BIO361H5 Lecture Notes - Lecture 6: Central Limit Theorem, Statistical Power, Sample Size Determination
Document Summary
What are the sample sizes: original study randomization. Population 10, the mean is a fixed value (parameter), the last value has to be such that you end with the parameter/population mean. So in this example you would have 9 values to choose from. Statistical power depends on sample size (small sample size may not detect difference even if its there = type 2 error) Sample size and difference between treatments affects the statistical power. Don"t want sample too large/approaching size of population bc then its no longer an estimation. Sample size makes you think about experimental unit: ex. sensory perception with hands tested both hands which aren"t independent of each other, should have used paired t-test. 10 mice, randomization test to run 1000 tests (do we increase sample size depending on the number of tests run?) no matter how many times you do the randomization, the sample size does not change. Graph histogram for 1 variable, boxplot for multiple.
