible when discussing the general structure of the study and is insufficient when considering the problem meaningfully: what actually changed as a variable. Let us give an example of J. Gibson's analysis of an experiment that served as a fulcrum for him in rethinking the factors included in the regulation of the perception of “depth” (as the third dimension in the perception of space).
Excursion 4.4
When it changes illumination of a plastered wall- from strong to barely noticeable - subjects in Metzger's experiments saw a uniform field because the light was not focused. In strong lighting conditions, the observer-subject saw a wall. In low light, the fine texture of the surface was not perceived by the eye, and the observer said that he did not see a two-dimensional surface, but fog, haze, or a “haze of light.” For the author of this study, the latter was the basis for asserting that the subject began to perceive third measurement, i.e. "space".

J. Gibson questioned the interpretation of the "light haze" as a phenomenon associated with the observation of a two-dimensional plane. He showed that a uniform field can be obtained in other ways: placing a hemisphere of diffuse glass in front of the subject’s eyes, brightly illuminating it from the outside, or putting frosted caps on both eyes. The subjects saw something that lacked depth. They perceived the "environment" rather than the third spatial dimension. Phenomenally, it was not reminiscent of a “light haze”, but rather of “looking at the sky”, where there are no objects or surfaces. These two circumstances - the ability to produce the same "uniformity" effect in different methodological ways and the possibility of observers reinterpreting what they perceive - allowed Gibson to argue that the essence of Metzger's experiments and their subsequent analogues was "not in the wall, not in the panoramic surface and not in diffusive caps." The controlled variable was optical structure, and in Metzger’s experiment the levels of this factor differed in that at the extreme poles of illumination the optical structure had a structure or turned out to be a “structure without structure.” The fallacy of the theory of depth perception was now associated with the analysis of the fact that the surface (or the sign of two-dimensionality) was perceived only in the case when the perception of differences for different directions is possible, i.e. its texture.


So, the independent variable in psychological research cannot be reduced to varying conditions. Different methods can bring to life, or initiate, processes similar to each other, and the very fact of varying conditions requires justification, which acts as a causal variable.

4.2.4. Fixed indicator and basic process

In psychological research, the identification of GP is associated with a description of the basic process on which NP acts and which manifests itself in the parameters of GP. Using the example of J. Gibson's discussion of Metzger's experiment, one can see another aspect of the problem - the reinterpretation of the characteristics of a controlled NP. In these and other experiments in the field of perceptual psychology, the subject is an "internal observer" (an observer of his own perceptual experience) who reports in one way or another on the phenomenally presented data. The experimenter is already dealing with descriptions of subjective experience, i.e. with recorded data, in relation to which he takes the position of an external observer.

When moving from the “psychological observation” method to the “psychological experiment” method, the position of an external observer becomes the position of an experimenter who manages the organization of experimental influences (and in this sense, an active researcher). The fact that he himself can be both a subject and an experimenter (for example, the experiments of Ebbinghaus, Sperling, etc.) does not change the principle of constructing experiments, where, as a subject-observer, the subject-experimenter reports to himself about the data of a phenomenal order . As a researcher, he takes the position of an external observer, for whom the data of subjective experience (even his own) is not direct psychological knowledge, but a subject of study and understanding.

Thus, further under psychological reality This refers to the scientistic position of the researcher as an external or internal observer who is not directly presented with the processes that mediate the manifestation of certain psychological patterns. Psychological reality is not directly revealed to the external observer, therefore psychological laws include assumptions of an indirect nature, for which the causal dependence of the “non-visual” plan is reconstructed as a certain functional relationship between “observable” variables. It is to update the process of interest to the researcher and record POs that are created psychological techniques as special “techniques” for obtaining data. The arbitrariness of constructing the PP is the arbitrariness of developing a methodology that is adequate from the point of view of fixing the essential characteristics of the psychological reality under study and from the point of view of those used by the experimenter report forms about how he reconstructs the relationship between indicators and the basic process being studied.


