difference between stratified sampling and random stratified sampling?

[vella97]

What exactly is the difference between Stratified sampling and random stratified sampling? Would you be able to provide me with examples. I don't quite understand how the textbook explains it..

[avcestudent]

Well, my understanding is that stratified sampling occurs when the population is divided into distinct groups, or “strata” and then samples of proportionate size are drawn from within each of the strata. For example, a stratified sample of 100 students from a 9-12 school might include 25 Year 9s, 30 Year 10s (more Year 10 students at the school than Year 11), 20 Year 11s and 25 Year 12s but participants are chosen to participate through a non-random sampling procedure i.e. convenience sampling (I might get my friends to participate).

Random stratified sampling involves the same process EXCEPT the people chosen to participate from each strata is done so randomly. So, I might split the population into year levels (strata) and if there are 100 year 9 students at the school I might use a computer to randomly pick 25 names and ask them to participate, then I repeat the process for the other year levels.

So whilst stratified sampling is more representative of the population than other sampling procedures (convenience etc.) it is still a biased sample as every member of the population does not have an equal chance of being selected to participate, so this is why random stratified sampling is used.

Hope this helps. 

[scout]

So whilst stratified sampling is more representative of the population than other sampling procedures (convenience etc.) it is still a biased sample as every member of the population does not have an equal chance of being selected to participate, so this is why random stratified sampling is used.

As you said, stratified sampling is a biased. That means that it isn't completely representative of the population, right? Then I'm confused as to why I've read that stratified sampling IS 'representative of the population'.

[avcestudent]

As you said, stratified sampling is a biased. That means that it isn't completely representative of the population, right?

Yeah, that's correct. I guess the best way to determine if any sampling procedure is representative of the population is to ask yourself if there is any bias. Even if the population is broken up into strata, if participants aren't chosen randomly, the sample cannot truly be considered to be representative. Stratified sampling is definitely more representative of the population compared to convenience sampling (and therefore a preferable option), however, it is not completely representative of the population as participants aren't chosen randomly.

Then I'm confused as to why I've read that stratified sampling IS 'representative of the population'.

I think this is to do with the wording. If you consider the literal meaning, stratified sampling is 'representative of the population' as it creates a sample where each group in a population is represented in a proportionate size. However, using psychological principles, 'representative of the population' means an unbiased sample, therefore stratified sample is not representative of the population.

[scout]

However, using psychological principles, 'representative of the population' means an unbiased sample, therefore stratified sample is not representative of the population.

Awesome timing (I just happened to be online)!

Thanks for the clarification avcestudent

I completely agree with what you said. What's really strange is the fact that I read 'stratified sampling is representative of the population' in the VCE Psychology Research Methods Workbook. Now I'm beginning to have doubts about published vce books... :/

[avcestudent]

Awesome timing (I just happened to be online)!

Thanks for the clarification avcestudent

I completely agree with what you said. What's really strange is the fact that I read 'stratified sampling is representative of the population' in the VCE Psychology Research Methods Workbook. Now I'm beginning to have doubts about published vce books... :/

No worries, glad I could help.  And yeah, unfortunately VCE books aren't always accurate

[Joseph41]

Awesome timing (I just happened to be online)!

Thanks for the clarification avcestudent

I completely agree with what you said. What's really strange is the fact that I read 'stratified sampling is representative of the population' in the VCE Psychology Research Methods Workbook. Now I'm beginning to have doubts about published vce books... :/

Stratified sampling is representative of the population insofar as it accounts for one variable accurately. But it is not the most representative sampling procedure because it doesn't account for any other variable.

So you could use stratified sampling to ensure, for example, a certain proportion of participants with red hair, but unless you use random stratified sampling, not each member of the population with red hair would have an equal chance of being selected as part of the sample. So that means that all of the participants in the sample with red hair might be from the same location, or might all be female, or biased in terms of another variable.

Does that make sense?

[scout]

Stratified sampling is representative of the population insofar as it accounts for one variable accurately. But it is not the most representative sampling procedure because it doesn't account for any other variable.

So you could use stratified sampling to ensure, for example, a certain proportion of participants with red hair, but unless you use random stratified sampling, not each member of the population with red hair would have an equal chance of being selected as part of the sample. So that means that all of the participants in the sample with red hair might be from the same location, or might all be female, or biased in terms of another variable.

Does that make sense?

So the one variable that stratified sampling does control is the proportion of specific characteristics in the sample (which should be equal to that in the target population).

I suppose what I'm curious about is whether or not it is right/safe at all to say that stratified sampling is representative of the population, when it is only to some extent (as you said).

[Joseph41]

So the one variable that stratified sampling does control is the proportion of specific characteristics in the sample (which should be equal to that in the target population).

That is my understanding, yes.

I suppose what I'm curious about is whether or not it is right/safe at all to say that stratified sampling is representative of the population, when it is only to some extent (as you said).

It might depend on context of the question. I think that it would be fairly safe to label stratified sampling as representative, but if you're concerned about it, there is no harm in being more specific. You could always write something as you have above, such as "ergo, stratified sampling is representative of the population insofar as [Quality X] is consistent in both the sample and the population."

[scout]

It might depend on context of the question. I think that it would be fairly safe to label stratified sampling as representative, but if you're concerned about it, there is no harm in being more specific. You could always write something as you have above, such as "ergo, stratified sampling is representative of the population insofar as [Quality X] is consistent in both the sample and the population."

Thank you Maverick Mac!