Experiment design (DOE , DOX or experimental design) is the development of some task that seeks to describe or explain the change in information under conditions that are hypothesized to reflect the change. The term is usually associated with experiments in which the design introduces conditions that directly affect change, but can also refer to the design of quasi-experiments in which natural conditions that affect change are chosen for observation.

In its simplest form, an experiment aims to predict outcomes by introducing a change in preconditions, which is represented by one or more independent variables, also referred to as "input variables" or "predictors." A change in one or more independent variables is usually hypothesized to result in a change in one or more dependent variables, also referred to as "output variables" or "response variables." The pilot design may also define control variables that should be kept constant to prevent external factors from influencing the results. Experimental design includes not only the selection of appropriate independent, dependent, and control variables, but planning the delivery of the experiment under statistically optimal conditions, taking into account the limitations of available resources. There are several approaches for determining the set of design points (unique combinations of explanatory variable settings) to be used in an experiment.

The main concerns in development work include action creation, reliability, and reproducibility. For example, these problems can be addressed in part by carefully choosing the independent variables, reducing the risk of measurement error, and ensuring that the documentation of the methods is sufficiently detailed. Related challenges include achieving appropriate levels of statistical power and sensitivity.

Properly designed experiments advance knowledge in the natural and social sciences and engineering. Other applications include marketing and policy development.

story

Systematic clinical trials

In 1747, while serving as a surgeon on HMS Salisbury, James Lind conducted a systematic clinical trial to compare remedies for scurvy. This systematic clinical study is a type of ME.

Lind selected 12 people from the ship, all suffering from scurvy. Lind restricted his subjects to males who "looked like I could them", that is, he granted strict entry requirements to reduce outside change. He divided them into six pairs, giving each pair a different supplement to their basic diet for two weeks. The procedures were all means that were suggested:

• A quart of cider every day.
• Twenty-five Gutts (drops) of vitriol (sulfuric acid) three times a day on an empty stomach.
• One half pint of sea water each day.
• A mixture of garlic, mustard and horseradish in a nutmeg-sized lump.
• Two tablespoons of vinegar three times a day.
• Two oranges and one lemon every day.

Citrus treatments stopped after six days when they ran out of fruit, but by then one sailor was fit for duty and the others had nearly recovered. In addition, only one group (cider) showed some effect of his treatment. The rest of the crew presumably served as controls, but Lind did not report results from any control (untreated) group.

Statistical experiments, next C. Pierce

The theory of statistical inference was developed by Ch. Peirce in Illustrations to the Logic of Science (1877-1878) and The Theory of Probable Inferences (1883), two editions that emphasized the importance of randomization based on inference in statistics.

randomized experiments

C. Pierce randomized volunteers to a blind, repeated measurements design to assess their ability to distinguish between weights. Peirce's experiment inspired other researchers in psychology and education, who developed a research tradition of randomized experiments in laboratories and specialized textbooks in the 1800s.

Optimal Designs for Regression Models

comparison In some areas of study it is not possible to have independent measurements on a traceable metrological standard. Comparisons between treatments are far more valuable, and are generally preferred, and are often compared to scientific controls or traditional treatments that act as a baseline. Randomness Randomization is the process of assigning individuals to random groups or to different groups in an experiment so that each person in the population has the same chance of becoming a study participant. Random assignment of individuals into groups (or conditions within a group) distinguishes rigorous, "true" experiment from observational studies or "quasi-experiment". There is an extensive body of mathematical theory that explores the consequences of decisions to allocate units to a treatment by some random mechanism (such as tables of random numbers, or the use of randomized devices such as playing cards or dice). The assignment of units to the treatment is random, usually to mitigate the puzzling effect that makes the effects due to factors other than treatment, presumably as a result of the treatment. Risks associated with random distribution (eg, having a major imbalance in a key characteristic between treatment and control groups) are calculable and therefore can be managed to an acceptable level using a sufficient number of experimental units. However, if the population is divided into several subpopulations that are in some way different, and the study requires that each subpopulation be equal in size, stratified sampling may be used. Thus, the units in each subpopulation are random, but not the entire sample. The results of an experiment can be reliably generalized from experimental units to a larger statistical population of units only if the experimental units are a random sample from a larger population; the likely error of such an extrapolation depends on the sample size, among other things. Statistical replication Measurements are generally subject to variation and measurement uncertainty; Therefore, they are repeated and complete experiments are replicated to help identify sources of variability, to better evaluate the true effects of treatment, to further strengthen the experiment's reliability and validity, and to add to existing knowledge of the topic. However, certain conditions must be met before a replication experiment is initiated: the original research question was published in a peer-reviewed journal or is widely cited, the researcher is independent of the original experiment, the researcher must first attempt to replicate the original data using the original data, and The review should indicate that the study conducted is a replication study that attempted to follow the original study as closely as possible. blocking Blocking is the non-random arrangement of experimental units into groups (blocks/lots) consisting of units that are similar to each other. Blocking reduces the known but irrelevant sources of inter-block variability and therefore provides greater accuracy in estimating the source of variation under study. Orthogonality Orthogonality concerns forms of comparison (contrast) that can be legitimately and effectively exercised. The contrasts can be represented by vectors and sets of orthogonal contrasts are uncorrelated and independently distributed if the data is normal. Because of this independence, each orthogonal processing provides different information to the others. If there is T- procedures and T- 1 orthogonal contrasts, all information that can be captured from the experiment can be obtained from the set of contrasts. Factorial experiments Use factorial experiments instead of a single factor-at-a-time method. They are effective in assessing the effects and possible interactions of several factors (independent variables). An experiment design analysis is built on the foundation of ANOVA, a collection of models, Partition of the observed variance into components, according to what factors the experiment is to evaluate or test.

example

This example is attributed to Hotelling. It conveys some of the flavor of these aspects of the theme, which involve combinatorial constructions.

The weights of eight objects are measured using balance panning and a set of standard weights. Each weighty measures the difference in weight between objects in the left pan versus any objects in the right pan, adding a calibrated weight for a lighter pan, until the balance is in balance. Each measurement has a random error. The average error is zero; on standard deviations according to the distribution of the probability of errors coincides with the number σ on different weightings; errors at different weighings are independent. Let us denote the true weights with

θ 1 , ... , θ 8 , (\displaystyle \theta _(1),\dots,\theta _(8).\)

We will consider two different experiments:

1. Weigh each object in one pan, with the other pan empty. Let X I be measured the weight of an object, I = 1, ..., 8.
2. There are eight weighings according to the following graph and let Y I difference to be measured I = 1, ..., 8:

left pan right pan First weighing: 1 2 3 4 5 6 7 8 (Empty) second: 1 2 3 8 4 5 6 7 third: 1 4 5 8 2 3 6 7 fourth: 1 6 7 8 2 3 4 5 fifth: 2 4 6 8 1 3 5 7 sixths: 2 5 7 8 1 3 4 6 sevenths: 3 4 7 8 1 2 5 6 eighths: 3 5 6 8 1 2 4 7 (\displaystyle (\ (begin array) (lcc) &(\text(left pan))&(\text(right pan))\\\hline(\text(1 weighting:))&1\2\3\4\5\6\7\8&(\ text((blank))) \\ (\ text(2)) & 1 \ 2 \ 3 \ 8 \ & 4 \ 5 \ 6 \ 7 \\ (\ text (3rd: )) & 1 \ 4 \ 5 \ 8 \ & 2 \ 3 \ 6 \ 7 \\ (\ text (4th :)) & 1 \ 6 \ 7 \ 8 \ & 2 \ 3 \ 4 \ 5 \\ (\ text (5th :)) : )) & 2 \ 4 \ 6 \ 8 \ & 1 \ 3 \ 5 \ 7 \\ (\text(6th:)) & 2 \ 5 \ 7 \ 8 \ & 1 \ 3 \ 4 \ 6 \ \ (\ text (7th: )) & 3 \ 4 \ 7 \ 8 \ & 1 \ 2 \ 5 \ 6 \\ (\ text (8th:)) & 3 \ 5 \ 6 \ 8 \ & 1 \ 2 \ 4 \ 7 \ end (array))) Then the calculated value of the weight θ 1 is θ ^ 1 = Y 1 + Y 2 + Y 3 + Y 4 - Y 5 - Y 6 - Y 7 - Y 8 8 , (\displaystyle (\widehat (\theta))_(1)=(\frac( Y_ (1) + Y_ (2) + Y_ (3) + Y_ (4) -Y_ (5) -Y_ (6) - Y_ (7) -Y_ (8)) (8)).) Similar estimates can be found for the weights of other items. For example θ ^ 2 = Y 1 + Y 2 - Y 3 - Y 4 + Y 5 + Y 6 - Y 7 - Y 8 8 , θ ^ 3 = Y 1 + Y 2 - Y 3 - Y 4 - Y 5 - Y 6 + Y 7 + Y 8 8 , θ ^ 4 = Y 1 - Y 2 + Y 3 - Y 4 + Y 5 - Y 6 + Y 7 - Y 8 8 , θ ^ 5 = Y 1 - Y 2 + Y 3 - Y 4 - Y 5 + Y 6 - Y 7 + Y 8 8 , θ ^ 6 = Y 1 - Y 2 - Y 3 + Y 4 + Y 5 - Y 6 - Y 7 + Y 8 8 , θ ^ 7 = Y 1 - Y 2 - Y 3 + Y 4 - Y 5 + Y 6 + Y 7 - Y 8 8 , θ ^ 8 = Y 1 + Y 2 + Y 3 + Y 4 + Y 5 + Y 6 + Y 7 + Y 8 8 , (\displaystyle (\(begin aligned)(\widehat (\theta)) _(2)=(&\frac(Y_(1)+Y_(2)-Y_(3 )-Y_(4)+(5 Y_)+Y_(6)-Y_(7)-Y_(8))(8)).\\(\widehat(\theta))_(3)&=(\ fracturing (Y_ (1) + Y_ (2) -Y_ (3) -Y_ (4) -Y_ (5) -Y_ (6) + Y_ (7) + (Y_ 8)) (8)).\\ ( \widehat(\theta))_(4)&=(\r hydraulic fracturing (Y_ (1) -Y_ (2) + Y_ (3) -Y_ (4) + Y_ (5) -Y_ (6) + Y_ (7) (-Y_ 8)) (8)). \\(\widehat(\theta))_(5)&=(\frac(Y_(1)-Y_(2)+Y_(3)-Y_(4)-Y_(5)+Y_(6)- Y_ (7) + (Y_ 8)) (8)). \\(\widehat(\theta))_(6)&=(\frac(Y_(1)-Y_(2)-Y_(3)+Y_(4)+Y_(5)-Y_(6)- Y_ (7) + (Y_ 8)) (8)) \\. (\widehat(\theta))_(7)&=(\frac(Y_(1)-Y_(2)-Y_(3)+Y_(4)-Y_(5)+Y_(6)+(7) Y_ ) -Y_ (8)) (8)). \\(\widehat(\theta))_(8)&=(\frac(Y_(1)+Y_(2)+Y_(3)+Y_(4)+Y_(5)+Y_(6)+ Y_ (7) + (Y_ 8)) (8)). \ (end justified)))

Experiment design question: Which experiment is best?

Estimation variance X 1 of & thetas 1 is σ 2 if we use the first experiment. But if we use the second experiment, the variance of the estimate given above is σ 2 /8. Thus, the second experiment gives us 8 times more than the accuracy for estimating one element, and evaluates all elements at the same time, with the same accuracy. What the second experiment achieves with eight would require 64 weighings if the items are weighed separately. However, we note that the estimates for the elements obtained in the second experiment have errors that correlate with each other.

Many experimental design problems involve combinatorial designs, as in this example and others.

To avoid false positives

False positives, often resulting from publication pressure or author's own confirmation bias, are an inherent danger in many fields. A good way to prevent skew that could potentially lead to false positives during the data collection phase is to use a double-blind design. When double-blind designs are used, participants are randomly assigned to experimental groups, but the researcher is unaware of which group the participants belong to. Thus, the researcher cannot influence the participants' response to the intervention. Experimental samples with undisclosed degrees of freedom are a problem. This can lead to conscious or unconscious "p-hacking": trying multiple things until you get the result you want. This typically involves manipulating - perhaps unconsciously - during statistical analysis and degrees of freedom until they return the figure below p<.05 уровня статистической значимости. Таким образом, дизайн эксперимента должен включать в себя четкое заявление, предлагающие анализы должны быть предприняты. P-взлом можно предотвратить с помощью preregistering исследований, в которых исследователи должны направить свой план анализа данных в журнал они хотят опубликовать свою статью прежде, чем они даже начать сбор данных, поэтому никаких манипуляций данных не возможно (https: // OSF .io). Другой способ предотвратить это берет двойного слепого дизайна в фазу данных анализа, где данные передаются в данном-аналитик, не связанный с исследованиями, которые взбираются данные таким образом, нет никакого способа узнать, какие участник принадлежат раньше они потенциально отняты, как недопустимые.

Clear and complete documentation of the experimental methodology is also important in order to support the replication of results.

Topics for discussion when creating development projects

A developmental or randomized clinical trial requires careful consideration of several factors before actually doing the experiment. Experimental design of the laying out of the detailed experimental plan in advance to do the experiment. Some of the following topics have already been discussed in the Experimental Design Principles section:

1. How many factors does design have, and are the levels of these factors fixed or random?
2. Are control conditions necessary, and what should they be?
3. Manipulation checks; did manipulation really work?
4. What are background variables?
5. What is the sample size. How many units must be collected for an experiment to be generalizable and have sufficient power?
6. What is the significance of the interaction between factors?
7. What is the influence of the long-term effects of the main factors on the results?
8. How do response changes affect self-report measures?
9. How realistic is the introduction of the same measuring devices into the same units, in different cases, with post-test and subsequent tests?
10. What about using a proxy pretest?
11. Are there lurking variables?
12. Should the client/patient, researcher, or even data analyst be conditionally blind?
13. What is the possibility of subsequently applying different conditions to the same unit?
14. How much of each control and noise factors should be taken into account?