Excursion 4.5
Let us give an example from the field of developmental psychology, demonstrating the indirect nature of the recorded indicators and their ambiguity as indicators of salary. A researcher who shares the views of J. Piaget on the nature of intelligence and cognitive development set the task of comparing problem-solving schemes as intellectual techniques used in youth (15-18 years old) and early adulthood (19–22 years old). The plan for comparing empirical data was quite simple. Two chemical problems that did not require the use of special knowledge were solved by two groups of subjects: 15-year-old tenth-graders and 20-year-old students. The controlled variable was represented by two levels of “adulthood”. The task required the subject to use combinatorics elements. A protocol of “reasoning out loud” was recorded.

The analysis of the protocols was carried out according to 5 points:

1) overall success of the solution as completeness, correctness and explanation of the answer (assessment on a four-point scale, correlated with the stages of development of specific operations, according to Piaget);

2) the result of general processing - the proportion of various types of “elementary components” of thought processes recorded in the protocol;

3) “heuristic” processing – the percentage indicator of the use of each of the heuristic techniques found in the decision protocol;

4) “strategic” processing – the proportion of different strategies used by the subject;

5) comparison of the solution algorithms proposed by the subjects with the template and analysis of the logic of constructing statements.
Let us consider the list of those formations that in the “reasoning aloud” protocol were classified as “elementary components”, “heuristics” and “strategies”. 24 “elementary” operations were identified: a person’s viewing of this information, making assumptions, listing possible questions, selecting evaluation criteria, determining preferences, etc.

When solving the problem, the subjects used 5 types of “strategies”:

1) feedback;

2) algorithmic;

3) extracting the sample;

4) hypothetico-deductive;

5) systematic evaluation.
Studied basic process– a person’s use of intellectual operations when solving a problem – was reconstructed through multiple qualitative and derived quantitative comparisons. The results confirmed the assumption that young men and adults use the same problem-solving schemes in their judgments. It was found that schoolchildren often display logical incompetence: they inadequately define and delimit problems and do not formulate hypotheses well enough.
The result of this study is important in another aspect. This study demonstrates the impossibility of a single GP and the need for a systematic analysis of a number of properties of thinking, i.e. multiple dependent variable
olderfiles -> Cognitive activity of students
olderfiles -> The influence of individual psychological characteristics on professional choice
olderfiles -> Seminar of psychologists, social educators and deputy assistants. “Prevention of deviant behavior among children and adolescents”

Independent Variables - these are the means that the experimenter has at his disposal to influence the subjects.

Dependent Variable - these are those changes that occur in the state, behavior, communication and activity of the subjects under the influence of the experimenter.

External variables (interference) – these are factors (conditions) that also affect the subjects against the will of the experimenter (fatigue, noise outside the window).

The experimenter should strive to reduce the impact of external variables on the subjects.

Independent Variables

                High quality – are expressed in the fact that any effect is either present in the experiment or absent.

(For example: experimenter's prompt (maybe, maybe not)).

                Quantitative – is expressed in various degrees (more than two) of the experimenter’s influence on the subjects.

Independent Variables may have the following forms:

    Appear in:

    In task elements;

    In the elements of the experimental situation;

    In the characteristics of the subject's condition.

In the task, the experimenter varies:

1. Characteristics of the impact (strong, weak impact).

2. The assignment material (written, on the computer).

3. The type of response of the subject (verbal and non-verbal).

4. Grading scale (money).

5. Instructions (may change, but should not).

6. The purpose and action of the subject.

7. Funds that the subject has.

8. Obstacles that are put in front of him.

9. System of rewards and punishments for subjects.

The situation varies:

1. Physical parameters:

    Equipment location;

    Exterior of the premises;

    Illumination;

    Sounds and noises;

    Temperature;

    Placement of furniture;

    Experiment behavior time

    • How long;

      Times of Day.

2. Social and psychological parameters:

        Work in the presence of an experimenter;

        Either work in isolation;

        In Group;

        Or when communicating with the experimenter.

3. Features of the subject’s condition:

    Mental stress;

    The level of experienced stress, fear, frustration: the level of conformity;

    Rivalry, etc.

