What is a scientific hypothesis? | Defining a Hypothesis

Hypothesis within psychological research

If we are talking about psychological scientific research based on the methods of mathematical statistics, then the hypothesis in this case must first of all meet such requirements as clarity and conciseness. As noted by E.V. Sidorenko, thanks to these hypotheses, the researcher, in the course of calculations, actually gets a clear picture of what he has established.

It is customary to distinguish between the null and alternative statistical hypotheses. In the first case, we are talking about the absence of differences in the studied characteristics, according to the formula X1-X2=0. In turn, X1, X2 are the values of the characteristics by which the comparison is carried out. Accordingly, if the goal of our research is to prove the statistical significance of the differences between the values of the features, then we want to refute the null hypothesis.

In the case of an alternative hypothesis, the statistical significance of the differences is stated. Thus, the alternative hypothesis is the statement that we seek to prove. It is also called the experimental hypothesis. It should be noted that in some cases, the researcher may, on the contrary, seek to prove the null hypothesis if this corresponds to the goals of his experiment.

The following examples of hypotheses in psychology can be given:

Null hypothesis (H): The tendency of an increase (decrease) of a characteristic when moving from one sample to another is random.

Alternative hypothesis (H1): The tendency of an increase (decrease) of a characteristic when moving from one sample to another is not random.

Let's assume that a series of trainings were conducted in a group of children with high levels of anxiety to reduce this anxiety. Measurements of this indicator were made before and after the trainings, respectively. It is necessary to determine whether the difference between these measurements is statistically significant. The null hypothesis (N) will have the following form: the tendency for a decrease in the level of anxiety in the group after the training is random. In turn, the alternative hypothesis (H1) will sound like: the tendency for a decrease in the level of anxiety in the group after the training is not accidental.

After applying one or another mathematical criterion (for example, the G-sign criterion), the researcher can draw a conclusion about the statistical significance / insignificance of the resulting “shift” in relation to the characteristic being studied (anxiety level). If the indicator is statistically significant, the alternative hypothesis is accepted, and the null hypothesis is accordingly rejected. Otherwise, the null hypothesis is accepted.

Also in psychology, there may be an identification of a connection (correlation) between two or several variables, which is also reflected by the research hypothesis. Example:

N: the correlation between the student’s concentration indicator and the indicator of his success in completing the control task does not differ from 0.

H1: the correlation between the student’s concentration indicator and the indicator of his success in completing the control task is statistically significantly different from 0.

In addition, examples of scientific hypotheses in psychological research that require statistical confirmation may relate to the distribution of a trait (empirical and theoretical level), the degree of consistency of changes (when comparing two traits or their hierarchies), etc.

Conclusion about acceptance or rejection of the main hypothesis

If the criterion value found on sample observation values belongs to the area of acceptance of the hypothesis, it is concluded that it is not possible to reject the main hypothesis.

If the criterion belongs to the critical region, it is concluded that it is not possible to accept the main hypothesis. In this case, the alternative hypothesis is accepted.

In the figure below, the axis of all possible values of the criterion R is shown in blue; other symbols illustrate how the criterion value falls into the area of acceptance of the hypothesis or the critical area.

In order to be more confident in their judgments about the correctness or incorrectness of a hypothesis, they prefer to apply several criteria and compare the results. If in several cases the criterion could not fall within the range of acceptance of the hypothesis, then they say that they have obtained a consistent result and most likely the hypothesis is false.

Types of hypotheses

When considering hypotheses, their types are identified, based on different classification principles. The main difference between hypothetical assumptions is determined by the cognitive functions presented, and also classified according to the object of study. According to cognitive functions, subtypes are distinguished: descriptive hypothesis and explanatory. Descriptive refers to the properties that are characteristic of an object, its structure, composition, and functioning features.

The descriptive one can also concern the existence of something (existential hypothesis), an example of such conclusions is the idea of the existence and possible location of Atlantis.

The explanatory type of hypothesis considers the mechanism and conditionality of the occurrence of an object, natural phenomenon or designated research events.

If you trace the historical chronology of the emergence of the described types of hypotheses, you will notice a characteristic logical pattern. Initially, in the course of scientific interest in a certain chosen field, conjectures of the existential spectrum arise. Subject to proof of the existence of something, descriptive hypotheses arise that study objects that exist in reality and their properties, and only then explanatory hypothetical assumptions arise that seek to clarify the mechanisms of formation and emergence. With further study of the object, the hypotheses become more complex and detailed.

Depending on the characteristics and scale of the object of study, general (this includes patterns of connection between natural and social phenomena, the functioning of the psyche, which have planetary confirmation) and specific (properties of specific individual manifestations, events, a selected separate group of objects, parts of the psyche) hypothetical conclusions are distinguished.

At the initial stages of constructing a study, a working hypothesis is formulated (the main one will be developed later), which is a conditional formulation, with the presence and help of which it is possible to collect and systematize primary data. With further analysis of the results obtained, the working hypothesis may remain and take a stable form, or undergo adjustments due to incompatibility with the facts discovered during the study.

Based on the type of origin, hypotheses are divided into:

— hypotheses based on reality (to confirm the relevance of a certain theoretical model);

— scientific and experimental (establishing the determination of various patterns);

— empirical (were formulated for a specific case and cannot be used for mass explanation);

— experimental hypotheses (necessary for organizing the experiment and actual confirmation);

— statistical hypotheses (necessary for comparing the parameters involved and influencing the reliability).

Statistical hypotheses and areas of their application

A statistical hypothesis is an assumption about some characteristics of a random variable. For example: is the change in the number of AI startups in Europe significant in two different years, etc.

Testing statistical hypotheses is the most important class of problems in mathematical statistics. Using this tool, you can confirm or reject assumptions about the properties of a random variable by applying statistical analysis methods to sample elements. If any terms in the previous sentence are not entirely clear, an explanation in plain language can be found below.

A random variable is a quantity that, depending on a particular situation, takes on specific values with certain probabilities. Examples: exam grade; result of dice game; number of AI startups by European countries. In general, almost anything!

The general population is the collection of all objects for analysis. For example: all AI startups in Europe in 2019.

A sample is a piece of data from a population. For example: officially registered AI startups in some European countries in 2019.

Statistical analysis is the use of various methods to determine the properties of a population from a sample.

Types of assumptions and requirements for them

The concept of hypothesis includes two types: primary and scientific. The first establishes the objectives and conditions of the study. It is primarily used when the area is little studied. Such an argument only helps to select and systematize information, and the conclusions drawn on its basis form the conditions for further study and contribute to the creation of a real hypothesis.

A scientific hypothesis is based on theory and is more specific than the primary hypothesis. This is an argument about the real connection between phenomena and their justifications, about the presence of a particular phenomenon, quality or consequence. However, the difference between primary and scientific hypotheses is conditional - during study, one transforms into the other.

