Samplind distributions
Sampling Distributions
Overview
Sampling distributions are critical in understanding how statistics behave in repeated sampling.
Key Topics
Sampling variability and distributions
Sampling distribution of a mean and Central Limit Theorem (CLT)
Sampling distribution of the sample proportion
Finite Population Correction (FPC)
Bias and noise
Descriptive vs Inferential Statistics
Descriptive Statistics:
Describe, show and summarize data in a meaningful way.
Inferential Statistics:
Infer or predict properties of a population based on data from a sample.
Statistical Estimation
Estimating an unknown mean (𝜇) by selecting a simple random sample:
Use the sample mean to estimate the population mean.
Different samples yield different sample means, showing sampling variability.
Goal: Understand what happens with many samples to capture this variability.
Sampling Variability
The variability in the values derived from repeated samples.
Two Main Areas of Inferential Statistics
Estimating Parameters:
Point estimate: Average wage in the UAE
Range of likely true population mean values.
Hypothesis Testing:
Compare the proposed hypothesis with observed data to see if results are significant (90%, 95%, or 99% confidence).
Statistical Estimation Examples
Average salary estimation for junior data scientists:
Point estimate and range example with computed values based on sample data.
Toward Statistical Inference
Understanding the sampling distribution of possible statistic values is essential for statistical inference.
The sampling distribution includes all possible values and their frequencies.
Sample and Sampling Distribution
Distinction between:
Sample Distribution: Distribution of observed or measured individual observations.
Sampling Distribution: Theoretical distribution of a sample statistic (mean).
Standard Deviation vs. Standard Error:
Standard deviation users sample spread, while standard error measures the estimate's precision.
Developing a Sampling Distribution
Assume a population size N = 4, with random variable X as the number of business meetings.
The random variables are independent and identically distributed (i.i.d).
Sampling Distribution Properties
Estimation of sample mean:
Defined by all sample means.
Standard Error of the Mean
Indicates variability in sample means:
Depends on sample size (n).
Characteristics of Normal Distribution and CLT
If the population is normal, sampling distributions will also be normal:
Sample mean is an unbiased estimator of the population mean.
Central Limit Theorem states that sample means are approximately normally distributed, even if the population is not.
Application Examples
Cereal Packet Weight Example:
Demonstrating sample means and probabilities using Z-scores.
Insurance Industry and the Law of Large Numbers:
The theorem assists in assessing risks over large numbers.
Bias and Noise in Estimation
Bias: Difference in accuracy of estimates.
Variability: Spread of the sampling distribution influenced by sample design.
Aim for low bias and low variability using random sampling and appropriate sample sizes.
Exercises and Applications
Problems to assess understanding of sample proportion distributions and probabilities in varying scenarios (e.g., tax increase vote support).
Statistical methodologies to determine required sample sizes based on desired probabilities.