This essay aims to explain why psychologists ‘fail to reject’ a null hypothesis rather than ‘accept’ it. In general, psychologists appear to avoid the verb ‘accept’ and they usually replace it with the phrase ‘fail to reject’. Furthermore, the only way to understand this tendency is through the science of philosophy. Therefore, the following essay is going to include the main philosophical theories that seem to be useful for the understanding of this phenomenon. These theories are Popper and Hume’s suggestions.
One of the most representative human characteristics could be said that is the strong human desire to explain every single event that takes place, a tendency that can be placed back in the ancient time. People always try to detect the reasons, the sources and the components that create particular behaviours. From this human desire was created philosophy, which tries to give an answer to all these questions. Philosophy aims to the improvement of the quality of life.
However, as it is based on observation and empiricism, there have been formed many theories and some of them are contradictious as they give totally different explanations to the same questions. Consequently, it has sometimes created great confusions and debates, instead of providing a clearer insight to their field of study. These ‘confusions’ were the basic motive for many scientists and philosophers to question the existing methods of research and try to provide certain patterns that could help them to the formation of more essential conclusions.
Karl Popper and David Hume, were the two philosophers who mostly influenced the way psychologists and of course most scientists commit their research and theorise on most subjects. Their suggestions had also a great impact in the way they express their results (‘fail to reject’ rather than ‘accept’). In order to understand Popper and Hume’s suggestions, which are going to follow, it is important to define first what a hypothesis is and afterwards explain its importance in a psychology-study.
Hence, hypothesis is ‘a formally stated expectation about a behaviour that usually relates to some other behaviour or influence’ (Heiman, 1998. p. 13). Furthermore, hypotheses are divided in two main types, which are known as the casual and the descriptive hypothesis. A casual hypothesis tries carefully to recognise a certain reason or influence on an attitude and a descriptive hypothesis carefully describes behaviour in terms of characteristics or the situation in which it occurs and in addition, it allows us to predict when it occurs.
However, debriefing the definition and meaning of the tem hypothesis is not enough, as it does not exist independently from research. Therefore, it is important to explain also the two main kinds of research, which is experimental research and descriptive research. As far as the experimental studies are concerned, they are usually conducted when the scientist wants to test a casual hypothesis. Testing a casual hypothesis is achieved by the manipulation or change of a situation (the manipulation of a situation is known as the independent variable (IV)).
This happens, so that experiment convenor will create the ideal conditions in which the subject’s behaviour is going to be elaborated (the reaction of the subject is known as the dependant variable (DV)) (Heiman, 1998. p. 13). On the other hand, ‘ descriptive studies are used to demonstrate a relationship, predict behaviour and describe behaviour or participant. ‘ (Heiman, 1998. P. 46). This kind of a study leads the experimenter to a better understanding of the participant’s personality and therefore he/she is able to infer at some level the participant’ s behaviour.
Finally, if reactions are observed into a wider group, then it is likely to reach the assumption that there is a relation between the particular cause and reaction. Hence, a descriptive study is mainly used to test a descriptive hypothesis. It is obvious that both experimental and descriptive designs obey to a particular skeleton and that the experiment convenor is careful about conjecturing based on the results of those studies. This situation of ‘wariness’ and deliberation could be argued to maintain Popper’s beliefs as far as scientific research is concerned.
In his work, which was published in the 1950’s and 1960’s, Popper divided theories in two wide categories, ‘good’ and ‘bad’ theories. He argued that, a good theory allows the experimenter to predict attitudes that will be able later to be proved either false or true. In contrast, to a bad theory, which can not be neither verified nor falsified. Therefore, he claimed that scientists have to form theories that can, with further research, be falsified. This approach ended to be known as Popper’s principle of falsifiability.
Moreover, it is important to be noted that by falsifiable Popper did not intend to say that a theory must be false to be a good one. Instead he meant that the nature of predictions generated that theory needs to be such so that, with subsequent trials or observations, they can be supported or rejected. Ambiguously, in that sense, the principle of falsifiability implies that a theory is good even if its predictions are clearly erroneous. Therefore, the suggestion that ‘the oceans will go dry tomorrow’ is a good theory, while the belief that ‘ God is omnipresent’ is a bad theory (Internet, p. ).
Moreover, it is not enough to say that a theory is good because it is falsifiable. All scientists are aware of the fact that it is required to provide strong evidence if they wish to conduct research about a particular phenomenon. In other words, they basically have to find a strong correlation between the cause and the effect, in order become right for their predictions. In this field David Hume had a great contribution, as he understood and stressed the importance of defining the causes of actions or other phenomena.