There are other types of hypotheses. Depending on the content, the assumption can be descriptive, explanatory and theoretical. The first is typical for experimental studies. A descriptive judgment is simply a consideration of the practical relationship between an effect and its consequences, a statement of reasons and intended conclusions. It leads to the assumption that one method will be more effective than another, but does not explain its operation.

An explanatory hypothesis differs in that it reveals the causes of reality and includes their intended conclusions, and also describes the criteria under which these conclusions are inevitable. A theoretical judgment is an assumption of the natural nature of a statement that is proven in the process of research. It requires a series of actions indicating that the interaction between causes is natural.

A hypothesis as a scientific assumption must meet certain requirements:

It cannot contain several theses.
The argument should not include concepts and judgments that have an ambiguous meaning and are not explained by the researcher.
When presenting an assumption, one should not use value judgments. A hypothesis must be supported by facts, tested and applied to a wide range of realities.
The judgment must be impeccably stylistically designed, understandable, and logical.
It must correspond to the topic, objectives and subject of the study. Often interesting assumptions are unnaturally tied to the topic.

The argument should not lead away from the topic. If many new facts have emerged during the study, it is preferable to develop an assumption rather than include in advance provisions that will take a lot of time to prove and ultimately will not be confirmed.
A judgment must respond to indisputable facts, interpret them, and identify new ones. The advantage goes to the assumption that explains the most facts equally convincingly.
A hypothesis cannot contradict proven theories. If the assumption still diverges from some of them, but extends to a wider range of phenomena, then the old theories become its special case.
The judgment must include a way to solve the problem in order to be part of the study.

Summarizing

The theory of abiogenesis says that life arose spontaneously under certain conditions from nonliving matter. How did scientists come to this hypothesis? Experiments show that organic molecules, RNA nucleotides and cell membranes can form spontaneously under conditions similar to early Earth. RNA is a nucleic acid that can act as a genetic library and catalyze reactions, so it is likely that the first life on Earth was able to function solely on RNA. Over millions of years, these RNA molecules developed new catalytic abilities and eventually evolved into the complex cells we know today.

Requirements for the hypothesis:

compatibility with existing knowledge, fundamental scientific principles, previously established facts;
clarity and consistency, absence of double interpretation;
validity (relevance), that is, the consistency of the proposed theory verified by analysis;
it must be verifiable (by observation, measuring instruments, experimental setups and other reliable available means).

The standard structure of a hypothesis consists of two parts: an empirical basis (premise) and an assumption based on it (conclusion). Its nomination is the result of extensive work, which includes studying theoretical foundations, collecting material, analyzing it, conducting experiments and observations. Main stages of preparation:

accumulation of material, assumptions, guesses about the object or phenomenon being studied;
formulating consequences arising from the hypothetical theory, putting forward preliminary answers and solutions to the problem posed;
refuting assumptions that turned out to be untenable, replacing them with reliable ones that correspond to the received factual data;
verification of the conclusions made in practice.

To confirm (or, conversely, refute) a hypothesis, it is necessary to follow the rules of logic. Thus, the conclusion (thesis or antithesis) must be precise and clear, unchanged during the research process. Only true facts that have already been established earlier are accepted as grounds (arguments).

The argumentation must be sufficient to formulate a final conclusion. If testing shows that the scientist’s presumptive statement is true, the hypothesis receives the status of a scientific theory, which requires further study and research, and comprehensive development.

A refutation of the hypothesis, justified by the falsity of its conclusion, is also possible. In this case, they go through falsification, establishing the discrepancy between the facts arising from the assumption and the consequences, or by proving the antithesis (opposite to the hypothesis of the consequence). If the antithesis is proven, logically this means the inconsistency (falsity) of the original thesis.

Concept and types of hypotheses. Version Concept of hypothesis

Reliable knowledge in a scientific or practical field is always preceded by a rational understanding and assessment of the factual material provided by observation. This mental activity is accompanied by the construction of various kinds of guesses and conjectural explanations of observed phenomena. At first the explanations are problematic. Further research amends these explanations. As a result, science and practice overcome numerous deviations, misconceptions and contradictions and achieve objectively true results.

The decisive link in the cognitive chain that ensures the formation of new knowledge is a hypothesis.

A hypothesis is a natural form of development of knowledge, which is a reasonable assumption put forward in order to clarify the properties and causes of the phenomena under study.

The most important among those noted in the definition will be the following characteristic features of the hypothesis.

(1) A hypothesis is a universal and necessary form of knowledge development for any cognitive process. Where there is a search for new ideas or facts, regular connections or causal dependencies, there is always a hypothesis. It acts as a link between previously achieved knowledge and new truths and at the same time as a cognitive tool that regulates the logical transition from previous incomplete and inaccurate knowledge to new, more complete and more accurate knowledge.

Thus, the development inherent in the process of cognition predetermines the functioning of the hypothesis in thinking as a necessary and universal form of such development.

(2) The construction of a hypothesis is always accompanied by an assumption about

the nature of the phenomena under study, which is the logical core of the hypothesis and is formulated in the form of a separate judgment or a system of interrelated judgments. It's always

has a weakened epistemic modality: it is a problematic judgment,

which expresses inaccurate knowledge.

To turn into reliable knowledge, a hypothesis is subject to scientific and practical testing.

of
testing a hypothesis, which takes place using various logical techniques, operations and forms of inference, ultimately leads to
a refutation
or
confirmation and its further
proof.
So, a hypothesis always contains probable knowledge that needs to be verified. A position proven on its basis is no longer a hypothesis itself, for it contains verified and unquestionable true knowledge.

(3) The assumption that arises when constructing a hypothesis is born as a result of an analysis of factual material, based on a generalization of numerous observations. An important role in the emergence of a fruitful hypothesis is played by the intuition, creativity and imagination of the researcher. However, a scientific hypothesis is not just a guess, fantasy or assumption, but a rationally based assumption based on specific materials, and not an intuitively and subconsciously accepted assumption.

The noted features make it possible to more clearly define the essential features of the hypothesis. Any hypothesis has initial data, or grounds,

and the end result is
a guess.
It also includes
logical processing of the initial data
and the transition to an assumption.
The final stage of knowledge is testing the hypothesis

Types of hypotheses

In the process of knowledge development, hypotheses differ in their cognitive functions and in the object of study.

1. Based on their functions in the cognitive process, hypotheses are distinguished: (1) descriptive and (2) explanatory.

(1)Descriptive hypothesis —

this is an assumption about the inherent properties of the object under study. It usually answers the question:

“What is this object?” or “What properties does this object have?”

