Hello Rahaman

1.Research studies are usually carried out on sample of subjects rather than whole populations. The most challenging aspect of fieldwork is drawing a random sample from the target population to which the results of the study would be generalized.

2.The key to a good sample is that it has to be typical of the population from which it is drawn. When the information from a sample is not typical of that in the population in a systematic way, we say that error has occurred. In actual practice, the task is so difficult that several types of errors, i.e. sampling error, non-sampling error, Response error, Processing error,… In addition, the most important error is the Sampling error, which is statistically defined as the error caused by observing a sample instead of the whole population. The underlying principle that must be followed if we are to have any hope of making inferences from a sample to a population is that the sample be representative of that population.

3.A key way of achieving this is through the use of “randomization”. There several types of random samples, Some of which are: Simple Random Sampling, Stratified Random Sampling, Double-stage Random Sampling... Moreover, the most important sample is the simple random sample which is a sample selected in such a way that every possible sample of the same size is equally likely to be chosen. In order to reduce the sampling error, the simple random sample technique and a large sample size have to be developed.

4. The above sampling methods are well-known and accurate, but choosing the best sampling method should be based on your problem, which should be related to a specific and different population.

Regards,

Zuhair

Again, please refer to the references :

1. William G. Cochran (1977). Sampling techniques.Wiley.

2. William G. Cochran, Gertrude M. Cox (1977). Experimental Design, Wiley.

3. Snedecor, W.G. and Cochran,W.G. (1989). Statistical Methods, Blackwell.

4. Geoffrey Keppel and Thomas D. Wickens (2004). Design and Analysis: A Researcher's Handbook (4th Edition). International Edition.

Simple answer: Random sampling is the best approach.

In a perfect world this is always the best answer. In the real world it is not always possible, mainly because we keep trying to reduce workloads and manage time constraints by using smaller sample sizes.

1) By random chance I could get all males in my sample. Bad luck on my part, but it was a random sample and is therefore representative? So if I do random sampling, but exclude some of the possible random outcomes is the sampling still random? There were 2.7*10^27 possible random samples and I excluded 1.8*10^4. Choose whatever values you like in the previous sentence, but where is this no longer random?

2) Some possible outcomes are rare. I have males-females, income levels, race, and education. In my random sample of 2300 surveys I find that I get 2 respondents that are female, Pacific islander with a Ph.D, making 20K/year. It becomes difficult handling and interpreting this data. Of course I also have 872 respondents that were Asian, male, B.S. making 60 K/year.

3) I ask 500 people to take my survey. They all agree. I assign all of them a unique number and then ask for them to wait while I go to a random number generator and select the 150 people that I need for my research. I return to the room to find that I have 2 people that stayed to take the survey.

Other approaches to random sampling are used to develop a sampling program that works. The real goal is to understand what population was sampled and to get a reasonably unbiased sample from that population. If I sample people at a mall on Wednesdays between 9am and 3pm, then my population is limited to people available to go shopping at this mall between these hours on this day. If that limitation is not useful, then I need to adjust my sampling protocol to get more of the population that I am interested in. In the end it is a balance between sampling the population that you would like and the available time and resources. Sample multiple malls randomly selected each week for 2 months, and randomly select a time interval of 3 hours for each sampling episode might be better but it will also require greater resources.

In the end, do your reviewers accept that you sampled the population you claim to have sampled and have done so in a reasonably unbiased method?

Prof. Al-Hemyari gave first rate references esp. the Cochran book. Let me add one more, Schaefer, Elementary Survey Sampling:

https://www.amazon.com/gp/product/0840053614/ref=dbs_a_def_rwt_bibl_vppi_i0

Both Cochran and Scheafer emphasize two points:

1. Probability samples are essential to doing meaningful statistical inference and

2. The only reason to depart from a simple random sample is to try to get smaller standard errors and hence shorter confidence intervals and larger t statistics. The rest of a sampling course is some mathematics on how to try to achieve 2.

The best method is the one that achieves 2, if possible, and a simple random sample otherwise. Unfortunately sampling courses are disappearing. Keep your sampling books. It took a lot of effort to figure these things out. They deserve to be remembered. All statistical methods used should come from the question being asked. Best wishes, David Booth

Thak you very [email protected]@Philip D Dewitt@Timothy A Ebert@David Eugene Booth.

Any sampling method that is simple, and easy to use is the best method. Hence, I agree that random sampling is the best method as long as it helps achieve objective of the research study under consideration.

It will varies research to research.