Background

Frostbury is a supermarket in a small conurbation in the north of England. It caters for a population of

around 3000 households. Although its main business is food, it has extended to other items such as clothes

and white/electrical/electronic goods. The management of Frostbury, anxious of the drop in retail takings,

wanted to understand the behavioural pattern of all its shopping population in order to respond correctly

to their need. It has introduced its own loyalty card and local people can apply for it using an application

form made available at the supermarket. Those shoppers opting for this card will earn loyalty points that

can be claimed via air miles and against future spending.

The application form for the loyalty card has a small questionnaire where customers can give their opinion

about five key motivations for their choice of Frostbury to do their shopping, namely, the location of the

supermarket, the presence of the car park, the opening hours, the till service and helpfulness of staff. The

questions are stated as follows:

Rate on a scale 1 to 7 (1 = Not at all important and 7 = Very important)

the following five motivations for shopping at Frotsbury.

As well as asking people for their personal details such as the gender of the applicant, their age, marital

status and the number of children still living with them, the applicants were also asked to give an estimate

of the time it takes them to get to the supermarket.

The supermarket management are interested in the following three key research questions and possible

extensions to them:

1. The demographic profile of applicants

2. Their motivation for choosing to shop at Frostbury

3. Their respective monthly spending

4. The prediction of future spending

The demographic data of applicants and their motivations for shopping with Frostbury are obtained from

the application forms whereas the data on spending was obtained from the loyalty card database. A sample

of 163 customers has elected to apply for the loyalty card. The whole information is made available on an

SPSS file named “LCholders.sav”. Some responses are missing and are represented by codes such as “99”

and “999”, therefore before you undertake any analysis, you should re-label the responses properly and

identify the missing data.

BA5001: Business Decision Making

Year: 2013-14

Coursework 1Page | 2

Part one: Analysis of data

You work for a market research agency and the management of Frostbury have approach your agency for

analysis of their data. They presented you with the data file and their research objectives as stated in their

research questions. For this part, they are interested in the first three research questions.

Since your CV stated that you have experience in working with SPSS software and a decent enough

education on statistical methods, your line manager has chosen you for this project.

Research question 1: The demographic profile of applicants

Required:

ï‚· Using SPSS, produce appropriate frequency tables of the demographic profile of shoppers who

applied for Frostbury loyalty card. The profile will have to include their gender, marital status, and

the mode of transport they use to get to the supermarket.

ï‚· Using SPSS, produce three appropriate charts to describe the age, the number of children still living

with the family and the type of accommodation they live in.

ï‚· Comment on your SPSS output.

Note: For this section, you need to report on what the frequency tables and charts tell you. The report

should be no more than 400 words in total and should include your output placed appropriately

within your report.

Research question 2: Motivation for choosing to shop at Frostbury

Required:

ï‚· Using SPSS, calculate the minimum, the maximum, the median, the lower quartile and the upper

quartile for the ratings of the shoppers’ motivations for choosing to shop with Frostbury. All five

motivational items should be considered.

ï‚· Comment on the single most important motivation. Also comment on the single most consistent

motivation among this sample of shoppers.

Note: Again for this section, you need to produce a small report of no more than 100 words and include

the SPSS output. The SPSS output, which is a single table, should be inserted with your report.

Research question 3: Monthly spending

Required:

ï‚· Using SPSS, draw a histogram and calculate the descriptive statistics for the January spending

among this sample of shoppers. Comment on the shape of the spending distribution as well as its

average and spread.

ï‚· Calculate the descriptive statistics of the January spending split by:

o Gender

o Accommodation

o Transport used

o Number of children living the house

You should include the sample size, the mean, median, skewness and standard deviation.

Comment briefly on the results.

Note: Again for this section, you need to produce a report of no more than 300 words (excluding the SPSS

output). The different parts of the SPSS output need to be inserted within your report to enable the

reader to check your results. In particular, you need to comment on the distribution of spending

across the whole sample and then comment on which profile is high spender and/or low spender.Page | 3

Part two: Hypothesis testing and linear regression

Extension to research questions: Hypothesis testing

Industry standards report that at least 40% of shoppers should be able to walk to their supermarket

without needing public or private transport of some sort. Equally, the same standards indicate that the

location of supermarkets is such that the average travelling time is about 12 minutes.

