Sampling inquiry for thesis

Question

Hi!

I have a mixed-method thesis ongoing and I plan collecting data on my own college (namely college X), specifically from students and faculty members on my department. Evidently, that would be stratified sampling. Specifically, it's the "proportionate stratified sampling". The study's aim is to determine the knowledge, attitude, and perceptions (KAP) of students and faculty members regarding a certain event and perform correlation of KAP between students and faculty members in college X in my department.

Upon calculating the sample size using the proportional allocation method, (for example) I'll be getting 200 students and 100 faculty members for the survey. I don't have any issue with allocating students in my department but I'll be dealing with the limited amount of faculty members. In our department, there are only 30 faculty members. My thesis adviser suggested that we could allocate other faculty members from other schools with the same department as I am, namely college A and B.



So in summary:

1. Allocate students in college X only (since that's the original objective of the study)
2. Allocate all 30 faculty members in college X, then the remaining 70 would be from college A and B.

My question:

1. Based on my study's objective, would there be a huge error or bias in my study since the objective only state college X? The plan is to solely focus on my college, then get the remaining faculty members from other colleges due to limited number of faculty members. Should I not mention any college at all?
2. If you will be suggesting to state college X, A, and B in the objective, that would be a huge error since the students will only be coming from college X and that won't be changed.
3. If you will be suggesting to use the disproportionate stratified sampling, there could be a huge error when doing the data analysis. The responses from students could overpower the responses of faculty members.
4. If you're thinking to make the students 30 sample size and 30 faculty members, the data analysis would not represent the whole department since there are a thousand of students in my department.

With this, I would like to extend my greatest appreciation to you all. I hope you could answer my question. Thank you!

Answer 1

Hi Jared,

I can think of three options for you, all of which would involve broadening your objective slightly. First, you could go with the option your advisor suggested--take faculty members from colleges X, A, and B, but that, as you pointed out, creates a problem if you want to focus solely on students from college X.

The second option is to add faculty from additional departments at College X instead of additional schools. I don't know what your department is, but let's say, for example, you are in statistics. There may only be 30 faculty available in the statistics department, but you could get additional faculty from similar departments such as math or economics. You would just need to change your objective slightly to say "in departments X, Y, and Z" or "in quantitative disciplines," as in the case above.

The third option is the one you suggested last: make the sample sizes 30 students and 30 faculty members. You are correct in that this is relatively small n--typically we want n to be at least 50; some textbooks say 75, but every study has limitations and, of those I've read, every study includes a specific limitations section, where you could mention your small sample size and explain the reason it is so small.

In my opinion, option 2 is the best in a vacuum. If it is practical to use faculty from other departments that your students have had classes with, I think it has the least impact on your analysis. However, I realize that it may not be practical at some colleges, in which case you would need to go with options 1 or 3.

I hope this helps.