Descriptive hypotheses can be put forward in order to identify the composition or structure of an object, reveal the mechanism or procedural features the functional characteristics

For example, the hypothesis about the wave propagation of light that arose in the theory of physics was a hypothesis about the mechanism of light motion. A chemist's guess about the components and atomic chains of a new polymer refers to hypotheses about composition and structure. The hypothesis of a political scientist or lawyer that predicts the immediate or long-term social effect of an adopted new package of laws refers to functional assumptions.

A special place among descriptive hypotheses is occupied by hypotheses about the existence of an object, which are called existential hypotheses An example of such a hypothesis is the assumption of the once coexistence of the continent of the Western (America) and Eastern (Europe and Africa) hemispheres. The hypothesis about the existence of Atlantis will be the same.

(2) An explanatory hypothesis is an assumption about the reasons for the emergence of the object of research. Such hypotheses usually ask: “Why did this event happen?” or “What are the reasons for this item?”

Examples of such assumptions: the hypothesis of the Tunguska meteorite; hypothesis about the appearance of ice ages on Earth; assumptions about the causes of animal extinction in different geological eras; hypotheses about the motivating reasons and motives for the accused to commit a specific crime and others.

The history of science shows that in the process of developing knowledge, existential hypotheses first arise that clarify the fact of the existence of specific objects. Then descriptive hypotheses arise that clarify the properties of these objects. The last step is the construction of explanatory hypotheses that reveal the mechanism and causes of the occurrence of the objects under study. The consistent complication of hypotheses in the process of cognition - about existence, about properties, about causes - is a reflection of the dialectics inherent in the process of cognition: from simple to complex, from external to internal, from phenomenon to essence.

2. Based on the object of study, hypotheses are distinguished: general and specific.

(1) A general hypothesis is a reasonable assumption about natural connections and empirical regularities. Examples of general hypotheses include: developed in the 18th century. M.V. Lomonosov's hypothesis about the atomic structure of matter; modern competing hypotheses of academician O.Yu. Schmidt and academician V.G. Fesenkova on the origin of celestial bodies; hypotheses about the organic and inorganic origin of oil and others.

General hypotheses serve as scaffolding for the development of scientific knowledge. Once proven, they become scientific theories and are valuable contributions to the development of scientific knowledge.

(2)A particular hypothesis is an educated guess about the origin and properties of individual facts, specific events and phenomena.

If a single circumstance served as the cause of the emergence of other facts and if it is not accessible to direct perception, then its knowledge takes the form of a hypothesis about the existence or properties of this circumstance.

Particular hypotheses are put forward both in natural science and in the social historical sciences. An archaeologist, for example, puts forward a hypothesis about the time of origin and ownership of objects discovered during excavations. A historian builds a hypothesis about the relationship between specific historical events or the actions of individuals.

Particular hypotheses are also assumptions that are put forward in forensic investigative practice, because here we have to make conclusions about individual events, the actions of individual people, individual facts causally related to a criminal act.

Along with the terms “general” and “particular hypothesis”, the term “working hypothesis” is used in science.

A working hypothesis is an assumption put forward at the first stages of the study, which serves as a conditional assumption that allows us to group the results of observations and give them an initial explanation.

The specificity of the working hypothesis is its conditional and thus temporary acceptance. It is extremely important for the researcher to systematize the available factual data at the very beginning of the investigation, rationally process them and outline ways for further searches. The working hypothesis performs the function of the first systematizer of facts in the research process.

The further fate of the working hypothesis is twofold. It is possible that it may turn from a working hypothesis into a stable, fruitful hypothesis. At the same time, it can be replaced by other hypotheses if its incompatibility with new facts is established.

Version

In historical, sociological or political science research, as well as in forensic investigative practice, when explaining individual facts or a set of circumstances, a number of hypotheses are often put forward that explain these facts in different ways. Such hypotheses

are called versions (from the Latin versio - “turnover”, versare - “to modify”).

A version in legal proceedings is one of the possible hypotheses that explains the origin or properties of individual legally significant circumstances or the crime as a whole.

During the investigation of crimes and trials, versions differ in content and scope of circumstances. Among them there are common

and
private versions.
(1) The general version is an assumption that explains all crimes as a whole as a single system of specific circumstances. She answers not one, but many interrelated questions, clarifying the entire set of legally significant circumstances of the case. The most important of these questions will be the following:

what crime was committed? who did it? where, when, under what circumstances and in what way was it committed? What are the goals, motives of the crime, and the guilt of the criminal?

The unknown real reason for which a version is created is not a principle of development or an objective pattern, but a specific set of factual circumstances that make up a single crime. Covering all the issues that need to be clarified in court, this version bears the features of a general summing assumption that explains the entire crime as a whole.

(2) A partial version is an assumption that explains certain circumstances of the crime in question. Being unknown or little-known, each of the circumstances can be the subject of independent research; for each of them, versions are also created that explain the features and origin of these circumstances.

Examples of private versions may be the following assumptions: about the location of stolen things or the location of the criminal; about accomplices of the act; about the method of penetration of the criminal to the place where the act was committed; about the motives for committing a crime and many others.

Private and general versions are closely interrelated with each other during the investigation process. The knowledge obtained with the help of particular versions serves as the basis for constructing, concretizing and clarifying the general version that explains the criminal act as a whole. In turn, the general version makes it possible to outline the main directions for putting forward private versions regarding the yet unidentified circumstances of the case.

Notes

Hypothesis // Explanatory dictionary of the living Great Russian language: in 4 volumes / author's compilation. V. I. Dal. — 2nd ed. - St. Petersburg. : Printing house of M. O. Wolf, 1880-1882.
↑ Merkulov I.P.
// New philosophical encyclopedia / Institute of Philosophy RAS; National social-scientific fund; Pred. scientific-ed. Council V. S. Stepin, deputy chairmen: A. A. Guseinov, G. Yu. Semigin, student. secret A. P. Ogurtsov. — 2nd ed., rev. and additional - M.: Mysl, 2010. - ISBN 978-5-244-01115-9.
Kirillov V.I., Starchenko A.A.
Logic: Textbook for law schools. — 5th, revised. and additional - M.: Yurist, 2002. - 256 p. — ISBN 5-7975-0059-0.
Popper, Karl.
Conjectures and refutations: the growth of scientific knowledge (English). - London: Routledge, 2004. - ISBN 0-415-28594-1.
↑ Kirsten Walsh.
The Idea of Principles in Early Modern Thought: Interdisciplinary Perspectives / Peter R. Anstey. - Routledge, 2017. - 304 p. — ISBN 9781315452678.
Ivlev Yu. V.
Logic. - M., Prospekt, 2015. - p. 269—270

Hypothesis Examples

So, how is a hypothesis correctly formulated in a course work? Examples from different fields of science will guide you to the right thoughts.

Direction of course work: business, entrepreneurship.

Topic: Motivating the activities of employees of the organization.

Hypothesis: It can be assumed that employee motivation is closely related to their perception of their own success in the workplace, as well as the expectation of immediate reward.

