minitab

Class 4 Assignment - Performing t tests with minitab

Use minitab to answer questions 1, 2, and 3 below.

For all questions:

A. Calculate the summary statistics for the dataset, using the minitab function.

B. Select the appropriate statistical hypothesis test (hint: they are all t tests, check the tables on slides 17 and 18 to help you pick).

AFTER you decided which test to use, go to the minitab Assistant and feed it your info (kinds of data, question you have) and see whether it comes up with the same test than you did. If yes, run the test. If no, run both tests and compare.

C. State your answer in statistical terms:

What was your null hypothesis? Your alternative hypothesis? What is the significance of your result? What did you set the power of your test to? Do you reject or not reject the null hypothesis?

Note: Minitab asks for or sets a default confidence interval, instead of the power of the test. These areinversely related: power of 0.05 means 5% chance or less that a test statistic (value) is part of the distribution for the null hypothesis, so there is a 95% chance that the test statistic is not part of the null hypothesis distribution. Or think about it this way: if you are 95% confident that a test statistics is within the range for the null hypothesis you don’t reject the null hypothesis. If the test statistic falls outside that range (that means they are in the 5% range NOT covered by the confidence interval) you reject the null hypothesis.

Question 1:

Is the protein content of a new, cheaper formula for cat food significantly different (smaller or larger) than the mean of the old formula?

Experiment: conduct protein assays on different batches of the new and old formula.

Dataset: Cat food protein data for class 4 HW.xls

Type of question: is HA > or < than H0?

Question 2:

Is the IQ of people living in a certain area (defined by ZIP code) higher than the (known) average IQ of the population?

Experiment: test IQ of a sample of people in the particular ZIP code area.

Dataset: IQs of 16 individuals living in a certain ZIP code area.xls

Type of question: is Ha > H0?

Question 3:

Does running on a regular basis affect your resting heart rate?

Experiment: enroll a sample of random people in the study. Determine resting heart rate. Let them carry out the running protocol for a year. Test resting heart rate again.

Dataset: Resting heart rate data.xls

Type of question: is HA > or < than H0?