The Student Room Group
Uni students - Complete our survey >>
Uni students - Complete our survey >>
Uni students - Complete our survey >>

Do all systematic samples have to be random?

So if we had data for rainfall on each day of the year and we had to take 12 samples, but we only sampled the first day of each month..would this be random?
Original post by dont know it
So if we had data for rainfall on each day of the year and we had to take 12 samples, but we only sampled the first day of each month..would this be random?


What do you think?
Original post by Muttley79
What do you think?

I would say no because you've picked the same day each month rather than various days of the month.
Original post by dont know it
I would say no because you've picked the same day each month rather than various days of the month.


I agree - should use a random number generator.
Original post by Muttley79
I agree - should use a random number generator.


In fact, would it even be a systematic sample? Maybe a better example is if we had data from 6 months and we had to take a sample of 6 and we ended up sampling the first day of each month, there wouldn't a fixed interval that allows for this since some months will have 30 days and some 31?
Original post by dont know it
In fact, would it even be a systematic sample? Maybe a better example is if we had data from 6 months and we had to take a sample of 6 and we ended up sampling the first day of each month, there wouldn't a fixed interval that allows for this since some months will have 30 days and some 31?


Do you mean stratified? Yes we'd need to factor in the different number of days in each month so for Jan (31/365) x whole sample size
Original post by Muttley79
Do you mean stratified? Yes we'd need to factor in the different number of days in each month so for Jan (31/365) x whole sample size


Nah I actually did mean systematic :smile:. I get why it might not be stratified but in the case of systematic sampling, if we chose between a number from 1-30 and we chose the number 1, then picking every 30th number thereafter wouldn't guarantee that we sample the first day of each month. So we might end up sampling the 31st day of that month. So essentially I'm saying that after selecting your starting point, there would be no fixed interval thereafter that means you only sample the first day of each month, and therefore what I used in my example wouldn't count as a systematic sample I don't think?
Original post by dont know it
Nah I actually did mean systematic :smile:. I get why it might not be stratified but in the case of systematic sampling, if we chose between a number from 1-30 and we chose the number 1, then picking every 30th number thereafter wouldn't guarantee that we sample the first day of each month. So we might end up sampling the 31st day of that month. So essentially I'm saying that after selecting your starting point, there would be no fixed interval thereafter that means you only sample the first day of each month, and therefore what I used in my example wouldn't count as a systematic sample I don't think?


A sample size of 12 is not a good idea anyway [too small] - what is the actual question you are trying to answer?
Original post by Muttley79
A sample size of 12 is not a good idea anyway [too small] - what is the actual question you are trying to answer?


Data on the daily total rainfall in Beijin during 2015 is gathered from the large data set. The daily total rainfall(in mm) on the first day of each month is listed below.

May 1st: 9.0
June 1st: 0.0
July 1st:1.0
August 1st: 32.0
September 1st: 4.1
October 1st: 3.0

a) State with a reason whether or not this sample is random.
b)Suggest 2 alternative sampling methods and give 1 advantage and disadvantage of each.
Original post by dont know it
Data on the daily total rainfall in Beijin during 2015 is gathered from the large data set. The daily total rainfall(in mm) on the first day of each month is listed below.

May 1st: 9.0
June 1st: 0.0
July 1st:1.0
August 1st: 32.0
September 1st: 4.1
October 1st: 3.0

a) State with a reason whether or not this sample is random.
b)Suggest 2 alternative sampling methods and give 1 advantage and disadvantage of each.


OK - so not random. For (b) what are you going to put?
Original post by Muttley79
OK - so not random. For (b) what are you going to put?


I was going to put simple random and stratified.
Advantage of simple random sampling: More likely to be free of bias.
Disadvantage of simple random: Clustered results meaning we won't get a representative sample.

Advantage of stratified: We get a proportionate sample from each month.
Disadvantage: Is a bit expensive to separate the months into different strata.
Original post by dont know it
I was going to put simple random and stratified.
Advantage of simple random sampling: More likely to be free of bias.
Disadvantage of simple random: Clustered results meaning we won't get a representative sample.

Advantage of stratified: We get a proportionate sample from each month.
Disadvantage: Is a bit expensive to separate the months into different strata.

Random is not more likely to be free of bias and you don;t necessarily get clustered results!
Original post by Muttley79
Random is not more likely to be free of bias and you don;t necessarily get clustered results!


I'm quite confused tbh, why not? Wouldn't a valid disadvantage be that you could potentially get clustered results e.g. we could sample more data points from June where there'd be less rainfall in comparison to October for example and therefore the picture we get wouldn't be totally correct. Also, wouldn't it be free of bias in the sense we would be sampling different days of the month as opposed to the 1st day of each month which could've been selected because there was a pattern for example.
Original post by dont know it
I'm quite confused tbh, why not? Wouldn't a valid disadvantage be that you could potentially get clustered results e.g. we could sample more data points from June where there'd be less rainfall in comparison to October for example and therefore the picture we get wouldn't be totally correct. Also, wouldn't it be free of bias in the sense we would be sampling different days of the month as opposed to the 1st day of each month which could've been selected because there was a pattern for example.

Ah, your wording here is different and I agree that a disadvantage is that the sample could be biased with too many from one month. Can you see a difference in what I've written? For an advantage - random sampling is all/any days of the month could be chosen.
Original post by Muttley79
Ah, your wording here is different and I agree that a disadvantage is that the sample could be biased with too many from one month. Can you see a difference in what I've written? For an advantage - random sampling is all/any days of the month could be chosen.


Thanks :smile:. Yeah I should've specified what I meant when I said it first time.

Quick Reply

Related discussions

Latest

Trending

Trending

Articles for you