Direction: Production management.

Topic: Document flow in the organization.

Hypothesis. It should be expected that with a deeper introduction of the latest computer technologies in a company, the level of organization of its document flow will significantly increase, bringing the number of losses of important documents to zero.

Direction: Pedagogy.

Topic: Increasing the curiosity of children of primary school age.

Hypothesis: It can be expected that the level of curiosity of younger schoolchildren will increase with proper motivation on the part of the teaching staff and an increase in the interest of the teachers themselves in the educational process.

About hypotheses

What's the hypothesis for me?

Why has the hypothesis tool become the most popular in assessing business results? Since any hypothesis requires its own proof, this has become the best tool for testing certain ideas for business development. To prove any hypothesis, an experiment is necessary. Thus, the formation of hypotheses alone is not enough; they must be tested and either confirmation or refutation obtained. In our case, the hypothesis must have a measurable goal in the end. Let's say that there is a hypothesis about the possibility of selling flowers at metro stations. We can formulate the task for testing the hypothesis as follows: We need to survey 25 men and 25 women who entered a metro station about their readiness to buy flowers in the subway at this metro station, and if 30% of respondents say “yes,” then the hypothesis will be confirmed. Please note that the hypothesis formulation includes indicators to measure the outcome.

What result do you want to get as a result of developing a hypothesis, what will become the basis for launching further implementations or changes in the business?

So, why are hypotheses worked out: 1. Search

the optimal solution for improving the economic performance of a business (for making changes, for finding new sales or promotion options, or studying the sales channel, etc.) 2.
Refusal
of illusions.
Until a hypothesis is tested, it is an illusion (something that does not exist) 3. At each stage, a business has its own tasks and hypotheses to test will be different. A hypothesis must be testable
and confirmed or refuted.

A living example: an online store of wine accessories decided to expand its sales market because it needed sales growth and an increase in product turnover. We developed a hypothesis: “That wine boutiques have difficulty increasing the average bill. Wine coolers and accessories can be an additional sale to increase the average check. We will call 10 wine boutiques. If 4 of them agree to such a project, then we consider the hypothesis verified.

IMPORTANT. The first step should be clear. Where to start actions. If only questions arise, then this is not a hypothesis for an entrepreneur.

A hypothesis is a risky assumption tied to an expected result.

The final formula is “hypothesis = we will do something, we will get something.” For example, if we make 10 calls, we will get one sale. If you received it, it was confirmed; if not, it failed.

Let me draw your attention to this point: testing hypotheses very well reveals the true needs of your clients.

Meaning

A hypothesis (from the Greek hypothesis, which means “foundation”) is a preliminary assumption that explains a certain phenomenon or group of phenomena; may be associated with the existence of an object or item, its properties, as well as the reasons for its occurrence.

A hypothesis itself is neither true nor false. Only after receiving confirmation does this statement become true and cease to exist.

In Ushakov’s dictionary there is another definition of what a hypothesis is. This is a scientific unproven assumption that has a certain probability and explains phenomena that are inexplicable without this assumption.

Vladimir Dal also explains in his dictionary what a hypothesis is. The definition says that this is a guess, a speculative (not based on experience, abstract) position. This interpretation is quite simple and brief.

The no less famous dictionary of Brockhaus and Efron also explains what a hypothesis is. The definition given in it is related only to the system of natural sciences. According to them, this is an assumption that we make to interpret phenomena. A person comes to such statements when he cannot establish the causes of a phenomenon.

HYPOTHESIS

HYPOTHESIS (from the Greek ὑπόϑεσις - basis, assumption), 1) scientific. a statement whose truth value is uncertain; 2) scientific method. cognition, which includes the formulation and subsequent empirical (experimental) testing of assumptions; 3) structural element of scientific. theory or a set of interrelated theories.

In logical The structure of a hypothesis is usually divided into a basis (premises) and a conclusion, which can only be confirmed by the premises with some degree of probability. Logical the structure of a hypothesis formally coincides with the similar structure of a plausible inference (induction, analogy, statistical inference). However, unlike the latter, the truth of the premises of a theory always remains uncertain, and these premises themselves change in the course of the evolution of science. knowledge. Therefore, the degree of confirmation of the conclusion of a hypothesis by its premises also turns out to be a historically changing value.

Being the most important scientific method. cognition, G. always comes forward in the course of the development of the department. fields of science in order to solve specific problems: for example, to predict or explain new experiments. data, remove the contradiction between theory and experimental results, build on the basis of fundamentals. theories, particular theories or applied models, etc. Therefore, any G. must be relevant in relation to such problems or experiments. data that it allows to explain or predict. It must also contain new conceptual information, have additional (compared to previous or competing theories and theories) theoretical. content. But even if this content does not receive empirical confirmation, G., directing scientific. research in a certain direction performs an important cognitive function, since when putting forward new theories, researchers necessarily rely on empirical results. checks (including negative ones) of their predecessors.

As a scientific G.'s statements must satisfy the criterion of fundamental empirical. verifiability (i.e., regardless of the given level of development of science). In modern empirical science verification of highly abstract G. is usually indirect in nature; it requires plural. intermediate links in the form of auxiliary. G., theoretical. models, experimental models. installations, etc. Fundamental empirical. testability scientific. G. means that they have the properties of falsifiability and verifiability (see Falsification, Verification). The falsifiability property fixes the presumptive nature of the field of application of scientific statements. Since the latter are statements of limited generality, they not only allow, but also directly or indirectly prohibit something in the field under study (a typical example is the Pauli principle in quantum mechanics, which prohibits the existence of two or more electrons in identical quantum states). The verifiability property allows one to establish and verify relative empirical data. content D. The greatest value is confirmation by such experiments. data, the existence of which could not be assumed before the nomination of the test G. The property of verifiability (as well as falsifiability) is absolute, since the authority that potentially confirms the G. may be both known and as yet unknown facts.

What is the essence of the hypothesis?

A hypothesis reflects objective reality. In this it is similar to different forms of thinking, but it is also different from them. The main specificity of a hypothesis is that it reflects facts in the material world in a conjectural manner; it does not assert categorically and reliably. Therefore, a hypothesis is an assumption.

Everyone knows that when establishing a concept through the closest genus and difference, it will also be necessary to indicate distinctive features. The closest genus for a hypothesis in the form of any result of an activity is the concept of “assumption”. What is the difference between a hypothesis and a guess, fantasy, prediction, guessing? The most shocking hypotheses are not based on speculation alone; they all have certain characteristics. To answer this question, you will need to identify essential features.

Testing the hypothesis about the population mean

It is often necessary to check whether the population mean is significantly different from some given value, such as a standard. In this case, the main and alternative hypotheses can be written as follows:

;

When testing the hypothesis about the sample mean, the Student t-test is often used as a statistical criterion, but it should be remembered that this criterion is applicable only when the sample data obeys a normal distribution law. The critical values of the criterion, according to the selected significance level α and the degree of freedom v, can be found in the appendices of books on statistics, and if the hypothesis is tested using a computer program, for example, STATISTICA, then the program selects it.

