DSc8450

Problem Set 1

Spring 2000

Brian Schott

In my directory (class/dsc845/files) are two files PROPERTY (the data set with 24 observations) and PROPERTY.PROG (a SAS INPUT statement for the data). PRICE is the appraised value of the residential property and the other variables are characteristics of the property. The data are quite dated but you may be able to meaningfully analyze and interpret the data. As a minimum, perform the following analyses.

1 Data examination: Choose **three** variables.

- a)Examine central tendancy, dispersion, skewness, shape of the distribution,
normality, outliers, etc.
- Apply statistical significance testing when it is appropriate for you data examination process, but do not limit yourself to this formal approach. Often there are multiple indicators for a particular feature and the indicators may not all be consistent; you may have to weigh incompatible evidence, objective or subjective.

- b)Analyze and comment on the pairwise relationships between the variables you selected.

2 Factor analysis (starting with **all** the variables).

- a)Use the MSA statistics to possibly limit the variables that you proceed with.
- b)Using principle components analysis, determine a number of factors to extract that will accomplish a modest level of data reduction while explaining a healthy proportion of the total variation in the data set.
- c)Extend the resulting principal components factor analysis with a varimax rotation. Was the rotation successful at achieving "basic structure?"
- d)Attempt to name the factors. How natural do you think the names are
in this case?

e)Output factor scores using PROC SCORE (you do not need to print the individual scores) and demonstate that they are standardized and uncorrelated.

Due Wednesday, February 21