The independent variable of a study often has many levels or different groups. In a true experiment, the researchers can get an experimental group, which is where their intervention is carried out to test the hypothesis, and a control group, which has all the same element in the experimental group, without the intervention element. Thus, when everything else except for one intervention is held constant, researchers can certify with some degree of certainty that this one element is what is causing the observed change. In some cases, having a control group is not ethical. Sometimes this is solved by using two different experimental groups. In some cases, independent variables cannot be manipulated, such as when testing for a difference between two groups that have different diseases, or testing for a difference between men and women (obviously a variable that would be difficult or unethical to assign to a participant). In these cases, quasi-experimental design can be used.

causal attributions

In pure experimental design, the independent variable (the predictor) is manipulated by the researcher - that is - each participant in the study is selected at random from the population, and each participant is randomly assigned to the conditions of the independent variable. Only when this is done is it possible to verify with a high degree of probability that differences in outcome variables are caused by different conditions. Therefore, researchers should choose an experiment design over other design types whenever possible. However, the nature of the independent variable does not always allow for manipulation. In cases, researchers should be aware of not certifying causal attribution when their design does not allow it. For example, in observational projects, participants are not randomly assigned to conditions, and therefore, if there are differences found in the outcome variables between conditions, it is likely that there is something other than differences between conditions that cause differences in outcomes, which is the third variable. The same goes for studies with correlational design. (Ader & Mellenbergh, 2008).

Statistical control

It is best that the process is under reasonable statistical control prior to conducting the designed experiments. If this is not possible, proper blocking, replication, and randomization allow for the careful conduct of designed experiments. To control for interfering variables, the researcher establish control checks as additional measures. Researchers must ensure that uncontrolled influences (eg, perceptions of a source of trust) do not skew research results. The manipulation check is one example of a control check. Manipulation testing allows researchers to isolate key variables to reinforce support that these variables are working as intended.

Some effective designs for evaluating several main effects were found independently and in the near succession of Raja Chandra Bose and K. Kishen in 1940, but remained little known until the Plackett-Burmese designs were published in Biometrics in 1946. About the same time, CR Rao introduced the concept of orthogonal arrays as experimental samples. This concept was central to the development of Taguchi methods by Taguchi, who took place during his visit to the Indian Statistical Institute in the early 1950s. His methods were successfully applied and adopted by Japanese and Indian industry, and subsequently also adopted by American industry, albeit with some reservations.

In 1950 Gertrude Mary Cox and William Gemmell Cochran published the book Experimental Designs, which became the main reference work for the design of experiments on statisticians for many years thereafter.

The development of the theory of linear models has embraced and surpassed the cases that concerned the early authors. Today, theory relies on complex topics in

© Saint Petersburg State University, 2018

© Dermanova I. B., Manukyan V. R., 2018

Introduction

This teaching aid reflects the main content of the first part of the course "Design of psychological research" for undergraduates of the first year of study. Its purpose is to reveal the scientific technology of planning and organizing psychological research. Achieving this goal is impossible without understanding the existing field of psychological research, without knowing their types, features and specific refraction through the prism of psychological science, as well as the general scientific principles on which they are based. These complex issues are presented in the manual in an exclusively applied aspect, which allows applying knowledge from the field of psychology methodology directly in independent research work. Any research has a number of stages: planning, actually conducting the research, presenting its results. The concept of "research design" implies a general organization of the study, including the type and methods of consistently searching for answers to the questions posed by the researcher [Breslav, 2010]. Research design covers the entire process of designing (planning) research and the result of this process. A psychological research project is a document that provides a coherent description of all the main elements of the planned research:

- formulation of the problem;

- formulation of the purpose and hypothesis of the study;

- setting goals;

- a method of forming a sample of subjects;

– choice of methods for collecting empirical material;

- selection of stages of the study;

– choice of methods for analyzing the obtained data;

– interpretation of the obtained results.

Designing a psychological study can be compared in importance to the work of an architect. As K. Hakim noted, “before a building of any scale is built, the initial design stage takes place. Architects are invited to submit their ideas, sometimes competitively, on the form, style and character of the building, taking into account its function, purpose, location, etc.” [cit. Quoted from: Research Design, 2017, p. 5].

When conducting research, a psychologist relies on his knowledge of the phenomena and phenomena that he studies, on those theories and concepts that describe and / or explain human psychology, as well as on the repertoire of data collection and analysis methods known and / or mastered by him. This knowledge gives the researcher a corridor of possibilities, the boundaries of which are largely determined by one or another approach - research design.

According to N. Blakey, research design should answer three main questions: what will be studied, why will it be studied, how will it be studied. The last question can be broken down into four sub-questions: what research strategy will be used, where the data will be obtained from, how the data will be collected and analyzed, when each stage of the research will be carried out.

He identifies eight research design elements that should be considered during the research design phase: research topic/problem; research questions and objectives; research strategies; concepts, theories, hypotheses and models; sources, types and forms of data; selection from data sources; data collection and timing; data processing and analysis .

The manual consistently discusses the preparatory stage of the organization of psychological research. The first chapter deals with the formulation of the research problem and its methodological elements (the object and subject of research, the purpose and objectives of the research, hypotheses); the second deals in detail with the problems of choosing research methods; the third analyzes in detail various sources of empirical evidence and methods for collecting them, and the fourth deals with ethical problems in psychological research.

To consolidate the acquired knowledge, the tasks proposed in the manual for independent work and lists of recommended literature on each of the topics will help.

We hope that the development of this course will allow novice psychologists-researchers to build their work more meaningfully and competently.

Chapter 1
Organization of psychological research

1.1. Psychological research: general idea, main types and stages

Scientific psychology was formed as an independent science in the second half of the 19th century and has come a long way in development, becoming a recognized branch of scientific knowledge. The main goal of scientific psychology is to search for new knowledge, which is achieved through scientific psychological research. The main objectives of research activities in psychology are:

– explanation of psychological phenomena;

- proof of certain theoretical positions (hypotheses);

– forecasting certain psychological facts [Karandyshev, 2004].

Scientific research, including psychological research, is characterized by objectivity, generality, systematicity, conclusiveness, reliance on scientific facts and concepts. Scientific psychological knowledge and research results are usually presented in the language of scientific psychology, understandable for professional psychologists, and are not always described in an accessible way for an unprepared reader.

There are various classifications of psychological research. The most common are dichotomous classifications based on various grounds: theoretical and empirical, fundamental and applied, laboratory and field, quantitative and qualitative research [Druzhinin, 2011; Nikandrov, 2007; Goodwin, 2004 and others].