The null hypothesis cannot be rejected with probability P = 1 - α, if , where

— sample average,

— some given average value, for example, standard,

s - standard deviation,

n—sampling power,

— critical value of Student's t-test.

Example 4. A kvass manufacturer decided to find out whether the bottle filling device works according to the standard. The main and alternative hypotheses are formulated as follows:

20 bottles were randomly selected for testing, the average empty level was mm, standard deviation mm.

Since the sample is very small (20 units) and the standard deviation of the population is not known, a confidence level of p = 95% was chosen.

We get the actual value of the statistical criterion:

Critical value of Student's t-test:

Since , that is, the actual value of the statistical criterion is less than the critical value, then the actual value falls within the area of acceptance of the hypothesis. Therefore, it is not possible to reject the main hypothesis that the average empty bottle level is not significantly different from 50 mm.

In research, there are often cases of the need to compare averages in groups, one of which can be called “normal” and the other far from the “norm”, that is, it is necessary to compare two groups

Statistics is not your specialty? Order statistical data processing

Working with a hypothesis

From introduction to conclusion, the hypothesis will relentlessly guide the course of your scientific work. In the first section of the main part, you will prove or reject hypotheses based on the collected facts. Analyze them and accompany them with your own opinion. The second section includes the results of your experiments and research, and the calculations you performed.

All interaction with the hypothesis is divided into the following stages.

Origin. Identifying facts and assumptions that do not fit into any known theory on your topic. These conclusions should cause heated debate in society and urgently require explanation, proof or refutation.
Formulation based on these conclusions.
Theoretical research. Search for opinions related to the hypothesis in different sources. Comparing the ideas expressed with your own ideas, analyzing and citing them.
Practical research. Carrying out thematic experiments related to the hypothesis. Analysis of the results obtained. Performing calculations, preparing all kinds of final charts and graphs.
Comparison of the research results obtained with the hypothesis, its subsequent refutation or confirmation.

Don’t forget to touch on the hypothesis in the conclusion, share your opinion on how true it is, and whether it can become a theory and become widespread in public opinion. Perhaps you will put forward and prove a hypothesis that will become a turning point in the development of your field of knowledge.

Stages of testing statistical hypotheses

A statistical hypothesis is not just a statement and its verification, but a whole mental and calculation process that requires compliance with certain rules. Now we will tell you about the main stages of evaluating the author’s static assumptions.

Stage No. 1. Formulation of the main and alternative hypothesis.

To begin with, the researcher needs to correctly formulate the main assumption - this is precisely the statement that requires verification. It must be plausible and logical. Moreover, this assumption will initially be considered correct, effective and true until it is refuted. The main statistical hypothesis is designated H0.

Types of hypotheses

An alternative hypothesis is a worthy “back-up airfield” when, during the course of research, the main assumption is destroyed and it is necessary to accept another property of the object being studied as true.

It is important to note that, in general, the main and alternative hypotheses can be closely intertwined with each other; they seem to back each other up. Let us give an example of such assumptions.

Alternative and main hypotheses

Stage No. 2. We determine the efficiency scale and parameters for assessing the effect of hypotheses.

A statistical test is a statement of certain parameters by which the main and alternative hypotheses will be evaluated. The assessment system can be based on general rules and principles, as well as mathematical operations, indicators, standards, etc. It is important to correctly compare the results, the actions of each hypothesis and formulate a reasoned conclusion.

What indicators are used to evaluate a hypothesis?

After rating on an approved scale, the researcher must decide on the effectiveness and ineffectiveness of the proposed assumption, its acceptance or rejection.

Most often, a statistical criterion is a random variable, a predictive value (that is, a desired result). The closer the real indicator is to it, the more effective and correct the assumption.

Stage No. 3. Level of significance, permissible errors and errors.

The significance level represents a certain acceptable error when evaluating a hypothesis. Here it is important to clarify in what ranges deviation is permissible, against the background of which it is possible to accept the author’s proposal, and in what interval the idea should be abandoned altogether.

When testing a statistical hypothesis, experts do not rule out obtaining a false result due to certain errors. Thus, errors of the first and second types are distinguished.

Any difficulties?

Need teacher help?

We are always happy to help you!

master's theses

An error of the first type assumes that the main hypothesis will be rejected by the researcher, despite the fact that it is true. In this case, the expert received unreliable data or made a mistake in calculations, analysis and received an incorrect result, although in fact the hypothesis is effective.

A Type II error suggests that the underlying hypothesis will be accepted despite false results.

It is important to note that the analyst can manage errors of the first type (he independently determines the likelihood of its occurrence, in what cases, predicts it and avoids it). Errors of the second type cannot be controlled and controlled. Therefore, when testing an assumption, it is important to subsequently plan the course of action and the potential result so that in case of failure the level of losses is minimal.

No less important is the level of trust. Initially, the main hypothesis is considered true. It will be rejected only if the researcher has received a reasonable rebuttal. Initially, the main hypothesis is a working one, and if no important refutations are received during the study, then the level of confidence and approval of the main idea will increase.

Example of assessing the level of trust

Stage No. 4. Determining the boundaries of accepting or rejecting a hypothesis.

Here the researcher will have to determine the specific parameters and values, the area that will form the basis of the test and will serve as the impetus for the final decision: accept or reject the idea.

Graph of the boundaries of acceptance or rejection of a hypothesis

For example, a hypothesis will be accepted if certain indicators (the largest part of them) are within the normal range and correspond to the predicted values, and rejected if the established goal is not achieved.

It is important to identify the areas of acceptance of the hypothesis and the critical area (rejection of the hypothesis). There must be specific restrictions and rules by which the effect will be regulated and assessed.

Stage No. 5. Conclusion of the results obtained: accept or reject the hypothesis.

At this stage, the author must emphasize the main factors: why the hypothesis is effective or ineffective, what criteria (parameters) indicate this. It is important to emphasize that the main hypothesis was found or not refuted, and then emphasize the conclusion: acceptance or rejection of the idea.

If the evaluation parameters fall within the acceptance range, then the author concludes that it is impossible to reject the main hypothesis by accepting it. If the results of the analysis fall in the critical area, then the researcher rejects the main idea.

Thus, testing a statistical hypothesis comes down to the fact that the author of the idea checks its effectiveness and validity. To do this, he will have to correctly formulate his assumption, develop a scale for its assessment and analysis, and then, based on the results obtained, formulate a conclusion and make a final decision. It is important to take the “test mission” responsibly, carefully check every action in order to minimize errors and mistakes, and also to obtain a reliable result.

