Teaching period 2, 2016
STA20007: Multivariate Statistics
Assessment 2: Workbook report 2
Word limit: n/a
Due date: 9am AEST Monday 5 September (Week 8)
The purpose of this workbook is to practice cleaning a data set and to practice all aspects of conducting a multiple linear regression, including reporting the results. For this assessment you will use two SPSS data files to answer questions.
In order to complete this assessment you will need to download three files:
• sta20007-wkbk2-answersheet (Answer sheet – which contains a template to write your answers to the questions in).
• wkbk2-q1.sav (this SPSS data file is for question 1).
• wkbk2-q2.sav (this SPSS data file is for question 2).
The management of a large insurance company were interested in identifying factors which might have an impact on workplace efficiency. In a hypothetical study, the following information was collected from a random sample of the company’s employees:
Efficiency Level of efficiency at work
Transport Mode of transport to work
ExerPartic Participation in workplace exercise program
Social Time spent socialising with colleagues
Morale Workplace morale
Stress Work related stress
Experience Level of experience in current position
The data is recorded in wbk2-q1.sav and more information about each variable is given in the labels in this data file. Assume that if a variable has no value labels, it is measured on a metric scale. It was suggested that people with more work experience, those who walk or cycle to work and those who feel less stressed tend to be more efficient.
There is also evidence to suggest that people with high morale, and those who have more social interaction with colleagues tend to be more efficient. Recently there has been speculation on the value of work place exercise programs. Researchers have suggested that providing an environment conducive to exercising at work would improve work efficiency.
Various links between these predictors have been proposed:
• Employees who take part in the work place exercise program, and those who walk or cycle to work are expected to have higher morale and to be less stressed
• Employees who are more experienced tend to be less stressed
• Those who have more social interactions with colleagues are expected to have higher morale.
(a) Look at the variables in the wbk2-q1.sav data file and decide how each variable has been measured (categorical or metric). Note that the level of measurement for all variables has been set as ‘nominal’, so you will need to look at the information given and redefine the ‘measure’ to ‘scale’ for some variables. DO NOT recode any variables. Fill in the level of measurement in the table provided in the Workbook 2 Answer document (sta20007-wkbk2-answersheet save it as SurnameAnswer.doc).
(b) Suggest a specific two stage hierarchical hypothesis that is consistent with the theory described above, and which can be tested using a hierarchical regression analysis. The dependent variable here is workplace efficiency, and you are expected to include four appropriate predictors. NOTE: It does not matter whether the data supports your hypothesis or not.
(c) In the hypothesis from part (b), which variables are acting as potential mediating variables?
(d) In order to test the hypothesis suggested in part (b), which predictors should be entered at stage 1 of the regression? Which predictors should be entered at stage 2?
There is no need to write a report for this question.
It has been suggested that the factors which predict job performance will be different across different types of professions. In one large accounting organisation researchers collected the following information from a random sample of employees:
Performance Job performance. Measured on a metric scale taking values from 0 to 100, with higher values representing a better performance.
Client Ability to form good client relationships. Measured on a metric scale taking values from 0 to 20, with higher scores representing a greater ability to form good client relationships.
Colleague Ability to work co-operatively with colleagues. Measured on a metric scale taking values from 0 to 20, with higher scores representing a greater ability to work co-operatively with colleagues.
Ascend Those scoring high in ascendancy attempt to influence and control other people. Measured on a metric scale taking values from 0 to 20
Social Participants were classified as either sociable or solitary depending on the extent to which they enjoyed social interactions.
Hypothetical data for these variables can be found in the wbk2-q2.sav data file.
It was hypothesised that ascendancy and sociability affect job performance indirectly through their effects on ability to form good client relationships and ability to work co-operatively with colleagues. Using SPSS, run a regression analysis to test this hypothesis.
Save your output file and upload it via the Blackboard website.
Write a concise report on the regression analysis. (USE the example and exercises in Section 3.5 of the textbook as a guide. The week 5 exercises might also be useful.) Type this report into the SurnameAnswer.doc document and also include all relevant sections of the SPSS output as an appendix at the end of this document.
This assessment will be submitted via Blackboard.
Please submit two files through Blackboard:
• Your Answersheet in Word document containing your answer to Questions 1 and 2 and your Appendix. (Save this document as YoursurnameAnswer.doc).
• An SPSS output file (.spv) created in Q2. Called SurnameOutput.spv
See the Assessment 2 section of Blackboard for more detailed information.
As the assessment is worth 20%, 20 marks will be allocated across the two questions as outlined
For Question 1 (worth 8 marks)
• Correctly identifies the type of variables for Q1,
• presents a sensible hypothesis,
• correctly identifies the mediating variables and
• correctly identifies the predictors for stage 1 and 2.
For Question 2 (worth 12 marks)
• Correct SPSS output produced (1 mark)
• Correct introduction and hypothesis, and type of analysis (1 mark)
• Correlation and regression table correct (1 mark)
• Correct wording of the relationship between DV and IV (1 mark)
• R-square and R-square change interpreted correctly (1 mark)
• Correctly identifies significant variables, which predictor is the most important, and the
direction of the correlation (2 mark)
• Correctly interprets mediation (4 marks)
• Well written (1 mark).