Geography fieldwork sampling techniques

Discovered in China
13 followers
2 Video Views·Nov 26, 2024
A quick over view of the three sample techniques you could apply to geography fieldwork. Normally I enjoy bringing the jelly beans into school to do this but as we are in lockdown I guess I will have to eat them all myself.
Okay today we are going to have a look at different sampling methods that you can use
on any piece of geography fieldwork and because I'm doing this at home in lockdown I have decided
to use these as my sample of data. There may be one or two less by the time I finish recording this.
Okay there we go we have Aldi's finest jelly beans. Okay there we go we have
Aldi's finest jelly beans but we could imagine that these jelly beans are people in a population,
pebbles on a beach, pebbles in a river, it doesn't matter what data set you are using
we can apply the same rules of sampling to any data set. So I'm going to create a null
hypothesis and hypothesis to test out as we go. My hypothesis is going to be
that all jelly beans are pink and my null hypothesis is going to be that not all jelly
beans are pink. So the theory always goes or the advice is to test against your null hypothesis
because when you're going out and collecting your fieldwork data all I need to do is find
one jelly bean that is not pink so this blue one to support my null hypothesis that they're not all
pink. If I was going to go out testing against my hypothesis I would need to collect all the
jelly beans or all the pebbles on a beach or all the people in your sample area to prove or disprove
the hypothesis so I'd need to get all of these jelly beans to prove that they are not all pink.
Okay let's begin first of all with systematic sampling. So systematic sampling works by going
on even intervals so it could be every 10th person in the population, it could be every 10th building
on a street, it could be every 10th pebble that you pick up, it could be every 10 meters distance.
So it doesn't really matter what scale you are using that would be systematic you're doing it
every 10 meters or every whatever. So using my jelly beans I'm going to pick every fifth jelly
bean to pick out of my population to be the jelly beans that I will sample. I might interview them,
I might survey them, I might give them a questionnaire but they're going to be my
sample set of data. So one, two, three, four, five, one, two, three, four, five, one, two, three, four,
five. So they are the three jelly beans that I would use as my sample set. Okay you might have
already spotted the first problem with systematic sampling is that it's purely by chance
that the three jelly beans selected from my systematic sample happen to be blue and that's
given me a misrepresentation of that entire population because not all those jelly beans
in that population are blue jelly beans but the ones I am sampling happen to be. I have overlooked
the four jelly beans in between. If I had done a smaller scale perhaps every three jelly beans
I would have got a better selection where I can see a better representation of the population. I
can see a better selection of colours. So you need to think really carefully when you're selecting
your systematic sample size that you get a broad enough range of data that you can cover a wide
area but you don't want it too big that you might overlook some little details in between and end up
with a sample set of all blue jelly beans. Okay so for that null hypothesis which was not all jelly
beans are pink I can accept my null hypothesis that not all jelly beans are pink because what
I selected out was blue jelly beans and I reject my hypothesis that all jelly beans are pink.
Let's try now stratified sampling. So stratified sampling is when you pick out a strata or a sub
group within the data you're collecting. So for population your strata could be age, gender,
ethnicity, religion, there's loads of things that you could use. For looking at my population of
beans again I want to make sure that I get a representative from every jelly bean colour
so that they are all equally represented. So I will need a yellow one, an orange, a red, a blue,
a little light green one, pink and a dark green. There we go. So that is now my
sample group and they're the ones I will use to interview or survey or assess.
So what I have ensured there is that I've got one representative from each of the subgroups
and from that I can now see that my null hypothesis will be accepted because not all
jelly beans are pink. So you would need to go into your research knowing full well
what your subgroup categories are. We don't have the issue of picking out just all blue jelly
beans like we did with the systematic sampling but there is quite a big degree of researcher
bias behind this and I've decided what the subgroup is going to be. So I decided that I
would choose this based on jelly bean colour but whatever you decide that's got bias behind it.
So you might choose age as a factor to investigate in population so you're putting bias behind that
by selecting out age as being a key factor when the key factor could be something else.
Okay the final method is random. Now this could be used, I find it easiest online using a random
number generator. If you're using this to try and identify a location, so a place or a site
where you're going to collect data, you can use grid references from an OS map or you could use
longitude and latitude or you could overlay any map with your own grid that you've designed
and put numbers on. You could use postcodes as well and what you'll do is input your
number ranges onto a random number generator and that will pick out your locations.
Okay so let's try that. So I have got 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15 jellybeans. So on the random number generator I will need it to select numbers from 0 to 15.
Okay so the first number picked out was 15 which is the end one, then it was 9, 1, 2, 3, 4, 5, 6,
7, 8, 9 and then the third one 1, 2, 3. Okay so what's great about this method is you have got
absolutely no bias as a researcher, you've got no control over what's being selected but again a
little bit like systematic, purely by chance I've got one blue jellybean and two reds and nothing
from the other colours being represented. So this is the best in terms of removing any researcher
bias but just be aware that you haven't got that control and your population that is selected
might not be representative of the whole population that you're researching.
So I hope that helped and I hope it made sense. Make sure that whatever you do you start off with
a hypothesis and a null hypothesis, that you really consider what sampling strategy is going
to be best for you, whether you need to have one representation from each subgroup in a population
or a beach, whether you want to be completely random about locations or what's chosen
and then you can test your hypothesis using your sampling techniques. So I hope it's helped,
I'm off now to enjoy my jellybeans and I will see you next time.
