## Types of Samples

I highly recommend this website: http://faculty.elgin.edu/dkernler/statistics/ch01/1-4.html

*All of the images from this page are from the above site.*## Overview

When you want to know the about an entire population of individuals, you examine a smaller group of individuals called a “sample.” There are five types of random samples that can be taken: Simple Random Samples, Stratified Samples, Cluster Samples, Systematic Samples, and Multistage Sampling. When sampling, you must select individuals at random because randomization tends to lead to less bias.

## Simple Random Sample

Every combination of what is being assessed has an equal chance of being chosen. First you choose where the sample comes from by sample framing, which is an organized list from which the sample will be drawn from, then assign each individual in that sample framing a random number. Then using a random number generator, generate the number of individuals you need for the sample.

## Stratified Sample:

A stratified random sample is random sampling within each strata. A strata is when a population is divided into homogeneous groups. For example, splitting the four grades of high school into four homogeneous groups 9th, 10th, 11th, and 12th grade.

## Cluster Sample

Cluster Sampling is random sampling within heterogeneous groups. Each cluster will look similar to the population, and this practice is done when the population is naturally separated into clusters that resemble the population.

## Systematic Sample:

Systematic sampling uses a list of individuals. With this list, the individuals are given numbers in numerical order. A random number generator generates a number that will select the start of the list. From that number, the list will continue with every nth number.

Multistage Sampling:

Multistage sampling is when you use a combination of different types of sampling methods. This method is used when the population is very large.

How to Sample Badly:

You can sample badly by taking a voluntary response sample, having undercoverage, and taking a convenience sample. In a voluntary response sample, individuals are invited to respond and only the ones who responded are counted. Undercoverage is the portion of the population that is not sampled at all or has a smaller representation in the sample than it has in the population. A convenience sample is when you only include the individuals who are convenient to us.

Final Reminder:

A valid survey yields the information we are seeking about the population we are interested in. You first must know what you want to know, use the right frame, tune your instrument, ask specific rather than general questions, ask for quantitative results when possible, and be careful in phrasing questions.

Examples

__Example 1:__To try to gauge the freshmen opinion about the food served on campus, Food Services suggests that you just stand outside a school cafeteria at lunchtime and stop people to ask them questions. What’s wrong with this sample?

*Solution: This is a convenience sample which is likely to be biased. There would not be responses from people who use the cafeteria for dinner, and there would be no response from the people who hate the food so much they don’t come to the cafeteria.*

__Example 2:__In Kinnick High School, I have decided to take a sample on how much students sleep every night. I’ve split the population in the high school into groups such as students who live in Negishi, Ikego, Off-base in Yokosuka, and on base in Yokosuka. What type of sample did I take?

*Solution: Stratified Sample.*