Data collection

Data to be taken on the system will include the following variables.

• Position of each hole in the grid.
• Size of each goby per hole
• Size of each shrimp per hole
• Number of goby and shrimp per hole
• Density of holes per cage
• Density of gobies per cage.

Each month the cages will alternate between caged and ¾ caged. It is difficult to determine the effects of predators on a system like this.

T-test
If I do it for 4 months with my 6 cage layout, then I will get a total of 12 ‘fully caged’, 12 ‘partly caged’, and (as a control) 12 non-caged. These can compared to each other using a t-test, and looking for the difference between the two set-ups. However, if this test will not be able to look to see if there is a flux in animals over time. In other words, if gobies are recruiting more on the summer and less in the winter, then all results in the winter will be lower anyhow. To look for this difference I will need to use another test.

Chi-squared test
A chi-squared test can be used to test the results against the expected. In this case, the variable to be tested is the change between caged and ¾ caged. In this case, there should be a difference of either positive or negative. This change can then be compared to the monthly change of the control plots. If a certain month has an influx of goby-shrimp pairs, then analysis will take this into account by calibrating the numbers.

For the chi-squared test, with 6 cages, there will be a total of 9 transitions from fully caged to ¾ caged and a total of 9 from ¾ caged to fully caged.