Statistical Sciences 1024A/B Chapter Notes - Chapter 21: Null Hypothesis, Standard Deviation, Statistical Inference
Document Summary
Can arise from a randomized comparative experiments that divides subjects into two groups/treatments. Compare to populations (survey: two separate srs, one quantitative measurement on each individual. Compare treatments (experiment: randomly divide individuals into two treatment groups, one quantitative measurement on each individual. Distributions of x 1 and x 2: x 1 - x 2. Since we do not know (cid:2869) or (cid:2870) we use s1 or s2: df = smaller of (n1 1) or (n2 1) Robust confidence level or p-value does not change very much when conditions of procedure are not met. Two sample t procedures are more robust than the one sample t methods. When sample sizes are equal the populations have similar shapes. P-values and confidence intervals are quite accurate: even when sample size is as small as n1 = n2 = 5. Two sample t procedures are more robust when the sample sizes are equal. Note: (n1 + n2) of 40 or more.