Therefore, he suggested three criteria that have to be fulfilled when inferring causation. The first is that there should be ‘temporal precedence of the cause’. In other words, the effect must come after the cause in time and not before. Secondly, there needs to be ‘covariance between the supposed cause and effect’. By this, Hume meant that whenever the cause occurs, the effect should occur as well and when the cause does not occur, neither should the effect. In case that this criterion is not fulfilled, the most logical thought is that the effect has more causes, alternative causes.
However, the most significant criterion of Hume for inferring causation is the third one, which is known as the exclusion principle. This criterion aims to define the cause of an effect, in order for this to happen it is necessary to exclude every other potential alternative. Although this criterion may seem easy to put it into practice, as the researcher is the controller of the variables, it is not. Psychologists are aware of the complexity of human personality and therefore they are able to understand the difficulties in completely satisfying the conditions of the exclusion principle.
Like Popper claimed, it is impossible to keep out every single-potential alternative cause. Therefore, psychologists avoid using radical statements for the demonstration of their results. Moreover, they usually employ statements such as ‘fail to reject a hypothesis’, ‘there is a strong suggestion that’ or ‘the data indicate that A causes B’. In addition, the exclusion principle complements Popper’s principle of falsifiability. That can be understood as, since we cannot exclude all possible alternative causes, we can only be certain when rejecting a cause establishing, therefore, the falsifiability of our theory (Internet, p. ).
Furthermore, what is important to be clear in this essay are the implications of Popper and Hume’s suggestions for these two kinds of research in the field of psychology, experimental and descriptive research. As it was explained in the above paragraphs, experimental studies must have falsifiable outcomes, therefore researchers introduce a new term, the null hypothesis. The aim of a null hypothesis is to be rejected so that the theory will be plausible, however the falsifiability of an outcome is also determined by other later research, which will either verify or contradict the result of the last experiment.
Furthermore, it is important for an experiment to satisfy the three criteria of Hume about causation in order to show correlation between the cause and the effect. The first two criteria for temporal precedence of the cause and covariance between cause and effect are easy enough to put into practice. That is also the reason for which, the independent variable must precede the dependent variable in time in a study and additionally, the changes made in the IV must be associated with changes in the DV.
In contrast to the last criterion that is almost impossible to be achieved, the only way to reach the best result is to try to establish experimental control through random selection of participants and their random assignment to the conditions of the study. Popper and Hume’s suggestions on experimental research are easier to be understood, if they are demonstrated in an example. In this case a good example could be the experimental study of ‘social projection’ that took place during a practical in the academic year 2000-01 for the requirements of the sp300 module at the UKC.
The main hypothesis in this study was that people have a tendency to assume that others share their own attitude position. The experiment was carried out to define not only if social projection occurs but also to define the causes of this tendency. Furthermore, so that the experiment convenor would achieve experimental control, the participants were assigned to different conditions and were asked to consider the attitudes of either a group of psychologists (with which the participants identified) or another group.
In this study the independent variables were the condition and the individual’s attitude, and the dependent variables were the consensus estimates. In this experiment, Popper’s principle of falsifiability is satisfied since the results will show whether social projection occurs or not. In addition, this study allows us to make inferences about cause and effect covariance, meaning that if the hypothesis is accurate the consensus appraises in the two conditions will be different.
Nevertheless, we cannot be completely sure about causality since there could be other extraneous variables, which the experimenter failed to exclude. Therefore, the experimenter is going to express his/her assumptions with a bit of deliberation, as there is a possibility for the results to change by the discovery of other extraneous variables in later studies. However, in descriptive studies the data are not the same as there is no manipulation of variables, consequently there is lack of experimental control and therefore the possible alternative causes can not be eliminated.
Hence, this kind of designs can not easily match with the theory of falsifiability and Hume’s criteria for inferring causation. As it can be assumed from the above information demonstrated in this essay the principle of falsifiability is extremely important for statistics, as it appears to be dangerous using radical term for expressing results. This happens because it is impossible to state something by being completely sure that it will not be rejected by later studies, so the principle of falsifiability in statistics is known as the ‘null hypothesis’.
The null hypothesis what the experimenter would expect if things were to happen randomly and he/she tries to reject. For example if we hypothesise that all the ships are white, the null hypothesis is going to be that there is at least one ship that is not white. When the research is conducted, the researcher will be able to reject the null hypothesis only if he/she has not found a single ship that will not be white in the sample that he studied. However, he/she want accept the hypothesis as it is much likely that later on another study might find a non-white ship and reject the hypothesis.
This very simple example is able to show the importance of the terminology that is used to explain statistical results. Evidently, Popper and Hume’s propositions, although they are not completely applicable in every kind of research, still have a great importance in the field of statistics. The reason of their suggestions’ importance is because their ideas contribute at high level in the way scientists perceive and conduct research. Finally, with their theories they make clear the importance of the existence of evidence and careful observation on which a theory should be based on.