I used Cochran (1963) sample size formula, but as it shows it so old and my advisor asks for some new references for finding an appropriate sample size.
First of of all, once a statical formula has been well established, then it will continue to be used until something better is developed, so classic work maintains it value, even if it is older. In other words, your advisor is wrong, but there is probably nothing to be gained by trying to argue with him or her.
One possible resolution is to use an online sample calculator. The statistical formulas underling these calculators are a century old, but the fact that they are online may disguise this fact. At any rate, these will allow you to establish a sample size for a given level of "precision" in your measures.
A different approach probably does involve an improvement over the older approaches, which is to do a "power analysis" via the online system, G*Power. This will assess your ability to detect the sample size necessary to find a statistically significant result for a given type of analysis (e.g., a t-Test)
First of of all, once a statical formula has been well established, then it will continue to be used until something better is developed, so classic work maintains it value, even if it is older. In other words, your advisor is wrong, but there is probably nothing to be gained by trying to argue with him or her.
One possible resolution is to use an online sample calculator. The statistical formulas underling these calculators are a century old, but the fact that they are online may disguise this fact. At any rate, these will allow you to establish a sample size for a given level of "precision" in your measures.
A different approach probably does involve an improvement over the older approaches, which is to do a "power analysis" via the online system, G*Power. This will assess your ability to detect the sample size necessary to find a statistically significant result for a given type of analysis (e.g., a t-Test)
You can various sample size calculation formula for unknown population
Very well explained by respected David L Morgan , whether it is possible to find a minimum sample size requirement (based on power etc.) through means like Gpower for a given research model, it is not possible to find the magical number fitting optimal sample size for all. Furthermore, either it is online or offline sampling, optimal sample size is very subjective and depends on the type of target population, research objectives, and design of the study.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size