Dear Colleagues,
currently, I am making a research proposal on studying the survey of LGBTIQ in Sri Lanka.I want to make a survey what is the population of LGBTIQ in Sri Lanka. In order to study this ,how can I make my sample size and I know snow ball is effective in this kind of survey.But I want to take the most scientific form of sample in this survey
The first questions that you need to address are:
1) How many people will refuse to answer one or more questions? They may be offended, or may not be ready to accept the answer let alone put it on paper.
2) How many people will give false responses?
Especially if I am being forced to participate, or you are simply making me angry, I will take your survey just to get rid of you. However, I will give creative answers designed to cause problems in your statistical results, or at least that is what I am trying for.
While the number of surveys is important, their distribution is also important. So sampling in a shopping mall might give one answer, while sampling at an airport, or at nightclubs might give an entirely different answer.
You will not get a flawless answer. Absolute truth is not an attainable goal. What you need to do is to think very carefully about the consequences of your choices when getting the best answer that you can. You could give everyone in Sri Lanka a number, and then use a computer to generate random numbers to identify people who will be surveyed. This is the only way to get a random sample. Failing that you will get a more-or-less biased sample depending on your compromise. If you only sample (or deliberately avoid) all LGBTIQ nightclubs you will get a biased result. If you include all of Sri Lanka, or only one city/town/village then that too will bias your answer. A small sample of a few hundred will insure that your sample is biased in some way. You can solve some of the problem by limiting your population. Instead of "everyone in Sri Lanka" consider a more modest population like everyone in Negombo, or even more modest like everyone attending the university in 2017, or just the freshman class. It won't tell you much about Sri Lanka, but it would be a good start in learning survey methods.
Of course the problems will be greater or less depending on the social status of LGBTIQ peoples. If there is no social stigma associated with these issues, then your survey will be much easier. That is not the case, in the USA. Just because there are problems does not mean that you should not do the survey. Just think about how your method influences your result and modify your claims of enormous scientific breakthrough accordingly.
Dear Dr. Timothy,
Thank you very much for your lengthy devoted answer for my question.
I see some insightful solution behind your explanation on the issue of sample for a survey to find the LGBTIQ population in Sri Lanka
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size