RES5115 Research Preparation: Principles And Approaches - Quantitative Assignment

QUESTION 2 [25 MARKS]

The Excel file “Credit.xlsx” contains data relating to credit card balance (US$) of 310 U.S. citizens. Included in the dataset are relevant socio-demographic variables, namely:

1. Income: Income in $10,000’s
2. Limit: Credit limit (US$)
3. Cards: Number of credit cards
4. Age: Age in years
5. Education: Number of years of education
6. Gender: Two levels – Female or Male
7. Student: Two levels – No or Yes
8. Married: Two levels – No or Yes
9. Ethnicity: Three levels – African American, Asian or Caucasian

* You may ignore the ID (1st) column in the dataset.
The objective of the study is to determine whether any of the 9 socio-demographic variables are predictive of an average U.S. citizen’s credit card balance.

TASK 1 - SELECTING A SUB-SAMPLE

• Open the Credit.xlsx data file as shown.
• Click on the “Data” tab and then run the Analysis ToolPak by clicking on Data Analysis.
• Select Random Number Generation and click OK.
• Select “Uniform” in the drop-down menu next to “Distribution”. Then fill out the other boxes as shown. In the box corresponding to “Random Seed”, make sure you type the last two digits of your student ID here. This will ensure the set of random numbers is unique to you (unless someone else shares the same two numbers).
• You should now be able to see 310 random numbers in Column L. Take note that your numbers will be different to those shown here.
• Highlight all the data from Columns A to L using the mouse/keyboard. Then click on and select Custom Sort...
• Now click on underneath Column → Sort by and select (Column L). Then click OK.
• The data are now sorted according to Column L. Again, note that what is shown below is very likely to be different to what you will have.
• Now, select the first 100 observations including the headings (i.e. Rows 1 to 101) from Columns A to K. Copy this sub-sample across to another spreadsheet (e.g. Sheet 2) in Excel. In the example below, Sheet 2 was renamed to Sub-sample.

Get Your Customize Task on any subject starting 10$/Page

Order Now Whatsapp Now

TASK 2 – SELECTING THE SOCIO-DEMOGRAPHIC VARIABLES TO ANALYSE

Now that you have your sub-sample, here you will determine which THREE (3) of the 9 variables to analyse. Again, go to the “Data” tab and then run the “Data Analysis” tool pack. Select “Random Number Generation” and click “OK”.

Now generate 9 random uniform numbers and store them in Column M of the sub-sample spreadsheet. In the “Random Seed” box, type the last two digits of your student ID again.

You should now be able to see 9 random numbers in Column M. Take note that your numbers will again be different to those shown here. Now, type the numbers 1 – 9 (to represent the variables) in Column N and right next to the random numbers that you have just generated.

Now, highlight all the numbers in Columns M and N using the mouse/keyboard. Then sort the numbers by
clicking and then Sort Smallest to Largest.

The first three numbers in Column N refer to the variables listed on page 8 that you will need to analyse. In this example, the variables are (1) Income, (3) Cards and (8) Married. Copy the relevant data (do not include the headings here) for the 3 variables and for Balance across to SPSS. These observations form the dataset that you will be working with for this question.

TASK 3 – ANALYSING THE DATA

1. Use SPSS and generate all necessary summary statistics and figures (e.g. boxplots, histograms and/or scatter plots) to describe the relationship between the credit card balance of U.S. citizens and each of your 3 selected variables (make sure you consider each item separately from the others). Proper interpretation of these output in the context of the problem is expected.
2. Use the appropriate test in SPSS and determine whether there is a relationship between the credit card balance of U.S. citizens and each of your 3 selected items. If so, comment on the nature and extent (e.g. via regression modelling) of this relationship. If you are comparing across groups (e.g. via t-tests, ANOVA or non-parametric tests), then you will need to comment on the significance and extent of these differences (if any). Hypothesis statements are not necessary.
3. Confidence intervals should be presented and interpreted where appropriate.
4. All relevant assumptions associated with your chosen test must be verified.