Some Guidelines for the Assignment
Some Guidelines for the Assignment
Approach to the assignment
Model Building
Given the information available to you, think about what factors may explain house sale prices.
To avoid a possible issue of the omitted variable bias you may include all possible factors as the explanatory variables in your model. This includes creation of interaction variables.
To search for the best possible explanatory mechanism for house sale price you may explore the logarithm transformation and/or dummy variable techniques.
Model Identification
Clearly demonstrate how you identify your final model.
Use the t tests technique in combination with the adjusted R-square technique to justify your identification of final model. The aim here is to try to find the best quality fit model according to the adjusted R-square criterion.
Model Interpretation
Discuss your findings drawn from your final identified model.
Identify factors that explain the outcome of house sale price. Pay particular attention on interpretation involving interaction variable(s) and data transformation if justified.
Structure of assignment
Executive summary (no more than 100 words)
Describe your approach and summarise your key findings. In this section, your major objective is to describe the approach and key findings of your study to non-technical readers of your report (such as policy makers, practitioners, investors etc). Therefore, you need to avoid very technical terms and concepts in the executive summary.
Introduction (about 150-200 words)
In this section you present motivations for your work, state your objectives/tasks, discuss your approach, dataset and present your main findings. Discuss limitations that may have adverse impacts on your findings.
Regression analysis (about 400 words)
In this section you discuss how you construct your model(s) and how you identify your model. You need to demonstrate how you can effectively apply the techniques learnt in the course in the construction and identification of your model.
You may have explored many possible explanatory mechanisms but should keep very brief on those models that do not support your findings. You should focus on your discussions on how you identify your best model(s) from which you draw your findings among many models you have tried.
Below are some questions you may ask yourself:
What are possible explanatory mechanisms for house selling price?
How would you model these possible explanatory mechanisms?
Log or non-log?
Different effects for different groups? Dummy technique including interaction term
How would you demonstrate that you are able to apply statistics techniques to identify the best model which perhaps nobody knows?
If there exist two or more very competitive models, would they give similar findings or contrast findings?
Are your findings likely exposed to limitations? Residual check?
Results (about 200 words)
Present your regression analysis result with rigorous justifications. Be mindful about possible limitations of your findings.
Conclusion
Summarise what you have done, discuss how your findings can benefit others, etc.
Other notes:
The presentation of your write-up is important. Poorly presented work may result in loss of marks.
Penalty can be applied for:
Excessive errors/typos
Poorly formatted tables and figures
If there are tables and figures not discussed in the text
Late submission
