If we survey patients for a descriptive study, how do we calculate for the sample size?
The sample size is generally dependent upon a number of factors such as type of statistics you are going to use for data analysis purpose, total number of items on all the scales, sampling frame etc. In general, it is recommended that your sample size should be at least 10 times of total number of items. If the participants are persons with a common health issue, the sample size should be more in order to obtain normality distribution. However, if the participants are persons with a rare disease, the sample size could be compromised with a justification for the same.
For sample size estimation calculating the effect size is the most important. The larger the effect size the smaller sample you need, and vice versa. However, in your case as you have only one patient group in a descriptive study, where you are not running compassion or correlation analyses, estimating the effect size might not be applicable. Some studies recommend sample by item role of thump for surveys, the recommendations ranged from 10 patients per survey item to 1 patient per survey item but the last is the lest recommended which will end with very small sample size. 10 patient per item is highly recommended but it depends on how rare is your target condition, where you can reduce the number if you face recruitment issues and justify in your paper.
Best wishes
The researcher should make sure the sample size is as large as possible and it most be representative of the population understudy.
John Paul Ben Silang, I think you need to think very carefully about what you are investigating, which will be reflected in your research questions. Rules of thumb such as 10 people for each item have been recommended for studies in which factor analyses are conducted, but I get the sense that you're not interested in factor analyses, so those rules of thumb might be inappropriate.
Furthermore (and I don't want to be disagreeable in terms of some of the above posts), "the larger the better" is not necessarily the best strategy when aiming for participants. If you can get the same results from a smaller sample, why inconvenience a lot more people and use up a lot more resources than you need to?
If I were you, I would look carefully at different sampling procedures (e.g., random sampling, stratified random sampling, etc.) in order to see what kind of sample might be best suited to your aims. A smaller number of people, appropriately sampled, might be much better than simply accessing a large number of people.
There are many ways to get the number you need. I use one simple way that requests knowing SD for the construct and negotiated level of marginal error with the founder. It is N95% = 1.96 x SD/H, with H representing one half of the marginal error (all values on the right side of equation need to be squared). SD of the construct you are about to measure is from your own pilot study or from some other relevant research. If you decide to go with more precision just include 2.58 instead 1,96 and your sample size will be larger for the difference between these two z-values. Hope this helps you.
simply question, you have to see my papers and my articles, Definitively you will find the exactly answers.
please follow me
Please define "descriptive survey." Are you just trying to measure something about your sample and not trying to infer anything, then, I guess 1. For example, if I want to know what a patient had for lunch, I survey them, find out. If you mean something else, or if you really plan to infer something from the data you collect, please say.
There are different formulas for calculating samples sizes. You may see these formulas with examples in following reference.
Business Statisitics by examples fifth edition by Terry Sincich.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size