Exploring Sampling Methods for Quality Testing and Surveys

so we've got these three different uh things we want to find out i'm going to see which sampling methods might be the

best ones to use and i've got a reminder of them at the top so the first one says you wish to test

light bulbs produced by a factory in a daily batch so there's a couple of different things that i think

might be useful to try here andrew what do you think would be a good one for this

the factory one yeah i tend to agree with you i tend to agree that i think systematic sampling might

be the best one now we'll talk about some alternatives but why do you think systematic sampling

could be good yeah it's very it's really convenient because with

if you imagine a factory and you've got like all these light bulbs and you want to just take every 10th light bulb or

one every hour that's quite an easy thing to do so it's pretty convenient and it's also going to be quite

inexpensive but there are some alternative ones that you could say that would that would also

be quite good were you going to say something different

yeah so i think the one that you'd suggested before amina was that it was going to be a simple random sample

a random sample technically you'd need to have a sampling frame and you'd need to be able to pick out

the random numbers that match the light bulbs or alternatively you could

just be like at the factory line and you could just randomly pick them but that's quite similar to the

systematic one i think the systematic one is a little bit better because you'll have them spread out over

the course of the day whereas the random one it's it's a bit difficult but

either random sampling systematic sampling there's just a few that obviously wouldn't be very good

you probably wouldn't want to do stratified sampling because what would the strata what would the

groups even be i don't know the times that they're being made but then i guess systematic

sampling is also going to spread that out as well yeah you're right there's

a systematic sampling you could end up if you did every fifth light bulb that might come from one particular machine

so there's flaws in all of these things and it's quite useful to think of what the flaws could be

because that's the kind of little one mark question they might say every light bulb

picked was from the same machine what is the floor and you can be like it's only testing one of the machines

so it's not always going to have like a hard and fast answer for these what about this one you wish to survey

consumer opinion on your new drink fizz guys released in the uk what did you think for this one

yeah i tend to agree with you i think that opportunity one opportunity sampling would be pretty

good for this um i know beforehand i mean you'd suggested stratified something which is

a good idea but it would be quite a lot more expensive and you'd need to have a

sampling frame of the whole population to think about how you wanted to split them up

so i think the good idea with an opportunity sampling is it's again it's going to be the same

advantages of it being convenient and it being inexpensive and the thing about it being inexpensive

is because you probably wouldn't want to do opportunity sampling for

when you really want it to represent the whole population but here i'm just interested if people like a new drink

like i don't i don't want it to cost a lot of money because it's not sorry to people whose job is to do

things like this but it's not that important in the grand scheme of life is it if someone likes your new

drink so it doesn't matter if it's not perfectly representative of the whole of

society because we just want to know if people like to drink or not so i think

opportunity sampling is because it's convenient and it's inexpensive you know you could just be stood outside

a supermarket but you know what if you can pick stratified sampling and you're able to

convince the examiner that stratified sampling was important then i think that would be good but

maybe they'd said something in the question like they don't have a large budget then in

that case stratified sampling wouldn't be very good because they don't have a sampling frame and then our last one

here when you want to find out what your favorite tv programs in the school that's representative of each year group

this one is pretty obvious this one has to be stratified sampling

and the reason it has to be stratified sampling is because we want it to be

representative and what else is something that's probably quite good to do stratified

sampling if you're in a school what do we have access to that we need for stratified

sampling good and we have the sampling frame so i think last lesson when we were

looking at these things it felt like there was an awful lot of things to learn

but actually if you just pause and think to yourself what would be good about this what would be bad about this most

of this is common sense it's just a few a few bits of language like sampling frames sampling units

if you can use that language it also makes it sound like you know what you're talking about as well

so just have like a sensible think about what would be useful um think about what could be good about

it and what could be bad about it as well okay