Answers: Inference
Answers to odd number problems
1. Ho in population
True False
--------------------
|
reject Ho FP | TP
|
--------------------
|
accept Ho TN |. FN
|
so we have:
* FP goes with Type I error
* FN goes with Type II error
3. test statistic is value of test (e.g., t-test, chi-square) calculated from results. Critical value is value from the appropriate distribution associated with a known probability.
5. lots of possible answers
7. Two tailed. No reason to specify a particular hypothesis, like, more red than green. Null hypothesis: candy colors in equal proportions; alternate hypothesis: candy colors not in equal proportions
9. Implies a two tailed test: Null hypothesis: Effectiveness of old and new drugs are the same; Alternate hypothesis: Effectiveness of old and new drugs are not the same.
Regarding side-effects, secondary hypotheses one tailed: Null hypothesis: new drug has the same or more side effects; Alternate hypothesis: new drug has fewer side effects
11. Calculate the standard error of the mean: 0.4110961, 0.2373464, 0.8221922, 0.4746929
approx lower, approx upper limits
12.17781, 13.82219
12.52531, 13.47469
11.35562, 14.64438
12.05061, 13.94939
11b. width of confidence interval proportional to size of variability; greater the variability the wider the interval, and therefore the less certain we are about our estimate.
11c.Width of confidence interval inversely proportional to sample size; greater the sample size the narrower the interval, and therefore the more certain we are about our estimate.
For 11, I used R as a calculator. Here’s some of my code
mySEM <- c(1.3/sqrt(10), 1.3/sqrt(30), 2.6/sqrt(10), 2.6/sqrt(30)) upper <- 13+2*mySEM lower <- 13-2*mySEM myDiff <- upper - lower sort(mySEM) sort(myDiff)