MSQ171 BUSINESS DATA ANALYSIS REVISON LECTURE : REVISION OF HYPOTHESIS TESTING Fill in the blanks and come up with

h the appropriate conclusion using the critical value method. (Please note you dont have to document the formal six step process here. However, if asked to perform a hypothesis test (one sample) in the exam, you will be expected to do so.) Question 1 A group of retailers is considering spending $150 million on a new shopping complex in the Melbourne suburb of Blue Bay. For the complex to be profitable, however, the mean weekly expenditure on a wide range of consumer products, for households within 5km, must exceed $55 per week. Setting up the test: Variable type (Numerical or Categorical): One or two sample: Key word(s): exceed. This points to (H0 or Ha) Test type: sign Ha must be an Null and Alternative: H0: _________________ v Ha: _________________ Level of alpha to use: Sample results: A sample of 400 households yielded a sample mean of $55.45 and standard deviation of $12. What conclusion would you draw? __________5%________ _________________ _________________ _________________ _________________ test.

MSQ171 Revision Lecture

Page 1

Question 2 A manufacturer has a machine that produces 10% defectives and this has proven very costly in the past. He has to decide whether to spend $400,000 on a new machine whose distributor claims will produce far fewer than 10% defectives and could save him money in the long run. Setting up the test: Variable type (Numerical or Categorical): One or two sample: Key words: fewer. That is, (H0 or Ha) Test type: sign Ha must be a Null and Alternative: H0: _________________ v Ha: _________________ Level of alpha to use: Sample results: A batch of 500 items produced on the new machine contained 32 defectives. What conclusion would you draw? _________5%_________ _________________ _________________ _________________ _________________test.

MSQ171 Revision Lecture

Page 2

Question 3 In 2001 a detailed market research survey showed that 45% of consumers were aware of Brand X Headache pill. In 2006, the manufacturers of the product wish to see if the same awareness level still applies. Setting up the test: Variable type: One or two sample: Key words: That is, Test type: sign Ha must be a Null and Alternative: H0: _________________ v Ha: _________________ Level of alpha to use: Sample results: Of 2000 consumers interviewed in 2006, 992 were aware of the product. What conclusion would you draw? _________5%_________ Categorical variable. One sample. same awareness level still applies. _______________ in 2006. This points to _______. ___________test.

MSQ171 Revision Lecture

Page 3

Question 4 It has been claimed that a new teaching method results in the improvement in marks for applicants sitting an exam. We wish to test see whether the New Method is better than the Old Method. Note p-value for Levenes test is greater than alpha. Setting up the test: Variable type: One or two sample: Key words: Test type: sign Ha must be a Equal Variances? ____________variable. _________sample. ______________ and _______. This points to Ha. ___________test. _______

Had we only wanted to test for a difference. Then we would use: H0: 1 -2 = 0 v Ha: 1 -2 0

But as we wish to test whether New is better than Old, we would use: Null and Alternative: H0: _______________ v Ha: _______________ Level of alpha to use: Sample results: Using the Old Method, the mean score in the test for 23 candidates was 52 marks with a standard deviation of 4 marks. A sitting of 15 applicants taught by the New Method gave a mean score of 55 marks with a standard deviation of 5 marks. Assuming the two groups form independent random samples is there any reason to believe that there is an improvement from the New Method? _________5%_________

MSQ171 Revision Lecture

Page 4