<p>1) What is the difference between stratified random samples and multistage random samples? They sound the same except for the fact that multistage random samples have groups that are redivided. </p>
<p>2) Also, can anyone explain exactly how to sample a multistage sample?
Here is my impression of what it is:</p>
<p>Group 1 divides into group 2 and 3 which then divide into 4,5,6,7. Then we sample the individuals in groups 4,5,6,7. Do we also sample group 2 and 3? But, in reality 2 and 3 are kind of already sampled when we sample 4,5,6,7.</p>
<p>Am I right? Can some one clarify my two questions?</p>
<p>Stratified sampling is when you take the population, divide it into certain groups, and then poll people from a randomly selected group. Multi-stage sampling involves continuous randomization ie “weeding out.”
The diagram of the levels is basically correct for multi-stage, except that you only sample one of the groups (4/5/6/7)- you dwindled it down so that you could pick one of them that has been randomized several times. If you polled all of them, it woudl be the same thing as level 1 on your diagram.</p>
<p>It is also important to know that the criteria upon which the stratified sample was based was meant to be equivalent to the population (if that makes any sense at all). </p>
<p>Say you have a village of 300 people, 100 of which are under the age of 18. If you wanted to organize a research study with 30 people, and stratify the sample based on the different age groups. You would probably want to have 10 people under the age of 18, and 20 over. </p>
<p>This is a really simple example, and most stratified samples are more complex than this.</p>