Beware of three main problems with Environmental Data: Its usually very noisy It is often unintentionally biased because the wrong variables are being measured to address the problem in question. A control sample is usually not available.

Elementary Data Analysis:

The first steps are to produce a distribution (histogram) of the variables and define the statistical components of that distribution (usually just the mean and the standard deviation.

After doing that one wants to attempt a simple linear regression to search for correlation in the data.

Benefits of Linear Regression:

• Defines the dependency of one variable on another in a relatively simple manner Linear: Y = ax + b; (most all relations in nature, however, are non-linear!)

• Allows for predictions to be made for values of X larger than the data set. Good for trend extrapolation (provided that the relation really is intrinsically linear).

• Using the dispersion around the relation, you can quantify your predictive power.

Linear Regression Applet