The Examining of Generalization Quantitative Scientific Findings by Using the Jackknife Method: an Application

Abstract

The outcomes which cannot be generalized are specific for a sample but am unable to be reflected to the rest of the population. The parameters that are reached at the end of the statistics that are scarce in sample arise doubts in the aspect of generalization. In these cases, parameter estimation may not be very stable and outlier values can produce false outcomes in the model. The situation that there is sample size is not adequate enough in numbers, reliability and generalizations that are reached at the end of restricted sets of data necessitate questioning with a dubious approach. Some statistical methods test the confidence intervals of the parameter values of the samples and the generalization of the parameter estimation values of the samples. These methods are known as 'resampling" or "domestic-self copying." Jackknife is one of these methods. The sample in this study is composed of 18 students. The status of the "self-respect" of individuals in making decisions has been tried to be determined for the sample by the application of the Melbourne Decision-making Scale I. In this research, the dependent variable was determined to be the total item score related to the self-respect manners of Individuals whereas independent variables were determined to be the academic success averages of the students (acdave), the financial income of their families (economy) and the number of siblings (sibling) they have. It has been seen that the academic success independent variable has a considerable effect at a significance level of .05 on the dependent variable of self-respect dependence of decision making (p < .05) and Jackknife has confirmed this generalization.