Types of dependent variables

(what the experimenter manipulates)

    One-dimensional – only one parameter is registered.

    It is this that is considered a manifestation of the dependent variable. (Reaction time) Multidimensional

    – a number of parameters are highlighted and recorded. .

Dependent Variable (For example: time to solve a problem, quality of solution, originality, etc.). Fundamental is considered as a function of individual parameters of its manifestation. (For example:

    level of aggressiveness

    – is considered as a function of the interaction of the following parameters:

    Verbal aggression;

    Non-verbal aggression;

Indirect aggression (slammed the door loudly); Manifestation of irritability, etc. , Where

A

- parameter. This principle is used to determine the level of aggression. Aggression questionnaire. A change in the dependent variable is determined by a change in the parameters of the subject's behavior. They are conventionally divided into formal-dynamic .

And meaningful TO

    formal-dynamic

    include:

    Task accuracy

    Latency (i.e., the hidden time from the moment the signal is presented to the choice of response).

    Execution duration Rate or frequency of action

Productivity – i.e. the ratio of the obtained result to the execution time.

Extrapunitiveness – external accusatory behavior (S. Rosenzweig)

Intropunity - self-accusation.

Dependent Variable Impunity – does not blame anyone, but tries to understand the situation. must be valid A change in the dependent variable is determined by a change in the parameters of the subject's behavior. They are conventionally divided into (degree of compliance),

reliable sensitive.

Validity – means that the level of manifestation of the dependent variable is stably fixed during the experiment.

Sensitivity – tells how much the dependent variable changes when the independent variable changes.

There are 2 variants of manifestation of the type of insensitivity of the dependent variable:

    Ceiling effect – manifests itself in the fact that the task is so simple that the level of its performance by the subject is higher than all levels of the independent variable. (For example: multiplication table).

    Floor effect – on the contrary, it occurs when the simplest level of the independent variable is so difficult for the subject that he is not able to perform it.

Changes in dependent variable manifest themselves in changes in a person’s state, communication, behavior, activity, facial expression, posture, activity, emotionality, etc.

Types of variables and their mixtures in a psychological experiment.

Variable- This:

1. any reality, the observed changes of which (according to specific parameters or indicators of the methodology) can be recorded and measured on any scale (Kornilov).

2. Any reality that can change, and this change is manifested and recorded in an experiment.

Experimental design includes an indication of the scheme for presenting the NP to different groups of subjects or the sequence of NP levels for one subject, the number of experiments, the number of subjects, and the plan for fixing the NP. The principle of isolated conditions – there is a general effect, the result of the action of the NP and the effect of the interaction of different levels of different variables. Principle of functional control of NP .

1) Independent variable (IV) – experimental influence (experimental factor, X-impact), controlled (actively changed by the e-tator) or functionally controlled variable, presented at 2 or more levels. Accepted in the hypothesis as a causal factor.

Aspects of NP allocation:

NP as a causal factor;

Rationale for NP as a psychological variable;

The question of the implementation of causal influence in the methodology;

Willingness to use the impact of NP for scientific purposes.

Types of NP (according to Campbell):

Controlled (complete situations, changing incentives);

Potentially manageable (eg: not governed by ethical principles);

Relatively constant (division into classes at school - NP levels);

Organic (gender, age, etc., usually DP);

Testable/measurable (in methods and tests).

Single NP– the result of isolating a variable in laboratory conditions; changes in the pattern are presented in isolation from other factors, which makes it possible to test precise hypotheses. In laboratory ex-ts.

Complex NP– the cut levels are set by a set of interrelated conditions; Usually they represent the ex-model in an artificial ex-those.

2 ) Dependent variable (DP) - response, or measured variable, changes that are causally determined by the action of the NP; represented by indicators of the subject’s activity; denoted as O - a fixed, observable and measurable indicator.