Target theoretical research– obtaining generalized knowledge about any psychological phenomenon. This study is based on the descriptions and explanations of the facts of mental life already available in science, previously put forward hypotheses and assumptions. In the process of theoretical research, the scientist interacts with the mental model of the research object [Druzhinin, 2011]. Theoretical research involves the analysis, synthesis, comparison and generalization of knowledge existing in science, as well as the derivation of new knowledge based on them with the help of inferences. It is based on a certain system of axioms, theories and empirical knowledge currently existing in a given scientific field, and uses the methods of logical derivation of new knowledge [Karandyshev, 2004]. The result of a theoretical study is presented in the form of more or less harmonious and demonstrative generalizations - hypotheses, concepts, theories. The level of these generalizations differs significantly. A hypothesis is a scientific assumption put forward to explain a phenomenon that requires further experimental testing and / or additional theoretical justification in order to become a reliable scientific theory. A concept is a system of reasoned views that formalizes one or another understanding of the phenomenon under study. Theory is a generalization of experience evidentiary level reflecting the essence of the studied reality. In psychology, we are talking about psychological reality, which includes both objective and subjective facts and patterns [Nikandrov, 2007]. Unlike a concept, a theory is a more strictly structured and substantiated system of theoretical propositions that describes the mechanisms, relationships, and structure of the object under study [Karandyshev, 2004].

empirical research aims to obtain factual material, subsequently either generalized by theoretical studies, or used for applied purposes. In the process of empirical research, the researcher carries out external real interaction with the object of research [Druzhinin, 2011]. In empirical research, they strive to obtain an extremely rigorous description of psychological facts, for which they collect data on the phenomenon under study very carefully. The main methods of empirical psychological research are observation, experiment, testing, questioning, conversation, modeling. Typically, these data are of a massive nature, i.e., they are obtained by multiple calls to the object of study, which increases the reliability of the final results [Nikandrov, 2007].

The dyad "basic - applied research" is formed on the basis of the ratio of the scientific and practical significance of the study. AT fundamental research the scientific significance significantly prevails over the practical one: the results of such studies cannot immediately be directly put into practice, but they contribute to the study of one or another major scientific problem. Fundamental research significantly expands the horizons of the scientific community and, most importantly, “opens up space and paves the way for the organization of narrower specific research of a practical orientation” [Nikandrov, 2007, p. fifteen]. In this regard, they play the role of a foundation both in the general system of knowledge of mankind and in conducting research aimed at obtaining practical results.

Applied Research aimed at obtaining an effect in specific situations of human life. Usually, these studies are carried out on a special order from interested persons or organizations (customers), dictated by the request of practice. Their goal is to solve a specific problem by “applying” known knowledge to it [Nikandrov, 2007]. In these studies, the theoretical and empirical knowledge of science is used, the methods and techniques developed and tested by it are applied. The main thing here is not obtaining new knowledge, but helping the customer in current life and practical affairs.

J. Goodwin also proposes to distinguish between studies according to the conditions of conduct (laboratory and field) and the nature of the methods used (quantitative and qualitative).

Laboratory research provide researchers with a high degree of control: the experimental conditions can be defined more clearly, and the selection and study of subjects can be carried out more systematically. In laboratory research, it is easier to obtain the informed consent of participants, it is relatively easy, unlike in field research, to strictly follow the standards of the code of ethics, while in field research there may be ethical problems associated with intrusion into the privacy of respondents.

Field studies are carried out in the conditions of everyday life and it is the similarity with real life that is their main advantage. J. Goodwin cites their other advantages: firstly, the conditions of field research often cannot be reproduced in the laboratory; second, field studies can validate laboratory studies and correct for errors caused by natural laboratory limitations; thirdly, it is possible to obtain data that can quickly affect the lives of the people being studied [Goodwin, 2004].

AT quantitative research data are collected and presented in the form of numbers - average estimates for various groups, the proportion (in percent) of people who entered one way or another, coefficients reflecting the relationship of various properties, states, processes, etc. At the same time, in modern psychology qualitative research are again in demand. They usually involve the collection of detailed information through interviews with individuals or focus groups, and sometimes include detailed case studies as well as basic observational research. What unites these types of qualitative research is that their results are not presented as statistical reports, but as a general analysis of the project [Goodwin, 2004].

However, it should be noted that the selection of these types of psychological research is still somewhat arbitrary and is rather an abstraction that allows a better understanding of the subject from the point of view of its constituent parts. Thus, many psychological studies are characterized by a combination of theoretical and practical aspects of research in a single process, since “any research is carried out not in isolation, but as part of a holistic scientific program or in order to develop a scientific direction” [Druzhinin, 2011, p. eight]. Theoretical aspects are characteristic of the initial and final stages of the process of psychological research, while empirical aspects are characteristic of the central stage. Conducting applied research is impossible both without the theoretical justification accumulated by fundamental science and without empirical procedures. At the same time, not only fundamental research leads to applied study of the issue, but the results of applied research often turn out to be important for fundamental research, confirming, refuting or setting boundaries for the theories put forward. J. Goodwin also cites cases in which laboratory and field experiments are combined by a common goal into one study, which makes it possible to achieve greater reliability of the results [Goodwin, 2004]. Modern psychology has examples of qualitative and quantitative studies in which the study of patterns in large samples is illustrated and supplemented by qualitative descriptions of mental processes and phenomena.

All psychological research has a certain logic - the sequence of their conduct. Like any scientific research, psychological research goes through three stages: 1) preparatory; 2) main; 3) final.

At the first stage, its goals and objectives are formulated, orientation is made in the totality of knowledge in this area, an action program is drawn up, organizational, material and financial issues are resolved. At the main stage, the actual research process is carried out: the scientist, using special methods, comes into contact (directly or indirectly) with the object under study and collects data about it. It is this stage that usually reflects the specifics of the study to the greatest extent: the reality under study in the form of the object and subject under study, the field of knowledge, the type of study, methodological equipment. At the final stage, the received data is processed and converted into the desired result. The results are correlated with the set goals, explained and included in the existing knowledge system in this area. If these stages are presented in more detail, we get the following scheme of psychological research:

The given sequence of stages should not be considered as a rigid scheme accepted for steady execution.

It is rather a general principle of algorithmization of research activities. Under certain conditions, the order of the stages may change, the researcher may return to the passed stages without completing or even starting the execution of the subsequent ones, individual stages may be partially performed, and some may even fall out. Such freedom in performing stages and operations is provided for in flexible planning of the study [Nikandrov, 2007].

1.2. Preparatory stage of research organization: problem statement

Psychological research, like any other, begins with a problem statement - the discovery of a deficit, a lack of information to describe or explain reality. In the philosophical encyclopedic dictionary, the term “problem” is interpreted as “an issue that objectively arises in the course of the development of knowledge or an integral set of issues, the solution of which is of practical or theoretical interest” [cit. by: Druzhinin, 2011, p. 16]. Thus, it is the lack of knowledge, information, the inconsistency of scientific ideas in social practice or as a result of scientific research that create the conditions for the emergence and formulation of a scientific problem. According to V. N. Druzhinin, “a problem is a rhetorical question that a researcher asks nature, but he must answer it himself” [Druzhinin, 2011, p. 12]. He also highlights the following stages of problem generation: 1) revealing the lack of scientific knowledge about reality; 2) description of the problem at the level of ordinary language; 3) formulating the problem in terms of a scientific discipline. The second stage, according to the scientist, is necessary, since the transition to the level of ordinary language makes it possible to switch from one scientific field (with its own specific terminology) to another and carry out a broader search for possible ways to solve the problem. Thus, already formulating the problem, we narrow the range of the search for its solutions and implicitly put forward a research hypothesis. L. Ya. Dorfman notes that problems are usually found at the intersection of different theories; theoretical provisions and empirical data; all sorts of empirical data; data relating to different populations; data obtained by some methods, and data obtained by other methods, etc. [Dorfman, 2005]. The productivity of future research largely depends on the ability of the scientist to see and formulate the observed contradiction.