Theory and practice

If you pay attention to the actual practice of forming and proving hypotheses, it becomes clear that the actual approach quite often differs significantly from the theoretical one. Often, the first thing that researchers receive is the parameters of the relationship between the observed phenomena, and based on this, attempts are made to establish the nature of the observed phenomena.

The very fact of the existence of a connection often does not become the object of formulating a hypothesis at all. More often, attention to this issue is paid to the scientist’s opponents or specialists who are entrusted with writing a review of a work devoted to a certain hypothesis.

The most significant question is to establish the fact of the existence of a connection between objects (that is, the phenomenon to which the hypothesis is devoted) within the framework of sciences that do not interact with real objects. A good example is mathematics. Scientists involved in this discipline spend years proving the existence of uncountable sets, irrational numbers, and regular heptahedrons. There is a certain logic in this: when interacting with real objects, there is no need to provide an evidence base for the existence of the observed phenomenon, since the researcher sees and feels it.

Example of statistical hypothesis testing

So, as you probably guessed from the examples above, we will test the hypothesis that there is a significant difference between the number of European AI startups created in 2019 and 2020. The example is simple enough to make it easier to understand how the algorithm works.

Figure 1 - initial data

First, let's pay attention to the initial sample (Fig. 1): the dataset is presented for 30 European countries, only startups officially registered in the country are included. Quantitative data for two years. It is worth noting that the samples are paired, that is, we observe the same indicator for the same countries with a difference of a year.

It is immediately worth noting that two statistical hypotheses will be tested in a row. In order to apply a criterion to compare the average samples of two years, you must first determine the law of data distribution. Thus, step 1 is testing the statistical hypothesis about the law of data distribution. Step 2 - testing the statistical hypothesis about equality between means.

Testing the hypothesis about the distribution law

For data from 2019, let's check the normality of the distribution.

H0: random variable is normally distributed
H1: The random variable is not normally distributed
Let the significance level alpha = 0.05 (as in 95 percent of statistical tests). Determining the level of significance is worthy of a separate post, so we won’t focus on it.
The Shapiro-Wilk test will be used.
At this step, you need to understand how the criterion works. In this case, the following statistics are calculated - a function of our sample:
, , , ;
As you can see, the formula is not very simple, plus there is a complex mechanism for determining the parameter a, so in such cases it is easier to use online calculators to calculate statistics. For example, I will use a good online statistical resource - https://www.statskingdom.com/320ShapiroWilk.html.
So, the calculator showed us that p-value = 1.20005e-9 , W = 0.435974 ; What to do next? There are two options:
You can compare the W statistics with the critical value Wcrit. The critical value is most often given in ready-made tables (the sample size and significance level are indicated in the rows/columns, and Wcrit lies at the intersection). If W>Wcrit., then we do not reject H0 and vice versa. But this is not very convenient, so the second method is more often used.
You can compare p-value with alpha (selected in step 2). If p-value < alpha, then we reject H0. If not, then DO NOT reject H0. In our case, p-value < alpha, therefore we reject H0 with 95% confidence.
H0 is rejected, the distribution of sample data for 2019 is not subject to the normal distribution law.

For data from 2020, let's check the normality of the distribution. Here the steps are exactly the same. It turned out that p-value = 3.41343e-9. The p-value < alpha, therefore we reject H0.

Thus, the values in both samples are not normally distributed. To compare the averages in two years, we will use the Wilcoxon test.

Testing the hypothesis about the difference in the number of AI startups in European countries for 2019 and 2020

H0: There is no statistically significant difference between the number of AI startups in Europe in two years.
H1: The statistical significance of the change in the number of AI startups in Europe between 2019 and 2020 is recognized.
Let the significance level alpha = 0.05.
The Wilcoxon test will be used.
At this step, you need to understand how the criterion works. Of course, there are also online calculators for this criterion, but it is quite easy to calculate it manually. The algorithm is very simple:
Step 1 - For each country, you need to calculate the difference between the values of the two years.
Step 2 - Next, understand which of the differences are typical, that is, they correspond to the direction of change in the indicator that is predominant in frequency.
Step 3 - Next, in ascending order, rank the differences of the pairs according to their absolute values. The smaller absolute value of the difference is assigned a lower rank.
Step 4 - Calculate the sum of ranks corresponding to atypical shifts. This will be the T-test value.
An example of a calculation for twelve countries is shown in the figure below (Figure 2). Don't be alarmed, the given ranks are calculated for all 30 elements of the sample, twelve countries are shown for illustration purposes only. Having carried out this calculation for all 30 countries and adding up the ranks for countries with atypical changes, it turned out that T = 28.
Let's compare T and Tcrit.=163. T < Tcrit, which means with 95th confidence the change in the number of startups is statistically significant.
H0 is rejected; the differences between the number of European AI startups in 2019 and 2020 are significant.

Figure 2 - example of calculating the Wilcoxon test

Interesting hypotheses and their refutation

Everything always starts small. All physics was built on countless shocking hypotheses that were confirmed or refuted by scientific practice. Therefore, it is worth mentioning some interesting ideas.

Some particles move from the future to the past. Physicists have their own set of rules and prohibitions, which are considered to be canon, but with the advent of tachyons, it would seem that all norms have been shaken. A tachyon is a particle that can violate all accepted laws of physics at once: its mass is imaginary, and it moves faster than the speed of light. The theory has been put forward that tachyons can travel back in time. The particle was introduced by theorist Gerald Feinberg in 1967 and declared that tachyons were a new class of particles. The scientist argued that this is actually a generalization of antimatter. Feinberg had a lot of like-minded people, and the idea took root for a long time, however, refutations still appeared. Tachyons have not completely disappeared from physics, but still no one has been able to detect them either in space or in accelerators. If the hypothesis were true, people would be able to contact their ancestors.
A drop of water polymer could destroy the oceans. This one of the most shocking hypotheses suggests that water can be transformed into a polymer - this is a component in which individual molecules become links in a large chain. In this case, the properties of water should change. The hypothesis was put forward by chemist Nikolai Fedyakin after an experiment with water vapor. The hypothesis has frightened scientists for a long time, because it was assumed that one drop of an aqueous polymer could turn all the water on the planet into a polymer. However, the refutation of the most shocking hypothesis was not long in coming. The scientist’s experiment was repeated, but no confirmation of the theory was found.

There were a lot of such shocking hypotheses at one time, but many of them were not confirmed after a series of scientific experiments, but they were not forgotten. Fantasy and scientific justification are the two main components for every scientist.

RNA came first

For years, scientists have debated whether DNA or RNA is more important. DNA serves as the primary means of storing genetic information. RNA is a ribonucleic acid that can act as a genetic library and catalyze reactions. This ability makes RNA an ideal candidate for the origin of the first life on Earth.

