ABSTRACT
In inferential statistics, we shi the focus from numerically describing data to examining relationships among the characteristics of interest and underlying phenomena. e task of classical inference is to learn something about an unobserved population through a sample statistic. Inferential statistics allows the generalization of ndings from the sample data to the parametric population from which the sample is drawn. Mohr (1990) stresses that in making such inferences, a sample is not good enough. Sampling distribution is an extra tool that bolsters condence in a study’s inferential power by calculating the probability distributions of all possible sample sizes to reect the distribution of the total population. Sampling distribution forms a
kind of bridge between the sample and the population. In the sampling distribution, certain bits of information about the sample and the population come together in such a way that one can make the desired statements about population parameters with varying degrees of condence and with varying degrees of accuracy. (p. 13)
Central Limit eorem: Sampling Distribution
In sampling distribution, the unit of analysis is no longer the person or variables, but the collective. e central limit theorem plays a key role in this transition of probability study. Sampling distribution is a hypothetical scenario in which innite samples are drawn from the same population. e central limit theorem states that the hypothetical mean of all randomly selected samples is also normally distributed, meaning that the sample mean, the mean of all sample means, and the population mean are all closely related. Inferential statistics, therefore, makes it possible for one to estimate the parameters (or characteristics) of a population from the statistics of a sample. e estimates, based on the laws of probability, reect best estimates rather than absolute facts.