Elena Zuchi, a researcher at the University of Milan, gives advice on how to formulate scientific problems - they concern the need to avoid too vague and general problems. Too general problems involve studies that cannot be implemented due to their time duration and breadth. Only problems that can be formulated operationally are subject to scientific analysis [Dzuki, 1997].

Operationalizations of concepts– the exact definition of terms in the formulation of scientific problems is given much attention in research. During the operationalization of concepts, as a rule, an indication is given of the way in which this phenomenon can be measured. J. Goodwin emphasizes that this is especially important in psychological research, where concepts are used for which many definitions can be given. The accuracy of operational definitions has another important consequence – they provide the possibility of reproducing experiments [Goodwin, 2004].

The process of developing and formulating a research problem is impossible without getting acquainted with publications on this topic and exchanging information with colleagues involved in this field. Usually, scientific research is preceded by a presentation of such an acquaintance with the problem in the form of a literature review. As L. V. Kulikov rightly notes, “you can convince your future reader that the problem really exists, based on your literary review” [Kulikov, 2001, p. eleven]. It characterizes the degree of research of the problem both in general and its individual aspects. Unexplored and little-studied issues, contradictions in understanding the phenomenon as a whole and its individual aspects, contradictions in the available empirical data are highlighted.

As a result of bibliographic preparation, the researcher needs to have an idea:

- about the number of publications on the topic of interest to him;

– about the time frame of publications;

– about the interest of scientists to this problem;

It is better to start reading with the most famous and cited authors and those who made the greatest contribution to the initial period of studying the problem - this way it will be easier to understand the content of further works.

The construction of a literature review can be chronological or logical. With the exception of certain topics where the subject is the history of the study of a phenomenon, a logical presentation of the material is preferable, since it allows to reveal and substantiate the research problem to a greater extent.

The structure of a theoretical review might be something like this:

1. Essence, nature of the phenomenon. Available definitions of this phenomenon. Characterization of the degree of diversity in its understanding by various authors.

2. Phenomenological description (description of manifestations) - the area of ​​manifestations, the frequency of manifestations, temporal, spatial, intensity, modality (if the studied phenomena have them) characteristics.

3. The structure of a phenomenon is a stable relationship between its components. In psychology, structure is most often understood as a functional structure, that is, stable relationships between individual functions. The consideration should be based on a scheme corresponding to the chosen approach (systemic, holistic, complex, environmental, situational, etc.).

4. The place of this phenomenon among other mental phenomena - its interrelationships, mutual influences (factors that determine it and the phenomena that it influences).

5. Patterns that govern the phenomenon [Kulikov, 2001].

The construction of a literature review depends on the specifics of the mental phenomenon under consideration, its study and many other factors, so the proposed scheme cannot always and must be followed.

In the literature review, the names of the authors whose ideas or experimental results are recounted or generalized must be mentioned. Reference should be made to sources by indicating the specific publications of the authors or publications of intermediaries, thanks to which the necessary information became available [Kulikov, 2001].

In addition, in the process of developing a research problem and preparing a literature review, special attention is paid to substantiating the relevance and novelty of the study.

Relevance research can be characterized from practical and scientific points of view. Practical relevance is determined both by the need to search for new knowledge to solve a practical problem, and by the importance of developing a system or method of practical psychological work to solve certain problems. The scientific relevance can be judged by the lack of certain knowledge, research methods in the relevant field of scientific psychology, by the need to solve a specific scientific problem.

At the first stage, the design is carefully worked out (from the English. design- creative idea) of future research.

First of all, a research program is developed.

Program includes the topic, purpose and objectives of the study, formulated hypotheses, definition of the object of study, units and scope of observations, glossary of terms, description of statistical methods for forming a sample, collecting, storing, processing and analyzing data, methodology for conducting a pilot study, a list of statistical tools used .

Name Topics usually formulated in one sentence, which should correspond to the purpose of the study.

Purpose of the study- this is a mental anticipation of the result of an activity and ways to achieve it with the help of certain means. As a rule, the purpose of medical and social research is not only theoretical (cognitive), but also practical (applied) in nature.

To achieve this goal, determine research objectives, that reveal and detail the content of the goal.

The most important component of the program are hypotheses (Expected results). Hypotheses are formulated using specific statistical indicators. The main requirement for hypotheses is the ability to test them in the research process. The results of the study can confirm, correct or refute the hypotheses put forward.

Prior to the collection of material, the object and unit of observation are determined. Under object of medical and social research understand a statistical set consisting of relatively homogeneous individual objects or phenomena - units of observation.

Unit of observation- the primary element of the statistical population, endowed with all the features to be studied.

The next important operation in the preparation of the study is the development and approval of the work plan. If the research program is a kind of strategic plan that embodies the ideas of the researcher, then the work plan (as an annex to the program) is a mechanism for the implementation of the study. The work plan includes: the procedure for selecting, training and organizing the work of direct executors; development of regulatory and methodological documents; determination of the required volume and types of resource support for the study (personnel, finance, material and technical, information resources, etc.); definition of terms and responsible for separate stages of research. It is usually presented in the form network graphics.

At the first stage of medical and social research, it is determined by what methods the selection of units of observation will be carried out. Depending on the volume, continuous and selective studies are distinguished. In a continuous study, all units of the general population are studied, in a selective study, only a part of the general population (sample) is studied.

General population called a set of qualitatively homogeneous units of observation, united by one or a group of features.

Sample population (sample)- any subset of observation units of the general population.

The formation of a sample population that fully reflects the characteristics of the general population is the most important task of statistical research. All judgments about the general population based on sample data are valid only for representative samples, i.e. for such samples, the characteristics of which correspond to those of the general population.

The real provision of representativeness of the sample is guaranteed random selection method those. such a selection of units of observation in the sample, in which all objects in the general population have the same chances of being selected. To ensure random selection, specially developed algorithms are used that implement this principle, either tables of random numbers, or a random number generator available in many computer software packages. The essence of these methods is to randomly indicate the numbers of those objects that must be selected from the entire population in some way ordered. For example, the general population "the population of the region" can be sorted by age, place of residence, alphabetical order (last name, first name, patronymic), etc.

Along with random selection, when organizing and conducting medical and social research, the following methods of forming a sample are also used:

Mechanical (systematic) selection;

Typological (stratified) selection;

serial selection;

Multistage (screening) selection;

cohort method;

The "copy-pair" method.

