Types of Hypotheses

What are the types of hypotheses?

A hypothesis is a probable, objective and specific answer to a scientific question, which must be verified.

There are different types of hypotheses: the research or working hypothesis, the alternative hypothesis, the null hypothesis or the statistical hypothesis.

However, it is possible for an investigation to have more than one hypothesis. This means that the different types of hypotheses are also related to each other. For example, a research hypothesis can act as the main hypothesis in a work, but in turn the null, alternative and statistical hypotheses help to clarify the central hypothesis.

To understand this better, let’s look at each type of hypothesis separately and its respective variants (with examples).

Research or work hypothesis

The research hypothesis aims to answer what is the relationship established between various variables. It is also known as a working hypothesis. It is the starting point of all scientific research.

According to its approach, it is divided into descriptive hypotheses, causal hypotheses, correlational or group difference hypotheses.

Descriptive hypotheses

They limit themselves to describing the relationship between the variables under study, but do not explain their causes. They anticipate the expected variable type, value, and qualities.

For example«Crime in the city of Caracas has increased 50% in relation to the year 2019».

Causal hypotheses

Causal hypotheses or causality hypotheses are those that propose to explain the cause-effect relationship between two or more variables. They can be explanatory or predictive.

  • Explanatory hypotheses. They offer a possible explanation about the cause that relates the variables. For example“Excessive alcohol consumption causes neuronal damage”.
  • Predictive hypotheses. They predict how one variable will behave in response to another. For example“Global warming will cause flooding in the coming years.”
You may be interested:  Meaning of Bacteria

Both explanatory and predictive hypotheses can be formulated inductively or deductively. Let’s see.

  • Deductive hypotheses: Based on a theory, the researcher formulates a hypothesis to explain a specific case. That is, deductive hypotheses are formulated from the general to the particular. For example, “All living things have DNA. Bacteria are living beings. Therefore, bacteria have DNA.
  • Inductive hypotheses: From the observation of a specific case or phenomenon, the researcher formulates a generalization or general principle. That is, inductive hypotheses are formulated from the particular to the general.
    For example, Newton observed that although the Moon and the apple are two spherical bodies, only the apple falls to the ground. Capturing this specific difference allowed him to induce the existence of a law that would explain such behavior. Thus, he formulated the hypothesis that there is a force of attraction (gravity) between bodies.

Correlational hypotheses

Correlational or joint variation hypotheses are those that establish the degree of mutual relationship between the variables, that is, how and to what degree one affects the other (and vice versa). In this type of hypothesis, the order of the variables does not matter.

For example, Newton’s theory of gravity is a correlational hypothesis, since its statement dictates: “The greater the mass, the greater the force of attraction.” Correlatively, it follows that: “The greater the attractive force, the greater the mass.”

Correlational hypotheses can be negative, positive or mixed.

For example,

  • Positive: “The greater impunity, the greater criminality.”
  • Negative: “The lower the fat intake, the lower the risk of suffering from coronary heart disease.”
  • Mixed: “The higher the altitude, the lower the temperature.”
You may be interested:  Optics

Group difference hypothesis

Group difference hypotheses are those that anticipate the difference in the behavior of various groups. It is based on statistical comparison. Group difference hypotheses are expressed in two variants:

  • Those that establish a difference between two groups, without determining which group falls on. For example“There is a difference in mortality rates from covid-19 between female and male people”.
  • Those that determine on which of the groups falls the difference. For example«The mortality rate from covid-19 is higher in males than in females».

Null hypothesis

The null hypothesis is one that denies the relationship between two or more variables based on a sample parameter. Your statement is negative, which means that it includes a “no.” It is represented by the symbol H0. The null hypothesis is not accepted, but is either rejected or not rejected.

The formulation of the null hypothesis is normally accompanied by the formulation of an alternative hypothesis that seeks to prove it false.

For example«The muscle mass index is not associated with the sex of people».

Alternative hypothesis

Every null hypothesis generates an alternative hypothesis, that is, an alternative answer to the null hypothesis that purports to prove it false. It is represented by the symbol H1. This type of hypothesis is accepted or not accepted.

For example,

  • H0: «The muscle mass index is not associated with the sex of people»
  • H1: «The muscle mass index differs between men and women».

Statistical hypothesis

Statistical hypotheses are those that translate the hypotheses into statistical symbols. Seek to assert or define the parameters of one or more populations. Therefore, they are formulated whenever it is expected to collect data in numbers, percentages or averages.

You may be interested:  Meaning of Glucose

They are subdivided into:

  • estimation hypothesis, which deal with single-variable descriptive hypotheses. This is analyzed in a context. The researcher formulates a statistical estimate of the result.
  • statistical hypothesis of correlationwhich deal with correlation hypotheses, which are those that study the relationship between two or more variables.
  • statistical hypotheses of differences in means, dealing with group difference. Compare numerical estimates between two or more groups under analysis.