Another report, published by a consumer research agency, suggested, among other things, that the

availability of parking at supermarkets is a significant incentive among shoppers, particularly among those

using their own private transport. The same report stated that male and female shoppers see the

supermarket’s opening hours as highly important, though they rate them differently. It was also stipulated

in the same report that average monthly spending per household at supermarkets depends on whether you

live in a flat or a house.

Required:

Using information from this sample of loyalty card applicants and an SPSS hypothesis testing facility to,

ï‚· Test one: Test if industry standards apply in relation to the proportion of all shoppers to Frotsbury

able to walk to the supermarket.

ï‚· Test two: Test if industry standards apply to travelling time of all Frostbury shoppers.

ï‚· Test three: Test the ratings of motivations for shopping at Frostbury against each of gender, marital

status, type of accommodation, number of children and mode of transport used, and identify the

significant associations only.

ï‚· Test four: Test if the January average spending of all shoppers is significantly different between

those who live in a flat and those who live in a house.

ï‚· Comment on each of the four hypothesis tests above in relation to the information obtained from

the two reports stated in the preamble above.

Note: For this section, you need to produce a statistical reporting of your results in no more than 200

words. Report on each test separately. For each test, you need to state clearly the hypotheses (Null

and alternative), indicate the level of significance you have chosen and then make a decision based

on your test result. You must only use SPSS to conduct your statistical test. Each “useful” SPSS

output should be placed within the report. No statements are acceptable unless they are followed

by appropriate statistical evidence taken from your SPSS output.Page | 4

Modeling future spending – Simple linear regression analysis

The spending by the loyalty card holders was also given for the month of February. We want to investigate

whether the amounts spent by cardholders in February can be explained by the amounts spent in January.

Required:

ï‚· Use SPSS to draw a scatter diagram of spending in February against spending in January.

ï‚· Use SPSS to find a linear equation between the two monthly spending amounts.

ï‚· Comment on your results and what they mean in this context. Your comments should include aspects

of the scatter diagram, the coefficient of determination, the significance of the model as a whole and

whether gradient coefficient is statistically significant.

Note: As evidence, your output should include tables of the Model summary, Anova and coefficients.

Comment on your simple linear model.

Modeling future spending – Multiple linear regression analysis

The intention here is to build a linear regression model for the February spending using the following four

variables:

 The spending by loyalty card holders in January (the previous month)

 The age of applicants

 The number of children still living with the parents

 The time taken to get to the supermarket

In essence, your dependent (or outcome) variable is the February spending and your independent (or

explanatory) variables are the four stated above (in bullet points).

Required:

ï‚· Use SPSS to produce the best multiple linear model for predicting the February spending.

ï‚· Comment on your multiple linear model. Your comments should include the variables

included/excluded by the model, aspects of the variation of the February spending amounts that are

accounted for by the model, the significance of the model (and its gradient variables) and the outliers

to this model.

Note: Your output should include tables of the model summary, Anova, Coefficients and the Casewise

diagnostics (outside 3 standard deviations). Also, your comments should be written as if they are addressed

to your manager at the market research agency. Your manager is familiar with the statistical terminology so

you can use technical language in this report. However, you will also need to explain the implications of

your findings in more general, formal English.

General expectations:

ï‚· You are expected to write in a formal academic English style when producing your report and you

should not include the data set. Your comments should be written in a report form with separate

paragraphs, titles and sub-titles (if needed).

ï‚· In the comments, you can refer to certain parts of your SPSS output to help the readers (Markers in

this case) check the evidence of your findings.

ï‚· Your SPSS output can be either screen-shot (or copied and pasted) on a word processor together with

your report so the reader can see your comments and check your results at the same time.

ï‚· The reports you are including in this coursework are purely evidence based and should not contain any

“theory” notes taken from books or the internet.

ï‚· Include a bibliography of any sources you have used/read to complete the coursework.

 Note that comments such as “there are more males than females” or “the most important motivation

is the opening hours” are not acceptable unless they are followed by appropriate statistical evidence

such as “Males represent 30% of the total” for example.