Mechanical (systematic) selection allows you to form a sample using a mechanical approach to the selection of units of observation of an ordered general population. At the same time, it is necessary to determine the ratio of the volumes of the sample and the general population and thereby establish the proportion of selection. For example, in order to study the structure of hospitalized patients, a sample of 20% of all patients who left the hospital is formed. In this case, among all the "medical records of an inpatient" (f. 003 / y), ordered by numbers, every fifth card should be selected.

Typological (stratified) selection involves a breakdown of the general population into typological groups (strata). When conducting medical and social research, age-sex, social, professional groups, individual settlements, as well as urban and rural populations are taken as typological groups. In this case, the number of units of observation from each group is selected in the sample randomly or mechanically in proportion to the size of the group. For example, when studying the cause-and-effect relationships of risk factors and oncological morbidity of the population, the study group is first divided into subgroups by age, gender, profession, social status, and then the required number of observation units is selected from each subgroup.

serial selection the sample is formed not from individual units of observation, but from whole series or groups (municipalities, health care institutions, schools, kindergartens, etc.). The selection of series is carried out using proper random or mechanical sampling. Within each series, all units of observation are studied. This method can be used, for example, to evaluate the effectiveness of the immunization of the child population.

Multistage (screening) selection involves a phased sampling. By the number of stages, one-stage, two-stage, three-stage selection, etc. are distinguished. So, for example, when studying the reproductive health of women living in the territory of a municipality, at the first stage, working women are selected, who are examined using basic screening tests. At the second stage, a specialized examination of women with children is carried out, at the third stage, an in-depth specialized examination of women with children with congenital malformations. Note that in this case of purposeful selection for a certain attribute, the sample includes all objects - carriers of the studied attribute on the territory of the municipality.

cohort method are used to study the statistical population of relatively homogeneous groups of people united by the onset of a certain demographic event in the same time interval. For example, when studying issues related to the problem of fertility, a population (cohort) is formed that is homogeneous on the basis of a single date of birth (a study of fertility by generations) or on the basis of a single age of marriage (a study of fertility by length of family life).

Copy-Pair Method provides for the selection for each unit of observation of the group under study of an object that is similar in one or more features (“copy-pair”). For example, factors such as body weight and sex of the child are known to influence infant mortality rates. When using this method, for each death of a child under 1 year of age, a “copy-pair” of the same sex, similar in age and body weight, is selected from among living children under the age of 1 year. This method of selection should be used to study risk factors for the development of socially significant diseases, individual causes of death.

At the first stage, research is also developed (ready-made is used) and replicated statistical toolkit (cards, questionnaires, table layouts, computer programs for controlling incoming information, forming and processing information databases, etc.), into which the studied information will be entered.

In the study of public health and the activities of the health care system, sociological research is often used using special questionnaires (questionnaires). Questionnaires (questionnaires) for medical and sociological research should be targeted, oriented, ensure the reliability, reliability and representativeness of the data recorded in them. During the development of questionnaires and interview programs, the following rules must be observed: the suitability of the questionnaire for collecting, processing and extracting the necessary information from it; the possibility of revising the questionnaire (without violating the system of codes) in order to eliminate unsuccessful questions and make appropriate adjustments; explanation of the goals and objectives of the study; clear wording of questions, eliminating the need for various additional explanations; fixed nature of most questions.

Skillful selection and combination of various types of questions - open, closed and semi-closed - can significantly increase the accuracy, completeness and reliability of the information received.

The quality of the survey and its results largely depend on whether the basic requirements for the design of the questionnaire and its graphic design are met. There are the following basic rules for constructing a questionnaire:

The questionnaire includes only the most significant questions, the answers to which will help to obtain the information necessary to solve the main objectives of the study, which cannot be obtained in any other way without conducting a questionnaire survey;

The wording of the questions and all the words in them should be understandable to the respondent and correspond to his level of knowledge and education;

The questionnaire should not contain questions that cause unwillingness to answer them. It should be strived to ensure that all questions cause a positive reaction of the respondent and a desire to give complete and true information;

The organization and sequence of questions should be subject to obtaining the most necessary information to achieve the goal and solve the problems set in the study.

Special questionnaires (questionnaires) are widely used, among other things, to assess the quality of life of patients with a particular disease, the effectiveness of their treatment. They allow capturing changes in the quality of life of patients that have occurred over a relatively short period of time (usually 2-4 weeks). There are many special questionnaires, such as AQLQ (Asthma Quality of Life Questionnaire) and AQ-20 (20-Item Asthma Questionnaire) for bronchial asthma, QLMI (Quality of Life after Myocardial Infarction Questionnaire) for patients with acute myocardial infarction, etc.

Coordination of work on the development of questionnaires and their adaptation to various linguistic and economic formations is carried out by an international non-profit organization for the study of the quality of life - the MAPI Institute (France).

Already at the first stage of the statistical study, it is necessary to draw up layouts of tables, which will later be filled with the data obtained.

In tables, as in grammatical sentences, the subject is distinguished, i.e. the main thing that is said in the table, and the predicate, i.e. that which characterizes the subject. Subject - this is the main feature of the phenomenon under study - it is usually located on the left along the horizontal lines of the table. Predicate - signs characterizing the subject are usually located on top of the vertical columns of the table.

When compiling tables, certain requirements are met:

The table should have a clear, concise title that reflects its essence;

The design of the table ends with the totals for columns and rows;

The table should not contain empty cells (if there is no sign, put a dash).

There are simple, group and combination (complex) types of tables.

A simple table is a table that presents a summary of data for only one attribute (Table 1.1).

Table 1.1. Simple table layout. Distribution of children by health groups, % of total

In the group table, the subject is characterized by several unrelated predicates (Table 1.2).

Table 1.2. Group table layout. Distribution of children by health groups, sex and age, % of the total

In the combination table, the signs characterizing the subject are interconnected (Table 1.3).

Table 1.3. Combination table layout. Distribution of children by health groups, age and gender, % of total

An important place in the preparatory period is occupied by pilot Study, the task of which is to test statistical tools, to verify the correctness of the developed methodology for collecting and processing data. The most successful is such a pilot study, which repeats on a reduced scale the main one, i.e. makes it possible to check all the upcoming stages of work. Depending on the results of the preliminary analysis of the data obtained during piloting, the statistical tools, methods of collecting and processing information are adjusted.

Experimental psychology is based on the practical application of the plans of the so-called true experiment, when control groups are used in the course of the study, and the sample is in laboratory conditions. The schemes of experiments of this kind are designated as plans 4, 5 and 6.

Plan with pre-test and post-test and control group (Plan 4). Scheme 4 is a classic "design" of a psychological laboratory study. However, it is also applicable in the field. Its peculiarity lies not only in the presence of a control group - it is already present in the pre-experimental scheme 3 - but in the equivalence (homogeneity) of the experimental and control samples. An important factor in the reliability of the experiment, built according to scheme 4, are also two circumstances: the homogeneity of the research conditions in which the samples are located, and the full control of factors affecting the internal validity of the experiment.