3) Additional Variables (AD) - influence the experimental conditions, but we cannot control them. DP included in the formulation of the experimental hypothesis (unlike PP) as clarification of conditions, under which NP action is expected . Their influence not statistically assessed , unless they are considered in factorial schemes as an independent NP. DP level determines the possibility of subsequent generalizations to reality and is usually indicated in the E-tal hypothesis . In this case, the selected levels of DP will always limit the possibilities of transfer, however, making them more conclusive. DP make the hypothesis less accurate , because suggest relationships between the basic variable and other influences that must be taken into account. At the same time, DP and specify the hypothesis , because indicate area of ​​action the pattern being studied. With the help of DP, the control of confusion, as well as clarification of the type of functional dependence.

Example : study of behavior under risk conditions based on lotteries. The subjects made hypothetical payments. DP – conditions for observing the behavior of players in real life. The decisions made in the conditions of real payments corresponded to the same patterns, cat. were found in hypothetical payment conditions.

DP present in EG: the population of potential subjects, the type of experimental influences, methods of recording PP - all these are potential sources for the development of factorial plans.

4) Basic variable – NP has its effect on it (according to Gottsdanker).

And according to Campbell ( basic side variable - BPP): This is a variable of internal conditions (interindividual difference factors) that is confounded with the underlying underlying process being studied (represented by the underlying underlying variable).

5) Latent variable – a hypothetical variable that cannot be measured in the study, but in the model of relationships between variables characterizes source of measured variables; unrefined influences (“disturbed” factors) affecting the measured variable.

6) Related Variable – defining placebo effect

7) Spillover variable (SP) or confounding variables – is not included in the EG, but under ex-ta conditions it can affect the basic process being studied or be mixed with the NP or ZP, which will distort the type of dependence being studied. The extor must try to avoid the influence of SP, otherwise poor internal validity is a low degree of the fact that the empirically established relationship is between X and Y, and not X and SP or Y and SP.

There are three types of mixtures:

1. Unsystematic mixing occurs when any of the factors (or combinations thereof) Interferes irregularly with the addiction being studied . The source of PP associated with the influence of the time factor can be: internal reasons (changes in the state of the subject himself, background fluctuations in the salary indicator), and external(accidental distraction by noise in the corridor, a shout from a colleague, a phone call, etc.). If they are unevenly distributed across the compared NP conditions, then there will be a distortion of the experimental effect (as a difference in GP rates).

One of the consequences of such irregular influences of PP is data unreliability, i.e. with a different scatter of PP levels - across samples over time - a different relationship between PP values ​​and NP levels is established. Typically this threat to the conclusion of experimental dependence controlled in two opposite directions. On the one hand, the experimenter strives reduce the number of samples in the overall experimental sequence to a minimum in order to conduct an individual experiment in the shortest possible time, leveling the time factor. On the other hand, the experimenter must provide a sufficiently large number of samples , i.e. try to approach an infinite experiment so that all mixing with oscillations on the part of the PP is randomly - and in this sense, equally - distributed between the NP levels.

Other sources of data unreliability. may take place variability of the NP itself when the ex-tor considers samples assigned to the same level, but in fact something happened in one or part of them that does not allow us to consider the conditions identical.

2. Systematic mixtures – the main type of threats to internal validity.

Non-random mixing of NP and PP levels, when active or inactive PP levels are regularly combined with a certain NP level , as a result of which it cannot be concluded that the established experimental effect is associated with the action of NP, and not PP.

The main factors threatening systematic mixing th: experimenter effects, task, time, sequence, and interindividual differences factors.

3. Associated Confusions - The occurrence of accompanying mixtures is associated with the by specifying the NP, fixing the PO or updating the basic variable being studied in conjunction with another basic process .

Concomitant confounding is distinguished by the immanent connection of the confounding variable with the methodological conditions for controlling the variables or with other variables.

Control– updating the active level of the accompanying variable in all experimental and control conditions. As a result, the effect is subtracted from the total result of the action of the ex-factor. The place of the accompanying variable can be taken by different, but factors associated with the main NP. An example is an experiment with craniotomy.

Related Variable– defining mixing the ORD of an experimental factor with the influence of the method of presenting its conditions . The effect of its active level can be represented placebo effect . It is controlled by a general scheme for setting active SP levels at all NP levels.