I would like to ask about sampling technique. My study aimed to collect data about physical activity among university students. To be representative to all students we decided to use multistage stratified sampling. So, we divide the faculties to practical & theoretical (1st stage) then we randomly select faculties (I.e, Law from theoretical , medicine from practical....and so on). From each faculty we select first and fourth academic years to represent early and late stages of university life (2nd stage). Students from each year will be selected randomly to fill the questionnaire. The number of students in each faculty will be proportionate to the total number (I.e, total sample size 850... theoretical faculties represent 56℅ of university students so the needed sample = 476).
My question , is that a correct way for sampling?? Is that a proportionate multistage stratified sampling or multistage stratified cluster sampling?
Thanks in advance,
How will students from each year be selected randomly?
This is not clear
I prefer the cluster sample (include all students in section/round). In this case you should multiply the calculated sample size by 1.5 or preferably 2, to compensate for the design effect. It is practically difficult to select simple random sample
What is the rationale behind selecting first & fourth years? It is better to include all educational years
Yes. I agree with Abdel when you are going for selection of students from sections amongst the stratified, it has to be cluster sampling preferable.
It is stratified cluster sampling. Stratified because you divide the faculties, and after it two stage cluster sampling (the first stage is for faculties in each division or stratum, the second stage is for students of a selected faculty).
Thanks for your response.... But I have a problem in some faculties where the students are not divided into small classes or sections (in theoretical faculties).
My sample size already *1.5 as primary sampling technique includes clusters.
The rationale for first and fourth year to compare between early stage after secondary school and late stages where the students are affected by university life.
You are dividing the faculties on theoretical and practical basis followed by selection of specific years only. What is the criteria to divide the faculty into theoretical and practical?
I would suggest you should first apply cluster technique to choose the faculties among all possible, then first year and fourth year may be selected in all the selected clusters, thereafter do a simple random sampling in each class. However if there is disproportionate number among the gender, please go for stratification of proportionate sample before applying simple random sampling after getting list of students from the authorities. Your technique is multistage cluster sampling. Please double the calculated sample size to account for sample size
I think it is not stratified sampling nor multistage sampling because the students of the courses first, third and fifth can not be represented nor selected in the sample with any nonnull probability.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size