BUSI 444 – Marking Summary – Case Study Exam
Name: / Date:Grader Signature:
BUSI 444 – Case Study Exam / Needs Improvement / Meets Expectations / Exceeds Expectations / Grade
Report Introduction & Structure
Transmittal, Contents, Appraisal Foundations / /10
Analysis of City, Neighbourhoods, Highest and Best Use
Report quality: organization (format/structure), quality of discussion and logic
Cost Model
Appropriate land valuation with supporting explanation / /20
Cost model calibration and initial testing
Data Analysis and Transformations
Comprehensive data analysis with sufficient support / /10
Transformations: comprehensive, correct, solid discussion
Variable Selection
Appropriate variables chosen, supporting explanation / /30
Appropriate criteria, multicollinearity considered
Market Model
Appropriate processes and supporting explanation / /10
Appropriate testing, treatment of outliers
Final Analysis & Conclusions
Final testing with supporting explanation / /20
Conclusions, clear and supported
TOTAL GRADE / /100
MEETS AIC/IAAO DEMONSTRATION REPORT REQUIREMENTS? / YES / NO
Explanation, if NO:
Note: for all sections other than Report Introduction and Structure, 40% of the marks are for the detailed explanation of the process. As a result, any section without adequate explanation will generally result in a failing grade for that section.
Report Introduction & Structure
BUSI 444 – Notes and Common Errors – Project 3 – Linear MRA—Income Properties
□Definition of appraisal problem: no definition of “fee simple”.
□Definition of appraisal problem: no definition of “market value”.
□Highest and best use analysis omitted or not directly related to the properties to be appraised.
□Highest and best use analysis of the properties to be appraised not specifically stated.
□Definition of the three approaches to value omitted or no reason given for not using all three approaches.
□No description of the three types of depreciation in the discussion of the Cost Approach.
□Other:
Cost Model
□Confused the requirements of a Cost model with those for an additive regression model.
□Incorrect characteristic adjustments.
□Error in the IMPRCNLD transformation.
□Did not test the Cost value against the market value levels indicated by the improvement residual.
□Other:
Data Analysis and Transformations
□Limited explanation of reasoning/process for transformations.
□Need to examine / transform the new variables.
□Other:
Variable Selection
□Did not test at least two groups of variables and include the main regression reports for each (i.e., Model Summary, ANOVA, and Coefficients—showing VIF/Tolerance statistics).
□Failed to include or discuss any of the new variables included in the database.
□In the modeling process, did not recognize that the homes with a full basement and those with only a crawlspace will exhibit a difference in value.
□Multicollinearity errors in groups for initial selection.
□Did not apply the criteria given in the course material to select variables from the final STEPWISE regression report (e.g., not-value analysis; inaccurate analysis of significance).
□Did not give appropriate consideration to appraisal or real estate factors for variables selected for or removed from the model.
□Selected final group for calibration without correcting multicollinearity.
□Other:
MarketModel
□All coefficients do not have the expected sign; the F and R2 values are low; the COV is high.
□Residuals outside of±3 standard deviations are not filtered out and the model recalibrated until no further outlier residuals are reported.
□Adjustments to ensure common level of assessment for all variables have not been done or are inappropriate (either statistically or from an assessment perspective).
□Inclusion of multicollinearity in the final model.
□ Other:
Final Analysis & Conclusions
□TEST database should only be used to test acceptable value level; remainder of tests should be done on the FULL database.
□Final testing should be done on the FULL database.
□Failure to make appropriate use of Mann-Whitney and Kruskal-Wallis tests for binary and discrete variables (you do not need to include all M-W or K-W tests, but you must provide summary totals of these for all binary and discrete variables—in and not-in the model). If any of these tests fail you must also include ratio statistics reports and take any action to correct).
□Incomplete testing: require complete graphic and statistical tests on land characteristics, binary, and discrete variables; as well as sufficient description of test purposes and accurate analysis of results for variables in and not-in the model.
□Did not carry out ratio statistics reports or take any actions to correct.
□Inadequate comparison of the results between the models and insufficient discussion with recommendations as to which is the better model and which should be used.
□Other: