Imagine you have collected data on the temperature and dissolved oxygen (mg/L) in a eutrophic lake over 5 years. You want to see if dissolved oxygen varies with temperature in this lake. What statistical test would you do? Explain why you chose this test.
1 Answer
-
You have a data-point for each temperature and dissolved oxygen at that temperature so you plot temperature on one axis and dissolved oxygen on the other axis, and you do a regression test from which you can get p-values, and a coefficient of determination (R^2) from which you can determine of there is a relationship.
However, you should plot the graph of Temperature vs Dissolved oxygen. Don’t just use linear regression if the graph doesn’t look like a line. If it looks exponential, then transform the axes to make it linear and then use linear regression or an iterative nonlinear regression method.
The most common linear regression method is least squares linear regression.
Essentially you are measuring how well one variable predicts the other (how close the points fall to the curve)
