Research Update 3: The Last Shell Has Finally Come

With thousands of oysters to analyze at the start of the summer, the end looked like it would never come.  However, I have officially finished analyzing every precontact shell in my collection!  Now that I have finished my data collection, it is time to begin exploratory data analysis, looking for trends in the results that tell interesting stories.  This has been my main objective of this past week, and likely the coming week.  At the start of this week, Professor Gallivan introduced me to a statistical software system called SPSS.  We debated over whether I could conduct my data analysis through Excel or SPSS.  However, due to my lack of knowledge on both platforms, it was decided I should put my effort into learning SPSS since I would likely use it more frequently in the future.  The first day or two of exploring SPSS left me very confused.  The initial process of data entry on SPSS is more complicated them in Excel, as there are two pages for each data set to manage–Data View and Variable View.  However, the more I have experimented with the software, the more comfortable I have become with it.  Over the week, I have input all my data into the system.  Then I went through all my data doing an accuracy check to ensure every column was completely filled in and there were no values or designations that seemed obscure from the rest of the measurements.  I only had a few spots that left me with questions, and for those, I went back through my oyster collection and reanalyzed the shells in question for all attributes.  After double-checking all my data, I created a table that simplified the data setup from a per shell factor of analysis to a “feature” factor of analysis.  Since there were over two thousand individual shells analyzed and only 14 feature they originated from, determining important trends between features is much easier than between shells.  With this being said, I calculated the mean, or average, of all the shells found in each feature for all of the attributes measured–mass, length, height, height-length ratio (HLR), presence or absence of an attachment scar, presence or absence of parasitism, percentage of parasitism, parasitism on attachment scar, and left-valve concavity (LVC).  By looking at the table, I have been able to identify possible trends to further explore with histograms and box plots that show the actual variations between the features.  One of the trends that I am extremely interested in is the relationship between shell size and parasite presence with the type of roasting pit in which they came from.  The site report for site 44YO0797 identified two types of roasting pits, some that show evidence of heavy firing and others that do not.  Currently, it appears that the roasting pits that have lesser evidence of firing contained the larger, more likely offshore oysters, and I wonder if this has anything to do with preservation techniques of the Native American.  Site 44YO0797 was a secondary base camp for the Native Americans used for gathering supplies, so I am curious in exploring the possibility that the roasting pits with evidence of firing that contained smaller, near-shore shells because those were the oysters eaten on-site, while the roasting pits without evidence of firing had more larger offshore oysters because they were preserved for later use.  At this point, this is merely a theory that I hope to explore more with continued data analysis.  Additionally, I plan to compare my findings to those of Jessie Jenkins and Martin Gallivan who conducted a similar study of oyster shells at a different Native American site along the York River that was a primary settlement instead of a secondary supply gathering camp.