Methodological and Operational Problems
 

• Not all the data was stored to the same degree of precision.  Some layers had approximately 250 rows and columns, while others had 450.  Since I had to have all parameters the same for the MCE, I changed the layers with fewer rows and columns to match ones with higher values.  I did this when I reprojected my data.  I realize now I should have decreased the precision of layers to match ones with the least values.
• Also, the lack of precision made it difficult to perform a more focused analysis.  The data used approximately 450 rows and columns, which in my opinion makes for a generalized analysis.  The distance module used was not always very useful since cell values were a very short distance to features, resulting again in a very generalized output.
• Additionally I do not know any details on the data itself.  I could not find any meta-data which would explain to me the resolution, positional errors and other pieces of information.
• When I re-projected the data from plane view to "clabsha", Idrisi's recommended projection, some values were skewed.  Precipitation and temperature values increased in range slightly.  While I recognize errors will inherently propagate more errors, I do not think this analysis was compromised considering its generalized nature.
• Lastly, the fact that a geography student made arbitrary decisions on factors weights would pose as a likely source for error.  It should be mentioned that all attempts were made to eke out reliable sources for this.  In failing to find such a source, I decided to consider the recurring comments on agriculture web sites of the importance of climate.  I used common sense in determining other factors, including slope and wind.  I also decided (from a marketing point of view) the importance of proximity to population and water sources.