Hello,
I`d love to ask you for a question regarding appropriate sample size of interobserver reliability study, I`ve working on a study that evaluate a reliability of a new classification of flatfoot between three raters in 92 patients. The classification is composed of numbers (0,1,2) and letters (A,B,C,D,E), which in turn make a combination of numbers and letters, e.g. (1A1B1C1D, 1A2B1C, 2E......) Each rater determines a class of each patient`s foot e.g (rater 1: 1A1B1C, rater 2: 1A2B, rater 3: 1B1C).
I used Fleiss`s kappa for interobserver reliability between multiple raters using SPSS which yielded Fleiss Kappa=0.561, p
Hello Hee Young Lee Take a look at the links below
https://statistics.laerd.com/spss-tutorials/fleiss-kappa-in-spss-statistics.php
https://www.researchgate.net/post/How_to_calculate_sample_size_based_on_Fleiss_kappa
The sample size has nothing to do with your Fleiss Kappa value, which is the result that shows the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items.
Sample size is related to the population size and confidence interval and level. You can use an online calculator to work out the sample size and it must be at least 30. Sample sizes equal to or greater than 30 are often considered sufficient for the central limit theorem (CLT) to hold. The CLT states that the distribution of sample means approximates a normal distribution as the sample size gets larger, regardless of the population's distribution :)
Andrew Paul McKenzie Pegman because Fleiss' kappa has a standard error, I would assume that sample size is indeed relevant.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size