The choice of an experiment plan with preliminary and final testing and a control group is made in accordance with the experimental task and the conditions of the study. When it is possible to form at least two homogeneous groups, the following experimental scheme is applied:

Example. For practical assimilation of the possibilities of implementing experimental plan 4, we will give an example of a real study in the form of a laboratory formative experiment, which contains a mechanism for confirming the hypothesis that positive motivation affects the concentration of a person's attention.

Hypothesis: the motivation of the subjects is a significant factor in increasing the concentration and stability of the attention of people who are in the conditions of educational and cognitive activity.

Experiment procedure:

• 1. Formation of experimental and control samples. Participants in the experiment are divided into pairs, carefully equalized by indicators of preliminary testing or by variables that are significantly correlated with each other. The members of each nara are then "randomly" (randomized) drawn by lot into the experimental or control groups.
• 2. Both groups are invited to work out the test "Correction test with rings" (O, and 0 3).
• 3. The activity of the experimental sample is stimulated. Suppose that the subjects are given an experimental stimulating installation (X): “Students who score 95 or more points (correct answers) on the basis of concentration and attention stability testing will receive an “automatic” credit this semester.
• 4. Both groups are invited to work out the test "Correction test with syllables" (0 2 and OD

Algorithm for analyzing the results of the experiment

• 5. Empirical data are tested for "normality" of distribution 1 . This operation makes it possible to find out at least two circumstances. Firstly, as a test used to determine the stability and concentration of the subjects' attention, it discriminates (differentiates) them according to the measured attribute. In this case, the normal distribution shows that the indicators of the features correspond to the optimal ratio with the development situation of the applied test, i.e. the technique optimally measures the intended area. It is suitable for use in these conditions. Secondly, the normality of the distribution of empirical data will give the right to correctly apply the methods of parametric statistics. Statistics can be used to estimate the distribution of data A s and E x or at .
• 6. The calculation of the arithmetic mean M x and root-mean-square 5 L. deviations of the results of preliminary and final testing.
• 7. A comparison is made of the average values ​​of test indicators in the experimental and control groups (O, 0 3 ; Oh OD
• 8. The average values ​​are compared according to Student's t-test, i.e. determination of statistical significance of differences in mean values.
• 9. The proof of the ratios Oj = O e, O, 0 4 as indicators of the effectiveness of the experiment is being carried out.
• 10. A study of the validity of the experiment is carried out by determining the degree of control of factors of invalidity.

To illustrate a psychological experiment on the influence of motivational variables on the process of concentration of attention of the subjects, let us turn to the data placed in Table. 5.1.

Table of experimental results, points

Table 5.1

The end of the table. 5.1

Subjects		Measurement before exposure X		Measurement after exposure X
experimental	Control group	experimental	Control group 0 3	Experimental group 0 2	Control group 0 4

Comparison of the data of the primary measurement of the experimental and control samples - Oh! and O3 - is made in order to determine the equivalence of the experimental and control samples. The identity of these indicators indicates the homogeneity (equivalence) of the groups. It is determined by calculating the level of statistical significance of the differences in the means in the confidence interval R t-test Styodeita.

In our case, the Studentent /-criterion value between the empirical data of the primary examination in the experimental and control groups was 0.56. This shows that the samples do not differ significantly in the confidence interval/?

Comparison of the data of the primary and repeated measurements of the experimental sample - Oj and 0 2 - is carried out in order to determine the degree of change in the dependent variable after the influence of the independent variable on the experimental sample. This procedure is carried out using the /-Styodeit test if the variables are measured on the same test scale or are standardized.

In this case, the preliminary (primary) and final examinations were carried out using different tests that measure the concentration of attention. Therefore, comparison of averages without standardization is not feasible. Let's calculate the correlation coefficient between the indicators of the primary and final studies in the experimental group. Its low value can serve as an indirect evidence that there is a data change. (Rxy = 0D6) .

The experimental effect is determined by comparing the re-measurement data of the experimental and control samples - 0 2 and 0 4 . It is performed in order to determine the degree of significance of the change in the dependent variable after exposure to the independent variable. (X) for the experimental sample. The psychological meaning of this study is to assess the impact X on the test subjects. In this case, the comparison is made at the stage of the final measurement of the data of the experimental and control groups. Impact Analysis X carried out with the help of /-Stuodent's criterion. Its value is 2.85, which is more than the tabular value of the /-criterion 1 . This shows that there is a statistically significant difference between the mean test values ​​in the experimental and control groups.

Thus, as a result of the experiment according to plan 4, it was revealed that in the first group of subjects, which does not differ from the other group in terms of setting psychological characteristics (in terms of concentration of attention), except for the impact of the independent variable on it x, the value of the indicator of concentration of attention is statistically significantly different from the similar indicator of the second group, which is in the same conditions, but outside the influence x.

Consider the study of the validity of the experiment.

Background: controlled due to the fact that events occurring in parallel with the experimental exposure are observed in both the experimental and control groups.

Natural development: controlled due to short inter-test and exposure periods and occurs in both experimental and control groups.

Test effect and instrumental error: are controlled because they appear in the same way in the experimental and control groups. In our case, there is a sample bias of 1.

Statistical regression: controlled. First, if randomization led to the appearance of extreme results in the experimental group, then they will also appear in the control group, as a result of which the regression effect will be the same. Secondly, if randomization did not lead to the appearance of extreme results in the samples, then this question is removed by itself.

Selection of test subjects: controlled because explanation of differences is ruled out to the extent that randomization provides equivalence of samples. This degree is determined by the sample statistics we have adopted.

Screening: controlled completely, since the period between tests in both samples is relatively small, and also through the need for the presence of the test subjects at the lesson. In experiments with a long exposure period (the period between tests), a bias in the sample and the effect of the results of the experiment is possible. The way out of this situation is to take into account, when processing the results of the preliminary and final testing data, all participants in both samples, even if the subjects of the experimental group did not receive experimental exposure. Effect x, will apparently be weakened, but there will be no sampling bias. The second way out entails changing the design of the experiment, since it is necessary to achieve equivalence of groups by randomization before the final testing:

The interaction of the selection factor with natural development: controlled by forming a control equivalent group.

Reactive effect: pre-testing really sets the subjects up to perceive the experimental impact. Therefore, the effect of exposure is "shifted". It is unlikely that in this situation one can absolutely assert that the results of the experiment can be extended to the entire population. Reactive effect control is possible to the extent that repetitive examinations are characteristic of the entire population.

Interaction of selection factor and experimental influence: in a situation of voluntary consent to participate in the experiment, there is a threat of validity (“bias”) due to the fact that this consent is given by people of a certain personality type. The formation of equivalent samples in a random order reduces invalidity.

The reaction of the subjects to the experiment: the situation of the experiment leads to a bias in the results, as the subjects fall into "special" conditions, trying to understand the meaning of this work. Hence, manifestations of demonstrativeness, games, alertness, guessing attitudes, etc. are frequent. Any element of the experimental procedure can elicit a reaction to an experiment, such as the content of the tests, the randomization process, dividing the participants into separate groups, keeping the subjects in different rooms, the presence of strangers, the use of an extraordinary X etc.