So where did RNA come from? Can RNA form spontaneously? First, consider the structure of RNA, consisting of four nucleotide bases:

Adenine
Guanine
Cytosine
Uracil

These four nucleotides are the building blocks of RNA. If they can be synthesized spontaneously under the conditions of the early Earth, then much of the puzzle of how life began could be solved. Now, it has recently been discovered that some molecules can actually form all four nucleotides in the presence of ultraviolet radiation or sunlight.

Occam's razor for testing hypotheses

There are principles, for example, Occam's Razor, which are not axioms, but presumptions, that is, they do not prohibit more complex explanations of phenomena in principle, but only recommend the order of consideration
of hypotheses, which in most cases is the best. Albert Einstein formulated the principle of Occam's razor this way: “Everything should be simplified as far as possible, but no more.” Reformulated in the language of information theory, Occam's razor principle states that the most accurate message is the message of minimum length. There are others too.
Among the most famous examples of the application of this principle is the answer given to Emperor Napoleon by the creator of the first theory of the origin of the Solar system, mathematician and physicist Laplace. Napoleon asked why the word “God,” constantly repeated by Lagrange, does not appear at all in his work, to which Laplace replied: “That’s because I didn’t need this hypothesis.”

Statistical tests for testing hypotheses

A statistical criterion is a statistical characteristic of a sample, calculated using a certain mathematical relationship (formula) based on the data available in the sample.

Based on the value of this characteristic, a decision is made whether to accept the main hypothesis or not. There are two types of statistical criteria:

one-sided criterion - a criterion whose values belong to the region (0; +∞);
two-sided criterion - a criterion whose values belong to the region (-∞; +∞).

Properties of the statistical test:

the statistical criterion is a random variable whose distribution law is known. Often the name of a statistical criterion mentions its distribution law. For example, the Pearson chi-square test follows the chi-square distribution law;
The closer the value of the statistical criterion is to zero, the more likely it is that the main hypothesis is true.

How is a hypothesis born?

Creating an argument in the human mind is not a simple thought process. The researcher must be able to create and update acquired knowledge, and he must also have the following qualities:

Problem vision. This is the ability to show the paths of scientific development, establish its main trends and connect disparate tasks together. Combines the problem vision with the already acquired skills and knowledge, instinct and abilities of a person in research.
Alternative character. This trait allows a person to draw interesting conclusions and find something completely new in known facts.
Intuition. This term refers to an unconscious process and is not based on logical reasoning.

Features of the hypothesis

If we talk about this concept, then it is worth establishing its characteristic features.

A hypothesis is a special form of development of scientific knowledge. It is hypotheses that allow science to move from individual facts to a specific phenomenon, generalization of knowledge and knowledge of the laws of development of a particular phenomenon.
A hypothesis is based on making assumptions that are associated with a theoretical explanation of certain phenomena. This concept acts as a separate judgment or a whole line of interrelated judgments, natural phenomena. Judgment is always problematic for researchers, because this concept speaks of probabilistic theoretical knowledge. It happens that hypotheses are put forward on the basis of deduction. An example is K. A. Timiryazev’s shocking hypothesis about photosynthesis. It was confirmed, but initially it all started from assumptions in the law of conservation of energy.
A hypothesis is an educated guess that is based on some specific facts. Therefore, a hypothesis cannot be called a chaotic and subconscious process; it is a completely logical and logical mechanism that allows a person to expand his knowledge to obtain new information - to understand objective reality. Again, we can recall the shocking hypothesis of N. Copernicus about the new heliocentric system, which revealed the idea that the Earth revolves around the Sun. He outlined all his ideas in the work “On the Rotation of the Celestial Spheres”, all guesses were based on a real factual basis and the inconsistency of the then still valid geocentric concept was shown.

These distinctive features, taken together, will distinguish a hypothesis from other types of assumption, as well as establish its essence. As you can see, a hypothesis is a probabilistic assumption about the causes of a particular phenomenon, the reliability of which cannot now be verified and proven, but this assumption allows us to explain some of the causes of the phenomenon.

It is important to remember that the term “hypothesis” is always used in a dual sense. A hypothesis is an assumption that explains a phenomenon.

A hypothesis is also spoken of as a method of thinking that puts forward some assumption, and then develops the development and proof of this fact.

A hypothesis is often constructed in the form of an assumption about the cause of past phenomena. As an example, we can cite our knowledge of the formation of the solar system, the earth's core, the birth of the earth, and so on.

What can be said about the concept of hypothesis and types?

In the process of the evolution of knowledge, hypotheses begin to differ in cognitive qualities, as well as in the object of study. Let's take a closer look at each of these types.

Based on their functions in the cognitive process, descriptive and explanatory hypotheses are distinguished:

A descriptive hypothesis is a statement that speaks about the inherent properties of the object under study. Typically, an assumption allows us to answer the questions “What is this or that object?” or “What properties does the object have?” This type of hypothesis can be put forward in order to identify the composition or structure of an object, reveal its mechanism of action or features of its activity, and determine functional features. Among descriptive hypotheses there are existential hypotheses, which speak about the existence of some object.
An explanatory hypothesis is a statement based on the reasons for the appearance of a particular object. Such hypotheses make it possible to explain why a certain event occurred or what are the reasons for the appearance of an object.

History shows that with the development of knowledge, more and more existential hypotheses appear that tell about the existence of a specific object. Next, descriptive hypotheses appear that tell about the properties of those objects, and finally explanatory hypotheses are born that reveal the mechanism and reasons for the appearance of the object. As you can see, there is a gradual complication of the hypothesis in the process of learning new things.

What hypotheses are there for the object of study? There are general and private.

General hypotheses help to substantiate assumptions about natural relationships and empirical regulators. They act as a kind of scaffolding in the development of scientific knowledge. Once hypotheses are proven, they become scientific theories and contribute to science. A partial hypothesis is an assumption with justification about the origin and quality of facts, events or phenomena. If there was a single circumstance that caused the appearance of other facts, then knowledge takes the form of hypotheses. There is also such a type of hypothesis as a working one. This is an assumption put forward at the beginning of the study, which is a conditional assumption and allows you to combine facts and observations into a single whole and give them an initial explanation. The main specificity of the working hypothesis is that it is accepted conditionally or temporarily

It is extremely important for the researcher to systematize the acquired knowledge given at the beginning of the study. Afterwards they will need to be processed and a further route to be outlined.

A working hypothesis is exactly what is needed for this.

The concept of a hypothesis and its structure.

In science and everyday thinking, we move from ignorance to knowledge, from incomplete knowledge to more complete knowledge. We have to make and then justify various assumptions to explain phenomena and their relationships with other phenomena. We put forward hypotheses that, when confirmed, can turn into scientific theories or into individual true judgments, or, conversely, will be refuted and turn out to be false judgments.

