The null hypothesis, HO is the accepted fact that researchers work to nullify or disapprove after a critical analysis of the data. Significant testing is a traditional form of analysis which has been said to hinder the advancement of information in scholars. Meta-analysis is the best method for determining statistical analysis. The point estimate can also be used to integrate multiple studies among learners.
Significant flaws in null hypothesis
The t-test, F test, chi-test and variance analysis are the traditional means used in interpreting data. These methods have majorly focused on Type I errors also called alpha error while less effort is put on Type II errors also called beta mistakes. The beta error shows a high level of mistakes ranging from 5%. The alpha error is always zero (Schmidt, 1996). Type I error increases chances of discovering false results in statistics. In another case, one cannot solve the multiple test problems using Bonferroni maneuver (Cohen, 1990). For instance, the division of 0.05 and 60 results to a figure that is not accurate of 0.00083 thus giving a double-solid t nearly reaching 4%. Such result may not be significant to provide any interesting ts.
Traditionally, getting the information of a learner was carried out using a method of point estimates. This usually caused a problem error of 50% confidence interval (Schmidt, 1996). Sir Ronald Fisher made psychological researchers to stop applying significance testing due to many problems that were experienced.
Many variables leads to information that was not supposed to be published in books or circulars for scholars to use. When the variables that do not depend on another one increase, its components will also double (Schmidt, 1996). A less-is-better idea of data analysis is preferred in representation as well as giving information. Most variables are balanced, and the required outcome is achieved by carrying out multiple regression analysis and checking on the z score predictor with the help of Type I errors. +1 is a positively related predictor while -1 is a negatively related predictor (Cohen, 1990). Null hypothesis results in artificial differences in sizes as well as correlation in entire learning.
Meaning of the flaws of null hypothesis testing in psychology
The meaning of data has experienced numerous critical errors as psychologists have been focusing on statistical significance testing hence retarding the growth of cumulative knowledge (Cohen, 1990). For example, the application of social science analysis in responding to essential issues occurs while using null hypothesis testing. No accurate results will be shown when determining the exact number of students in public universities who are above age 25.
The null hypothesis is not appropriate in psychology since there are contradicting results and sometimes leaves problems unsolved. For instance, it may be difficult to answer whether the age and education level determines the type of job one has to do (Schmidt, 1996). Furthermore, if medical practitioners say that some medicine may not treat patients efficiently, null hypothesis may mislead by stating that such medicines are poisonous to human health. Meta-analysis is today used to publish journals of primary study (Cohen, 1990). Social, psychology, medicine, and finance are also the fields that have shifted from primary survey to meta-analysis.
Summarily, statistical significance testing should be abandoned in teaching graduate students since it has numerous deficiencies in analyzing data. Instead, meta-analysis should be used because it is efficient in analyzing correct data. However, it is true to state that null hypothesis has triggered scientists to make further discoveries in social sciences as well as physical sciences basing on the arguments it provides.
References
Cohen, J. (1990). Things I have learned (so far). American Psychologist; American Psychologist, 45 (12), 1304.
Schmidt, F. L. (1996). Statistical significance testing and cumulative knowledge in psychology: Implications for training of researchers. Psychological Methods, 1(2), 115
