# chi squared (hermit crabs and pieces of plastic)

#### patricksean

##### New Member
Good day all,

new to this and stats overall so I hope i phrase this appropriately

one of our students created an experiment to look at the relationship between number of pieces of plastic and number of hermit crabs on a beach using quadrasts. They would like to first do a chi squared to look at plastic present (y/n) versus crabs present (y/n) but i'm having trouble explaining how to find the expected values.

1. would this make sense to do?
2. Could I ask someone experienced how you find the expected values?

much appreciated

thank you

sean

#### gianmarco

##### TS Contributor
Hello,
I am not an expert in quadrants analysis; I recall having read something about that some time ago, but it was related to when you have counts of "things" in each quadrant, not just presence/absence.
Maybe some of the guys here expert in natural sciences will jump in to provide some guidance.

Did you try to search a viable strategy in books dealing with spatial analysis in natural sciences?

#### gianmarco

##### TS Contributor
Upon further thought, I think that one way of dealing with your data is to build a simple 2x2 cross-tabulation. In each cell you may want to put the count of how many quadrants had crabs and plastic pieces, how many had only grabs, how many had only plastic pieces, and so on.

Something like this:
Code:
...........n  pieces pres n plastic abs
n crab pres	10	         5
n crab abs	5	         10

Then, a chi-square test (or Fisher's test) can be put to work.

gm

#### patricksean

##### New Member
Gianmarco, thank you for the help, you have been very very helpful. Have a great day, Sean.