1- and 2-sample t-tests Flashcards
The smaller the sample size is…
- the more variable the sample SD is as an estimate of σ
- hence the distribution of the test statistic will be “wider”
The t-distribution with v degrees of freedom
- v = ∞ results in the standard normal distribution
- The standardised chance error (if standardised with the sample SD) follows a t-distribution with v = n−1 = sample size − 1
The “best” choice of weighted average is the one that…
Gives the resultant estimate the smallest possible SE.
Assumptions for 2-sample t-test
- All observed individuals are independent
- The two samples contain different people
- The two populations are normal and have equal spread (SD/var)
var.equal = for t.test
T - indicates we are doing 2-sample
F - would run a Welch t-test
We can relax the assumption that 2 pop. have to have equal spread by using a…
Welch 2-sample t-test
One-sample t-test
Looks at diff between sample mean and a fixed value
Two-sample t-test
Looks at the diff between sample means from 2 diff group
* Take the difference and see if it is 0
z-test
Diff between a population proportion and fixed proportion value
95% CI’s for population proportion
CI= ov ± se x qnorm(0.975)