Uncertainty in Sample Estimates Flashcards
confidence intervals, bonferroni correction
What does the uncertainty about the sample mean depend on?
- the variability in the wider population
- sample size used to calculate the sample mean
- distributional shape of the wider population (doesn’t matter in larger samples)
When does the variability in the sample decrease?
When sample size increases (more data = better idea of the true mean).
What type of distributions describe the sample means from different samples?
Normal dist, t-dist.
What is a standard error?
The standard deviation of a sampling distribution, a measure of how likely the sample mean is to be to the population mean.
large se = low precision
Where is the true mean usually found?
Within 2 se of the estimate (95% of the time).
What is a confidence interval?
A range of values for the true mean within which we are confident the true parameter value lies.
When are empirical (bootstap-based) confidence intervals used?
Small sample size, wider population is very skewed, cannot safely assume normal/t-distribution.
How does the Bootstrap approach work?
i) select n values at random (with replacement) from our original sample of n values
ii) for each sample, calculate the mean
iii) examine the distribution of these sample estimates and locate the values that define the central 95% of the values (2.5 and 97.5 percentiles)
When finding the estimate of the difference of two means, what decides what approach we must use?
Depends on whether we can assume the variability of the two groups is the same or not.
How do we decide if the variances of two groups are equal or not?
Use the F-test.
What is the Bonferroni correction?
To maintain an overall type I error rate (α) of 5% we divide α by the number of comparrisons k.
