The 16 types of hypotheses (and their characteristics)

The scientific method and, therefore, modern science, was born in the seventeenth century thanks to Galileo Galilei, Italian physicist, mathematician and astronomer who was the first person to apply this methodology based on the observation of reality to progress in knowledge, thus marking the beginning of the scientific revolution, the divorce between science and religion and the establishment of modern science.

All the progress that we have achieved, are achieving and will achieve has, as a fundamental basis, the scientific method, a form of hypothetical-deductive reasoning that has essential properties so that knowledge can be classified as scientific: falsifiability (it can be refuted in a future) and reproducibility (the experiment, test or investigation can be replicated always having the same results). Thanks to the scientific method, there is science.

And although this scientific methodology has a total of ten sequential steps, all of it revolves around a key concept: hypotheses. Venture to give an explanation to the why of a phenomenon that we do not know by establishing predictions based on the data that we do know that, with experimentation, we are going to confirm or refute. Every great scientific discovery was, in its day, a hypothesis.

But are all hypotheses the same? No. Not much less. The hypotheses, depending on the environment in which they are developed and the procedures for working with them within the framework of the scientific method, can be classified into different groups. And in today's article, from the hand of, as always, the most prestigious scientific publications, we are going to see what kinds of scientific hypotheses exist.

What are the different kinds of hypotheses?

Hypotheses are assumptions based on data that serve to initiate an argumentation or a scientific investigation. In other words, they are predictions based on prior knowledge about a field that seek to give an explanation to a phenomenon that we do not know and that, through the steps of the scientific method, we will put to the test to confirm or reject them.

Thus, a hypothesis can be viewed as speculation or conjecture that, at first, lacks both confirmation and refutation. Therefore, these propositions, based to a greater or lesser extent on what we know for sure, serve as provisional formulations about something that we do not know that will be tested through experimentation.

These hypotheses, when confirmed as true (although they can always be denied in the future) allow us to formulate an interpretation of reality. And it is that when the hypothesis is always fulfilled, a scientist can deduce (remember that the scientific method is hypothetical-deductive) that the conclusion reached is universal.

The scientific method is based on the formulation of hypotheses and deductions. And these hypotheses, which, as we have said, are an attempt to explain something we do not understand and which are good for making predictions, can be, depending on their scope of applicability, of different types. Let's see them.

1. Descriptive hypotheses

Descriptive hypotheses are those that have the objective of discover the relationship between the variables within an investigation, but without focusing on explaining its causes.

2. Causal hypotheses

Causal hypotheses are those that claim explain the cause-effect relationship between two or more variables. They can be predictive (they predict how one variable will behave in response to another) or explanatory (they explain how one event is the cause of another).

3. Correlational hypotheses

Correlational hypotheses, also known as joint variation hypotheses, are those that determine how and to what extent one variable affects another and vice versa. These relationships can be positive (the higher A, the higher B), negative (the lower A, the lower B) or mixed (the higher A, the lower B; or the lower A, the higher B).

4. Group difference hypothesis

The group difference hypotheses are those that seek to anticipate the difference with respect to the behavior of different groups based on statistical comparison. For example, such a hypothesis could be that "the incidence of multiple sclerosis is higher in women than in men."

5. Estimation hypothesis

Estimation hypotheses are a type of statistical hypothesis (those that introduce statistical symbols to define parameters) that are responsible for make statistical predictions of an outcome. They are analyzed in a concrete context of a descriptive hypothesis of a single variable.

6. Statistical correlation hypotheses

Statistical correlation hypotheses are those that are responsible for statistically analyzing how one variable affects another and vice versa. Therefore, it is to apply the statistics to the correlational hypotheses that we have analyzed before.

7. Statistical hypothesis of mean differences

The statistical hypotheses of mean differences are those that are responsible for comparing the numerical estimates between two (or more) groups that we are analyzing. In line with the previous two, it is to apply statistics to the group difference hypotheses that we have seen previously.

8. Null hypotheses

The null hypotheses refer to all those situations in which, After conducting the experiment or investigation, no relationship was found between the variables that we have used as an object of study. For example, we arrive at a null hypothesis if our research concludes that "there is no relationship between red meat intake and increased risk of developing cancer."

9. General or theoretical hypotheses

General hypotheses, also known as theoretical hypotheses, are all those that are established prior to the study and conceptually. It is what most resembles speculation, since it arises from certain preliminary observations but without having quantified the variables. It is a prediction that becomes an object of study.

10. Conditional assumptions

Conditional hypotheses are those that are formulated assuming that the value of one variable depends on the value of two other. Let's say there are two cause variables and one effect variable that depends on both. For example, we would have "if a person does not exercise (cause 1) and follows a poor diet (cause 2), their risk of osteoporosis increases (effect)".

11. Relative hypotheses

Relative hypotheses are those that study the influence of two or more variables on another. We have a dependent variable and two independent variables, so we analyze and evaluate what relationship the dependent one follows with the independent ones.

12. Singular hypotheses

Singular hypotheses are those that, in contrast to the universal ones that we will see below, focus on a specific fact. They seek to be unique and specific to a very specific context, without the intention of becoming universal concepts that are always applicable.

13. Universal hypotheses

In contrast, universal hypotheses are those that do not claim to be simply singular, but rather they try to demonstrate something that is always applicable. Hence, they are called universals, since, if demonstrated, their range of application will be in the totality of that investigated.

14. Inductive hypotheses

Inductive hypotheses are those that are obtained through induction, so they have a less logical but more probabilistic nature. Starting from the observation of some particular cases, we want to establish some general conclusions. We go from the specific to the universal. We apply what we see in a very concrete case (a specific premise) to what, according to reasoning, should always apply.

15. Deductive hypotheses

Deductive hypotheses are those that are obtained through deduction, so they have a less probabilistic but more logical nature. Starting from universal premises, we want to reach specific or particular conclusions. As we have said, the scientific method is a methodology based on hypothetical-deductive reasoning.

Thus, this scientific method is divided into two components: hypotheses and deductions. The hypothetical part is based on analyzing specific cases to reach potentially universal conclusions that will define our hypotheses. After seeing something many times, we arrived at a hypothesis that could be universal.

After arriving at these hypotheses, the second part of the reasoning begins: the deduction. And these hypotheses are used as universal premises to see if, from that moment of the investigation, all the specific cases that we see meet these hypotheses. Only then, When the premise that made up our hypothesis is always fulfilled, we can deduce from there the name of the method that our conclusion is universal.

An example to understand it much easier. After seeing that many birds lay eggs (a succession of particular cases), we hypothesized (a potentially universal conclusion) that all birds lay eggs. And with this hypothetical conclusion, we will have to analyze whether each and every one of the bird species lays eggs to deduce that, indeed, our universal premise can be applied to all specific cases.

16. Analogue hypotheses

Analog hypotheses are those that are obtained by using the analogy resource. That is, we transfer the content of a hypothesis that we know has become universal to that of our study as long as they are similar enough so that they can be compared.