This document discusses the concepts of sampling distribution, standard error, and sample size. It explains that sampling distribution refers to the distribution of sample means when multiple samples are taken from a population. The standard error is derived from the sampling distribution and provides a theoretical basis for estimating sample size without actually taking multiple samples. It also describes how standard error decreases as sample size increases, following the central limit theorem. Finally, it discusses how standard error can be used to calculate the minimum required sample size for studies using means or proportions as key variables.
This document discusses the concepts of sampling distribution, standard error, and sample size. It explains that sampling distribution refers to the distribution of sample means when multiple samples are taken from a population. The standard error is derived from the sampling distribution and provides a theoretical basis for estimating sample size without actually taking multiple samples. It also describes how standard error decreases as sample size increases, following the central limit theorem. Finally, it discusses how standard error can be used to calculate the minimum required sample size for studies using means or proportions as key variables.
This document discusses the concepts of sampling distribution, standard error, and sample size. It explains that sampling distribution refers to the distribution of sample means when multiple samples are taken from a population. The standard error is derived from the sampling distribution and provides a theoretical basis for estimating sample size without actually taking multiple samples. It also describes how standard error decreases as sample size increases, following the central limit theorem. Finally, it discusses how standard error can be used to calculate the minimum required sample size for studies using means or proportions as key variables.
Shital Bhandary Assistant Professor DCHS, PAHS-SOM Review Preview • Sampling distribution of means • Standard Error of Mean • Sample size and standard error Objectives-1 Explain the concept of standard error: – Describe a sampling distribution of means – Define the standard error of the mean – Discuss the use of the standard error of the mean in relation to a sampling distribution of means. – Identify other situations in which standard errors are used Sampling distribution of means • Statistics uses samples to make inferences about the population • And It is always better to make inferences using many sample (than taking a simple sample) from the population under study (?) • If many sample are taken then each sample will have a mean (same or different) • The distribution of such means is known as “sampling distribution of means” Mean, Variance and Standard Deviation of Sampling Distribution Central Limit Theory: Revisited • It states that: Given a population with a finite mean μ and a finite non-zero variance σ2, the sampling distribution of the mean approaches a normal distribution with a mean of μ and a variance of σ2/N as N, the sample size, increases. • In other words, regardless of the shape of the parent population, the sampling distribution of the mean approaches a normal distribution as N increases. Sampling distribution w.r.t. sample size Quick think! • Why CLT is important? – We need to take multiple samples to make our sampling distribution normal – We need increase our sample size to at least 30 - 50 samples to get a normally distributed data (for a variable under study) Why normal distribution is important? • Normal distribution is most widely studied and used “continuous probability distribution”
• It is the basis of many “parametric” statistical
tests: t-test, ANOVA, correlation, regression etc. N.B. • We use the “standard error” along with the “standard score” to calculate the minimum sample size required for the study • The “standard error” provides the theoretical lieu for NOT taking multiple samples as it is actually derived from the sampling distribution of means • Yet, this minimum sample size is usually fulfilled by taking multiple samples (e.g. clusters, stratas) whenever possible Quick Think! • Can other statistic like proportion, median, quartiles, skewness etc. can have standard error?
• If yes, how can we interpret them?
Objectives-2 Describe how sample size affects standard error – Identify the formula for the standard error of the mean and use it to find the sample size – Illustrate the relationship between sample size and standard error – Calculate sample size using proportions – Discuss various methods used to calculate sample size in medicine and health research (integrating with the epidemiology, demography and other disciplines) Standard Error and Sample Size • Two most commonly used sample size determination are based on: – Mean – Proportion • Both of them use standard error (why?)
• Both show that as N increase standard error
decreases and vice-versa (why?) Review 1 • Sample size calculation using mean • Key variable: Age at first marriage = 17.5 years • Representative Population: Nepal • Assumptions: – age at first marriage follow a normal distribution with larger sample size – Level of significance = 5% = 0.05 – Margin of error = 1 – 5% • Minimum sample size = ? Review 2 • Sample size calculation using proportion • Key variable: Childhood Overweight Prevalence = 4.5 % • Representative Population: Nepal • Assumptions: – Childhood overweight prevalence follow a normal distribution with larger sample size – Level of significance = 1 – 5% – Margin of error = 1 – 5% • Minimum sample size = ? N.B. • We will come back to the concept of minimum sample size many times • Because we need to analyze the data in terms of not only a single sample but also multiple samples (and single variable with multiple categories!) • In multivariate analysis like regression analysis, it is essential to know how many sample should each variable MUST have during the modelling Thank you!