Tip:
Highlight text to annotate it
X
So because it usually doesn't make sense for us to, you know, try to talk to
everybody to do a census. usually we need to create a sample.
And so there's a variety of ways in which we create a sample and so let's talk
about those. And let's talk about them through
examples. So suppose that we could somehow identify
all the likely voters in the state and we want to, create a sample of all those
voters. In order to ask 'em how they're going to
vote in the next election. So, one option would be to write all
their names on a piece of paper and toss those in a really, really big hat.
and then you know, put the bunch, all these little slips of paper, all these
names in the hat. And then draw a thousand slips out of the
hat. kind of like a lottery.
So, this sampling method is called A simple random, sample.
Simple random sample. Simple random sample means that all
options are equally likely to get selected no matter what, it is,
equivalent to putting a bunch of names in a hat.
Now of course, people don't actually use large hats, instead they Do things like
put all the names in the computer and have the computer randomly select name
and that would be a simple random sample. Now that's not always practical though or
sometimes it's not even desire. So for example, suppose that in a
particular state previous data has adjusted that the electorate was
approximately 39% Democrats, 37% Republicans and 24% independents.
And so in a sample of 1,000 people, let's say we're giving a political poll, we
want to make sure that we go ahead, get about 390 Democrats, 370 Republicans and
240 independents. Now one option would be to go and do our
simple random sample. and just sort of hope that everything
works out okay. because changes are it probably will come
out pretty close. But another option would be to select,
randomly select from 390 people known to be Democrats, 370 who are known to be
Republicans, and 240 for those no political affiliation.
This is assuming that people have somehow indicated their party affiliation.
So this is a method called, this is a method called stratified sampling.
The idea behind stratified sampling is that we take our entire population and we
divide it into groups. So we divide our entire population into
groups, and then we sample and proportion it to our, our, our desires.
And so I'm going to randomly select, I'm going to randomly going to select 390 of
these, of these democrats. And I'm going to randomly select 370 of
these, of these republicans, and 240 of these independents.
And so, we've. Divided our population in to groups and
then we randomly select from inside each of those groups.
Very similar to this is an idea called quota sampling so this is called quota
sampling. And it works the same way, except instead
of starting by dividing the population into groups, we just sort of start
randomly sampling people. So we go, okay, I'm going to grab you and
you and you and you and you and you and you, and we keep on going...
Until we meet our quota of 390 Democrats. So let's say I sampled, you know, 600
people, and I've already talked to 390 Democrats.
I call up the next person and say, hi, are you a Democrat or Republican.
And they say, I am a Democrat. And I say, I'm sorry, I already have
enough Democrats, so I'm not going to talk to you.
And that is the idea of quota sampling.