If the study population is large then random sampling is ideal. But if it is small, then you would to have the entire population in the study, in which case it would like a census. You can only have a sample size of one in using a qualitative method for impact analysis as shown in this paper: https://pdfs.semanticscholar.org/9f60/ffd76787dc881d2818dcdc123ada0f915207.pdf

Best,

Debra

Quasi-experimental design involves selecting groups, upon which a variable is tested, without any random pre-selection processes. For example, to perform an educational experiment, a class might be arbitrarily divided by alphabetical selection or by seating arrangement.

Hello Everyone:

I am resposting as my first post above had some typos that made my response faulty.

If the study population is large then random sampling is ideal if you're doing quantitative research. But if it is small, then you would want to have the entire population in the study, in which case it would be like a census survey. You can also have a sample size of one in using a qualitative method for impact analysis as shown in this paper: https://pdfs.semanticscholar.org/9f60/ffd76787dc881d2818dcdc123ada0f915207.pdf

Best,

Debra

The type of the sample is one of the major differences between the true experimental and the quasi-experimental study. whereas it should be totaly randomized in the former, it can be either intended or random in the latter.

I think the choice depends on the goals. If you want to compare the quality of training of teachers of the humanities and science, invite both to an interview . These should be teachers who have received the same training or the same courses - sample intentional. Be sure to ask questions on the subject of teaching, then he can build comparative diagrams. If you research general issues, then by accident - random. I work mainly on the first scenario. I conduct a lot of workshops, seminars, webinars and interview their participants with reviews, as well as how to increase the relevance of the content of my classes.

Hi, acording to the research gols, of course, It would be better the random option among individuais who have receveid the same treatment.

Do you mean in a quasi-experimental design? If so, the sample should be the natural group or intact groups.

There are two different purposes in sampling: (1) to generalize to a population and (2) to generalize to a theory (hypothesis testing or experimenting). For (1), a random sample of sufficient size is needed (see Sample Size Table on the Internet). For (2), an intended sample of certain characteristics is needed; sample size is determined by (a) chosen confidence level, say 95%; (b) chosen confidence interval, say 5%; (c) desired statistical power, say 0.8; and (d) expected effect size, if unknown 0.5.

in quasi experimental research, it is normally advisable to choose intact classes depending on your research objectives. dealing with human samples is different from material samples. there is a possibility of the experimental group sharing the intervention with the control group who are either in the same class or same school community. you should therefore purposively choose an intact class from one school and have your control in another to reduce diffusion effect.

Hello, please find attached a chapter we published recently about experimental research design. I hope you'll find it interesting for your research.

Random

If you mean to measure the effectiveness of the teacher training program, you should compare it with another non-experimental teacher training program. Before the experiment starts, you need to ensure that the baseline is equal, the participants are similar in their representations, and the programmes have the same intention, though they may vary considerably. Therefore you need to sample a group of teachers and, based on the results of their pre-test scores, divide the participants into two groups: control and experimental. You also need to examine both an experimental and non-experimental teacher's training programmes.

Teachers' interviewing could be regarded as an additional test since it could be informative only if teachers can compare at least two teacher training programmes.

Notably, in an experiment , one can adopt either a truly random sampling method or an intended , purposive type of sampling. However, in your case, since you are using a semi-experimental research method , an intended sampling method will suffice.