I am analyzing a group of 102 healthy participants, stratified by two levels of education (n1 = 50, n2 = 52), for the development of normative data for a reading and writing assessment battery. The number of items per task varies between 15 and 75 items. Some tasks are evaluated according to the execution time (ranges from 62 to 194 seconds) or the number of words (ranges from 42 to 165 words). The cutoff score will be the 16th percentile for each measure. I have averages, standard deviations, percentile distribution. I want to know if the sample size is sufficient to have normative data representative of the population. I would also like to know if adding 15 participants per level of education would improve my results. Thanks for your help!
Valérie
You can use the free program G*Power to assess your ability to detect a given effect size with this sample.
Thank you for your suggestion David L Morgan ! As I'm not sure how to use G*Power for percentile distribution (no comparison group and no known reference score), I am planning to add information about the interval estimate of the percentile ranks. If you have any advice, you can let me know.
I found those two interesting papers on the subject.
Ana J. Bridges & Karen A. Holler (2007) How Many is Enough? Determining Optimal Sample Sizes for Normative Studies in Pediatric Neuropsychology, Child Neuropsychology, 13:6, 528-538, DOI: 10.1080/09297040701233875
Crawford, J. R., & Garthwaite, P. H. (2009). Percentiles please: The case for expressing neuropsychological test scores and accompanying confidence limits as percentile ranks. The Clinical Neuropsychologist, 23(2), 193-204. DOI: 10.1080/13854040801968450
Thanks for your help!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size