General instructions

▪ The maximum word limit for the present assignment (excluding tables, contents page,  footnotes,  charts,  graphs,  figures,  reference  lists  but  including  in-text references) is 1500 words. Markers will stop marking once the word count limit has been reached. Students who submit work that is below the word limit will not be penalised.

▪ The quality of your report is much more important than the word count.

▪ It is recommended (not mandatory) putting captions for your tables and figures, as posted sample scientific papers do.

▪ It is recommended to support your statements by referring to your textbooks, or any other  references you would find  in the  literature.  Minimum two  references are expected with at least one reference of a non-textbook nature.

▪ Clarity of statements, appropriate technical language and use of tables and figures are all important for this subject and for your assignment. Recall one figure talks more than a thousand words!

Guideline on marking for the assignment

Available marks: 100

➢ Overall clarity and structure of the  report,  including  referencing,  English, and quality of tables and figures (total 20 mark)

➢ Accuracy of answers for Statistics questions (total 80 mark)

Task 1. Coursework Brief (total 40 mark)

For this coursework task you are required to perform an Analysis of Variance using

SPSS. Your  solution  should  be  word-processed  and  submitted  electronically.  Your solution should include any output produced from the analysis, and an account of the methods you used to obtain that output.

Problem description:

Discounts and Expected Prices

Does the frequency with which a Tesco supermarket product is offered at a discount affect the price that customers expect to pay for the product? Does the percent reduction also affect this expectation? These questions were examined by researchers in a study conducted on students enrolled in an introductory Marketing Analytics course at Swansea University. For 10 weeks, 160 subjects received information about the products. The treatment conditions corresponded to the number of promotions (one, three, five, or seven) during this 10-week period, and the percent that the product was discounted (10%, 20%, 30%, or 40%), called discount rate. Ten students were randomly assigned to each of the 4 × 4 treatments.

For your case study, you examine the data for two levels of promotions (1 and 3) and two levels of discount (20% and 40%). Thus, we have a two-way ANOVA with each of the factors having two levels and 10 observations in each of the four treatment combinations.

The data are in the SPSS Supermarket_SZ40.sav file. Load the data into SPSS and investigate the meaning of the values of the variables.

Perform a suitable two-way Analysis of Variance to investigate the results. If appropriate, include a suitable multiple comparisons analysis and/or profile plot(s). Explain why it is appropriate to include (or exclude) such analyses.

You should address the following questions, explaining your answers in detail:

(a)  Are the main effects of the number of promotions and discount rate significant,

and is there any significant interaction between the number of promotions and discount rate? (10 mark)

(b)  What percentage of the variation in the DV product price is accounted for by the

model? Please briefly interpret the “tests of between-subjects effect” table. (10 mark)

(c)  Is any level of promotion significantly higher or lower than the others? If so, which level  of  promotion  is  higher?  Which  level  of  discount  rate  produces  higher expected price? (10 mark)

(d)  If your  requirement  is to  maximize  expected  price, which combination of the number of promotions and discount rate would you select? (10 mark)

Task 2. Coursework Brief (total 40 mark)

For this task you are required to develop a multiple linear regression model using SPSS.

Your solution should  be word-processed and submitted electronically. Your solution should include relevant output produced from the analysis, and a detailed account of the methods you have used, the reasons you have chosen those particular methods, and the conclusions you have drawn.

Problem description:

Car price prediction using multiple linear regression

A Chinese automobile company Geely Auto aspires to enter the US market by setting up their manufacturing unit there and producing cars locally to give competition to their US and European counterparts. They have contracted an automobile consulting company to understand the factors on which the pricing of cars depends. Specifically, they want to understand the factors affecting the pricing of cars in the American market, since those may be very different from the Chinese market. The company wants to know:

➢ Which variables are significant in predicting the price of a car

➢ How well those variables describe the price of a car

➢ Based on various market surveys, the consulting firm has gathered a data set of different types of cars across the America market.

Business Goal

We are required to model the price of cars with the available six independent variables. It will be used by the management to understand how exactly the prices vary with the independent variables. They can accordingly manipulate the design of the  cars, the business strategy etc. to meet certain price levels. Further, the model will be a good way for management to understand the pricing dynamics of a new market.

Import the Car_Price_SZ100.sav data into SPSS and develop an appropriate regression model to investigate the relationship between the dependent variable “car price” and other predictors, and report your findings. Your report should address (but not necessarily be confined to) the following questions:

Produce the Pearson correlation matrix for all variables .

(a)  If we are trying to model the relationship between target variable “car price” and

the  other  predictors,  from  these  correlations,  which  predictors  should  be considered for inclusion in the model? (10 mark)

Perform a regression analysis to predict the “car price” variable from all the other numeric variables, using the “Enter” method.

(b) What  percentage of the variation in the “car price”  is accounted for by other predictors? Should we keep all predictors as significant variables within the final regression model? What conclusions do you draw from the ANOVA table? (10 mark)