The hypothesis testing of a null hypothesis depends on its method of meaning testing. The final remarks for such a test are either to define the null hypothesis as true if it exceeds the critical value based on an alpha risk that the null hypothesis was false, but the hypothesis is false if it falls short of the critical value. (Cohen, 1990) Argues that the conclusion of such a method of research is not at all conclusive, as it raises other questions rather than providing a solution. In this case, if the null hypothesis is false was is the true nature of the research, at the same time if the null hypothesis is true to what extent does it facilitate a conclusive remark to be generated for the study.
Both literary sources by (Cohen, 1990) and (Schmidt, 1996) find several loopholes within the utilization of null hypothesis as a research method, while (Cohen, 1990) focuses on the validity of the probability tools utilized and the significance testing methodology utilized, (Schmidt, 1996) concentrates on the development of the study based on the two types of errors that are persistent with this method of research study.
There are two types of errors namely type I and type II. Type II error occurs when the null hypothesis is deemed false yet in actual existence based on facts it is true. (Schmidt, 1996) questions the validity of these errors as he says the result of analysis of almost all null hypothesis is that they are deemed false hence type I error cannot occur in such instances as one cannot declare an event that is not in existence.
Work by (Cohen, 1990) later on brought into light the non-centrality nature of variables for the hypothesis testing method currently in play. He then found a way to decompose this non-centrality into aspects of effect and sample size that is when he developed the power analysis technique that utilized four parameters; the alpha significance criterion, the power of the test, the sample size and, the population effect size. This method showed that it was now possible to prove a null hypothesis.
In the psychological criteria of statistics, results and analysis go beyond the standard figures as if it were so computers could analyze the provided data and give the necessary output of the study. This being social science different factors have to be considered when conducting any research and the validity of the methods used. Null hypothesis if viewed from a mathematical perspective would realize the values that correlate in approval or disapproval of the statement based on the sample size. However, to what account do the values help us make an informed judgment on the topic matter. This makes it necessary to associate the method with an analysis that proves the null hypothesis from a social science perspective. Basing a conclusion from the majority in an aspect of true or false does not facilitate the realization of the set objectives for any study or research. The statistical analysis deployed that utilized the measure of p values also does not give any theoretical account of human behavior that is vital in any psychological research (Cohen, 1990).
Significance testing never account the for the effect size that is critical in social sciences, but a utilization of correlation analysis facilitates this process through the output of an r value, which is a measure of effect size and can be later transformed to a T or F and utilized for significance testing while in the previous testing an output of an F value could not give any account of the effect size.
Cohen, J. (1990). Things I have learned (so far). American Psychologist, 1304.
Schmidt, F. L. (1996). Statistical significance testing and cumulative knowledge in psychology: Implications for training of researchers. Psychological methods, 115.