What type of sampling is this....`

98107

New Member
#1
First of all sorry if this is in the wrong section but i think is the section that applies most!

I want to distrubute questionaires to five groups of people with different educational back grounds;
• Group 1; No GCSE’s
• Group 2; Some GCSE’s
• Group 3; Higher education (A-levels, college course, NVQ etc)
• Group 4; University Graduates
• Group 5; Academic Professional

To get people to complete the questionaire from group 1-3 i will send them to local schools in my region and ask teachers to make the appropaite kids fill them in.

Is this 'non-probabilisitic sampling?

As i know who they are going to be filled in by (as i am aiming it at people with particular levels of education levels, BUT; i dont know who exactly will be filling it in. as in im not choosing thr particular kid.

So its not random as there all people i want to ask, but i dont know who will be asked in advance( im not picking and choosing the induviduals of that particular level of education_)

hope this makes sense.

Cheers any help appreciated!

Alex
 

CowboyBear

Super Moderator
#2
Is this 'non-probabilisitic sampling?
I would say so, yes. This looks like convenience sampling to me. The common factor amongst different types of probability sampling is random sampling (e.g. simple random sampling from a specific population, random sampling from within specific sub-populations in stratified sampling, etc.) There is no random sampling here - you've picked out schools presumably on a non-random basis, and even though you don't know exactly who will be surveyed, teachers will still hand out surveys on a non-random basis.
 

98107

New Member
#3
thanks mate thats a great post;

So obviously convenience sampling has its drawbacks (as in not random) but for my case would you say its the only method? as the schools ive picked are just schools which are close to me and they are the source which has the highest abundance of the required educational level.
as the alternative would be to stand in the street and ask random people untill i have enough people from each group (as this clearly is not possible)

From each group say if i get a 100 or so, does that then enable me to say its 'representitive' or due to it being non random slection i can not say it is representitive?

thanks once again
 

CowboyBear

Super Moderator
#4
So obviously convenience sampling has its drawbacks (as in not random) but for my case would you say its the only method? as the schools ive picked are just schools which are close to me and they are the source which has the highest abundance of the required educational level.
as the alternative would be to stand in the street and ask random people untill i have enough people from each group (as this clearly is not possible)
Hi again!

Asking people in the street wouldn't really be random sampling either ;) For (simple) random sampling, you need to have a clearly defined population within which every individual has an equal chance of being selected, usually via use of random number tables etc.

Is convenience sampling the only method you could use? Realistically a genuine probability sample is very difficult to accomplish. What you could maybe do is to take steps to make your sample as representative as possible within the practical constraints of the project - e.g. you could make sure you have a clearly specified population of interest, and then select a handful of schools that are representative of the population on key variables of interest. E.g. if half of the schools in the population are urban and half rural, you would want to have a similar mix in the selected sample of schools. Your advisor might have other suggestions about what changes might be feasible or necessary.

From each group say if i get a 100 or so, does that then enable me to say its 'representitive' or due to it being non random slection i can not say it is representitive?
Unfortunately you can't really make a strong claim that it's representative without random sampling. You would instead really need to put together a good argument to justify assuming it's representative based on things like the sample matching the population on key demographic variables, study procedures being designed to avoid teachers selecting students with biased characteristics on study variables, etc. (And would still need to list non-probability sampling as a limitation in your discussion section! :))
 

98107

New Member
#5
thanks mate again, could i ask you one more question.

Basically for my study i want to analyse how people of different educational back grounds react and interpret different methods of presenting them with information.

For the groups as i stated before. I was just going to use schools as they have a high abundance of people with each requirement for the groups, however after thinking a bit more i dont want to really use schools as it will become more of a study about age and not really educational background.

How do you think the best way to sample of the desired groups would be? considering i need a good amount of people to do the questionnaires to get good data.

For groups 4+5 i will probabaly have to use my university.

But how do you think the best way to get data from the other groups will be?

Im thinking of getting rid of group 1 altogether, as not many people with no GCSE's or equilievent.

Thanks mate, owe you one.
 

CowboyBear

Super Moderator
#6
Hi again! To be honest I'm not sure if I can give you much of a useful response here - the best way to sample will depend on practical issues in your area, constraints of your project, etc. Your advisor or supervisor should be able to help though :)