Do I need to conduct a pilot study for research where the instrument is taken from other researchers and the sample size is 250 employees out of 550 employees?
I, personally, do not do pilot studies. Though many say it has benefits. I beg to differ. If you know your research subject matter well, then i dont feel its neccesary. Pilot studies can also consume time, energy and resources. However, thats just my opinion.
You use pilot research to address questions you have about how you will conduct your main study, particularly about your materials. Do you have such questions? If the pilot research would have caught a problem that you could fix, and you did not run it, then your n=250 study essentially becomes your pilot. So, you need to consider the risks. If, for example, this were for a dissertation and you could not complete a second study after the error, then you would have to consider how important it is for you to pass versus fail the dissertation.
Hello Ahmad
If the instruments are standardized and culture-free for the intended population of your study (I am assuming cross sectional, quantitative data), then you may not require pilot study. However, if you are doubtful about the aspects such as wordings, relevance, or meaning of the items of the scales, then doing a pilot study would serve the purpose. Having said that, I guess a sample size of N = 100 or something would be sufficient for pilot study.
Depends on circumstances of intended research; principle says "yes"; pragmatism says "maybe".
Ahmad -
One reason for conducting a pilot study for quantitative data is if you do not have a good guess for population or strata standard deviations for estimating sample size needs for a probability-of-selection-based (randomized/design-based) sample to follow, or for an idea of sample size needs for model-assisted design-based, or a prediction/model-based approach. So whatever your methodology, a pilot study can help you decide on the sample size needed in the effort for which you are preparing. However here it seems you have already picked a rather large sample size compared to the population size. Are you sure that you need that?
Remember to account for the finite population correction factor for the design-based case, or it's equivalent in a model-based approach. See https://www.researchgate.net/publication/262970761_FINITE_POPULATION_CORRECTION_fpc_FACTOR.
Sample size needs may be determined separately for each important data item, and then you could use the largest estimated size which results.
Cheers - Jim
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size