The way out of this difficulty is to "mask" the study, i.e. drawing up and strict adherence to a system of legending experimental procedures or their inclusion in the usual course of events. To this end, it seems most rational to conduct testing and experimental exposure under the guise of regular verification activities. In the study of even individual members of the group, it is desirable to participate in the experiment of the team as a whole. It seems expedient to carry out testing and experimental influence by staff leaders, teachers, activists, observers, etc.

In conclusion, it should be noted that, as D. Campbell pointed out, “common sense” and “considerations of a non-mathematical nature” can still be the optimal method for determining the effect of an experiment.

R. Solomon's plan for four groups (plan 5). In the presence of certain research conditions that allow the formation of four equivalent samples, the experiment is built according to scheme 5, which was named after its compiler - "Solomon's plan for four groups":

Solomon's plan is an attempt to compensate for factors that threaten the external validity of the experiment by adding to the experiment two additional (to plan 4) groups that are not pre-measured.

Comparison of data for additional groups neutralizes the impact of testing and the influence of the experimental setting itself, and also allows for a better generalization of the results. Identification of the effect of experimental exposure is reproduced by statistical proof of the following inequalities: 0 2 > Oj; 0 2 > 0 4 ; 0 5 > About b. If all three relations are satisfied, then the validity of the experimental conclusion much increases.

The use of design 5 determines the probability of neutralizing the interaction of testing and experimental exposure, which facilitates the interpretation of the results of studies according to design 4. Comparison of Ob with O, and 0 3 reveals the combined effect of natural development and background. Comparison of means 0 2 and 0 5 , 0 4 and 0 0 makes it possible to estimate the main effect of preliminary testing. Comparison of the averages () 2 and 0 4 , 0 5 and 0 D) makes it possible to estimate the main effect of the experimental exposure.

If the pre-test effect and the interaction effect are small and negligible, then it is desirable to perform a covariance analysis of 0 4 and 0 2 using the pre-test results as a covariate.

Plan with control group and testing only after exposure (plan 6). Very often, when performing experimental tasks, researchers are faced with the situation of the need to study psychological variables in the conditions of the impossibility of conducting a preliminary measurement of the psychological parameters of the subjects, since the study is carried out after exposure to independent variables, i.e. when an event has already occurred and its consequences need to be identified. In this situation, the optimal design of the experiment is a plan with a control group and testing only after exposure. Using randomization or other procedures that provide optimal selective equivalence, homogeneous experimental and control groups of subjects are formed. Testing of variables is carried out only after experimental exposure:

Example. In 1993, by order of the Research Institute of Radiology, a study was made of the effect of radiation exposure on the psychological parameters of a person 1 . The experiment was built according to plan 6. A psychological examination of 51 liquidators of the consequences of the accident at the Chernobyl nuclear power plant was carried out using a battery of psychological tests (personality questionnaires, SAN (Health. Activity. Mood), Luscher test, etc.), EAF according to R. Voll (R. Voll) and the automated situational diagnostic game (ASID) "Test". The control sample consisted of 47 specialists who did not participate in radiological activities at the Chernobyl nuclear power plant. The average age of the subjects of the experimental and control groups was 33 years. The subjects of both samples were optimally correlated in terms of experience, type of activity and structure of socialization, therefore the formed groups were recognized as equivalent.

Let us make a theoretical analysis of the plan according to which the experiment was built, and its validity.

Background: controlled because the study used an equivalent control sample.

natural development: controlled as a factor of experimental influence, since there was no interference of experimenters in the process of socialization of the subjects.

Test effect: controlled, since there was no pre-testing of the subjects.

Instrumental error: it was controlled, since a preliminary check of the reliability of methodological tools and clarification of their normative indicators after the experiment was carried out, and the use of the same type of “test battery” was carried out on the control and experimental groups.

Statistical regression: was controlled by working out the experimental material on the entire sample, formed at random. However, there was a threat to validity due to the fact that there were no preliminary data on the composition of the experimental groups, i.s. the probability of occurrence and polar variables.

Selection of test subjects, was not fully controlled due to natural randomization. Special selection of subjects was not carried out. In a random order, groups were formed from the participants in the liquidation of the accident at the Chernobyl nuclear power plant and chemical specialists.

screening of test subjects, was not present during the experiment.

Interaction of the selection factor with natural development", no special selection was made. This variable was controlled.

Interaction of composition of groups and experimental influence", Special selection of subjects was not carried out. They were not informed which study group (experimental or control) they were in.

The reaction of the subjects to the experiment, uncontrollable factor in this experiment.

Mutual interference (superposition) of experimental influences: was not controlled due to the fact that it was not known whether the subjects participated in such experiments and how this affected the results of psychological testing. By observing the experimenters, it turned out that, in general, the attitude towards the experiment was negative. It is unlikely that this circumstance had a positive effect on the external validity of this experiment.

Experiment results

• 1. A study was made of the distribution of empirical data, which had a bell-shaped form, close to the theoretical normal distribution curve.
• 2. Using the Student's t-test, the averages Oj > 0 2 were compared. According to ASID "Test" and EAF, the experimental and control groups differed significantly in the dynamics of emotional states (in the liquidators - higher), the effectiveness of cognitive activity (in the liquidators there was a decrease), as well as the functioning of the motor apparatus, liver, kidneys, etc. due to chronic endogenous intoxication.
• 3. Using Fisher's ^-criterion, the influence of "fluctuations" (dispersion of the independent variable) was calculated X on the variance of the dependent variable 0 2 .

As a conclusion of this study, appropriate recommendations were made to the participants in the experiment and their leaders, the diagnostic battery of psychological tests was validated, and psychophysiological factors that affect people in extreme radiological conditions were identified.

Thus, the experimental "design" 6 represents the optimal scheme for psychological research when it is not possible to make a preliminary measurement of psychological variables.

It follows from the foregoing that the basis of the experimental method in psychology is the so-called true plans, in which almost all the main factors affecting internal validity are controlled. The reliability of the results in experiments designed according to Schemes 4-6 does not raise doubts among the vast majority of researchers. The main problem, as in all other psychological studies, is the formation of experimental and control samples of subjects, the organization of the study, the search for and use of adequate measuring instruments.

• The symbol R in the scheme indicates that the homogeneity of the groups was obtained by randomization. This symbol can be conditional, since the homogeneity of the control and experimental samples can be ensured in other ways (for example, pairwise selection, preliminary testing, etc.). The value of the correlation coefficient (0.16) reveals a weak statistical relationship between measurements, i.e. it can be assumed that there has been some change in the data. Post-exposure readings do not match pre-exposure readings. EAF - Voll's method (German: Elektroakupunktur nach Voll, EAV) - a method of electrified rapid diagnostics in alternative (alternative) medicine by measuring the electrical resistance of the skin. The method was developed in Germany by Dr. Reinold Voll in 1958. In essence, it is a combination of acupuncture and the use of a galvanometer.
• Assessment of the psychological status of military personnel - liquidators of the Chernobyl accident using the dynamic situational game "Test" / I. V. Zakharov, O. S. Govoruha, I. II. Poss [et al.] // Military Medical Journal. 1994. No. 7. S. 42-44.
• Research B. II. Ignatkin.