Inference Procedures
Type / Assumptions / Confidence Interval / Hypothesis Testt-procedure on means / · Have a random sample of ____
· Distribution is approximately normal because
- Large sample size (n >30)
- Graph data / / H0: m = ___
Match pairs t- procedure / · Have a random sample of ____
· Distribution is approximately normal because
- Large sample size (n >30)
- Graph differences / / H0: mdif = 0
2 sample t- procedure on means / · Have 2 independent random samples or 2 randomly assigned treatments
· Distributions are approximately normal because
- Both are large samples (n >30)
- Graph both sets of data / / H0: m1 = m2
z- procedure on proportions / · Have a random sample of ____
· Distribution is approximately normal because
- np 10 & n(1-p) 10 / / H0: p = ___
2 sample z- procedure on proportions / · Have 2 independent random samples or 2 randomly assigned treatments
· Distributions are approximately normal because
- n1p1 10 & n1(1-p1) 10 AND
- n2p2 10 & n2(1-p2) 10 / / H0: p1 = p2
Type / Assumptions / Confidence Interval / Hypothesis Test
t- procedure on slope / · Have an SRS of ____
· X & y are linear because the residual plot is randomly scattered
· sy is the same for all x-values because the points are evenly spread over the LSRL on the scatterplot
· Responses are normal because the boxplot of the residuals is approximately symmetrical / / H0: b = 0
Chi-square test– Goodness of fit / · Have an SRS of ____
· All expected counts 5.
(Be sure to list expected counts) / df = categories -1 / Ho: p1=__, p2=__, …, pk=__
Chi-square test– Independence / · Have an SRS of ____
· All expected counts 5.
(Be sure to list expected counts) / df = (r-1)(c-1) / Ho: ____ & ____ are independent
Chi-square test– Homogeneity / · Have an SRS of ____
· All expected counts 5.
(Be sure to list expected counts) / df = (r-1)(c-1) / Ho: the categorical proportions of the 1st pop. are the same as the categorical proportions of 2nd pop.