A hypothesis is an assumption that arises in the course of intellectual practice about the causes or natural connections of any phenomena or events of nature, society, or thinking, the basis for testing which is inferential operations followed by a truthful assessment of the resulting consequences. During the development process, a hypothesis can become the object of evidentiary procedures that allow one to compare arguments that confirm or refute the proposed assumption. (From this point of view, the origin of the term hypothesis, which from Greek through Latin penetrated into European languages, is worthy of attention. The word hipothesis is made up of the prefix hipo (“under”) and the term thesis, which does not need translation. Thus, etymologically, the word hypothesis can be interpreted as subthesis or prethesis, i.e. something that in the course of research tends to turn into a thesis, become the object of proof).

The specificity of a hypothesis - to be a form of development of knowledge - is predetermined by the basic property of thinking, its constant movement - deepening and development, a person’s desire to discover new patterns and causal connections, which is dictated by the needs of practical life. The assumption that arises when constructing a hypothesis is born as a result of an analysis of factual material, based on a generalization of numerous observations. This means that a hypothesis is not any guess, fantasy or assumption, but only a well-founded position based on specific materials. In accordance with this, the emergence of a hypothesis is not a chaotic or subconscious, but a natural logical process. To turn into reliable knowledge, an assumption is subject to scientific and practical verification.

One of the specific features of a hypothesis is associated with the idea of its rational inclusion in a certain fragment of knowledge. Judgments (groups of judgments) in which hypotheses are expressed must be able to play the role of logical foundations for more or less complex inferential constructions.

Any hypothesis has initial data, or grounds, and the final result of reasoning is an assumption. It also includes processing of initial data and a logical approach to making assumptions. The final stage of knowledge is testing the hypothesis, turning the assumption into reliable knowledge or refuting it.

However, ideally, any hypothesis is oriented towards passing through a certain cycle, consisting of the following stages: 1) origin (promotion), 2) development (derivation of consequences), 3) verification (proof, justification, refutation).

Types of hypotheses.

Depending on the degree of generality, scientific hypotheses can be divided into general, particular, individual, scientific, and working.

A general hypothesis is a scientifically based assumption about the causes, laws and patterns of natural and social phenomena, as well as the patterns of human mental activity. General hypotheses are put forward with the aim of explaining the entire class of phenomena described, deducing the natural nature of their relationships at any time and in any place. Examples of general hypotheses include: developed in the 18th century by M.V. Lomonosov hypothesis about the atomic structure of matter, modern hypotheses of Acad. O.Yu. Schmidt and acad. V.G. Fesenkov about the origin of celestial bodies, hypotheses about the organic and inorganic origin of oil.

Once proven, they become scientific theories and are valuable contributions to the development of scientific knowledge.

A particular hypothesis is a reasonable assumption about the causes, origin and patterns of some objects isolated from the class of objects of nature, social life or human mental activity under consideration.

Particular hypotheses find application both in natural science and in social and historical sciences. An archaeologist, for example, puts forward a particular hypothesis about the time of origin and ownership of objects discovered during excavations. A historian hypothesizes the relationship between specific historical events or the actions of individuals.

Particular hypotheses are also those assumptions that are used in forensic investigative practice, because here we have to make conclusions about individual events, people’s actions, individual facts causally related to the crime.

A single hypothesis is a scientifically based assumption about the causes, origin and patterns of individual facts, specific events or phenomena. The doctor builds individual hypotheses during the treatment of a particular patient, selecting medications and their dosage individually for him.

In the course of proving a general, particular and single hypothesis, people build working hypotheses.

A working hypothesis is an assumption put forward, as a rule, at the first stages of research. The working hypothesis does not directly pose the task of elucidating the actual causes of the phenomena under study, but serves only as a conditional assumption that allows us to group and systematize the results of observations into a specific system and give a description of the phenomena that is consistent with the observations.

A scientific hypothesis is a reasonable assumption, consistent with existing knowledge and facts, put forward to explain a phenomenon.

A scientific hypothesis indicates the direction of scientific research; it stimulates and directs the development of knowledge. In this connection, it is often characterized as a form of knowledge development. This obviously implies what is called the hypothetico-deductive method of cognition. And hypotheses, along with deduction, are the main elements of this method.

Hypothesis testing

Please note that all the above statements are verified. The main feature of a hypothesis is that something can be tested and that these tests can be reproduced

An example of an untested statement is: “All people fall in love at least once.” The definition of love is subjective. Moreover, it would be impossible to survey every person about their love life. A non-simple statement can be modified to make it testable. For example, the previous statement could be changed to: “If love is an important emotion, some may believe that everyone should fall in love at least once.” With this statement, a researcher can survey a group of people to see how many believe that people should fall in love at least once.

A hypothesis is often examined by multiple scientists to ensure the integrity and validity of the experiment. This process can take years, and in many cases hypotheses do not advance further in the scientific method because it is difficult to gather sufficient supporting evidence.

“As a field biologist, my favorite part of the scientific method is in the data collection area,” Jaime Tanner, a biology professor at Marlboro College, said on the air. “But what really makes it interesting is trying to answer the question, so the first step in defining questions and creating possible answers (hypotheses) is also very important and a creative process. Then, once you collect the data, analyze to see if your hypothesis is supported or not.”

A null hypothesis is a hypothesis that may be false or have no effect. Often, during a test, the scientist will explore another branch of an idea that might work, called an alternative hypothesis.

During a test, a scientist may attempt to prove or disprove only the null hypothesis or test both the null and alternative hypotheses. If a hypothesis points in a certain direction, it is called a one-sided hypothesis. This means, according to the scientist, that the result will be either with an effect or without an effect. When a hypothesis is created without predicting an outcome, it is called a two-sided hypothesis because there are two possible outcomes. The result may or may not have an effect, but until testing is complete there is no way to know what the result will be.

Statistical hypotheses: basic concepts. Hypothesis testing steps

A statistical hypothesis is some assumption about the properties of a population that needs to be tested. Statistical hypotheses are put forward when it is necessary to check whether an observed phenomenon is an element of chance or the result of the influence of certain measures.

For example, you need to find out whether the average sales volume after an advertising campaign is significantly different from the average sales volume after an advertising campaign. If the answer to this question is positive, then we can conclude that the changes are the result of an advertising campaign.

Conclusions obtained by testing statistical hypotheses are probabilistic in nature: they are accepted with some probability. A statistical hypothesis can also be an assumption about the properties of two populations if, for example, during the course of an intervention there was an impact on only one population and it is necessary to conclude whether this impact was effective.

The steps for testing statistical hypotheses are as follows:

the main hypothesis H0 and the alternative hypothesis H1 are formulated;
a statistical criterion is selected with which the hypothesis will be tested;
the value of the significance level α is set;
the boundaries of the area of acceptance of the hypothesis are located;
a conclusion is made about accepting or rejecting the main hypothesis H0.

Let's look at these steps and related concepts in more detail.