Parametric t-test

See Also:

Description:

The parametric t-test is implement in ChemStat as either a group comparison following U.S. Navy (1999), whereby compliance well measurements are compared as a group to background measurements as a group, or as an individual well comparison as described in USEPA Statistical Training Course for Ground Water Monitoring Data Analysis (1992) on page 28-29. In the USEPA method, each individual compliance well is compared to background.

Because the test is parametric, the data must follow a normal distribution. Non-detects should be minimal. The test can be run at the 90%, 95%, 97.5%, or 99% confidence levels.

Use:

The parametric t-test provides a simple group comparison whereby each individual compliance location is compared to background or all compliance measurements as a group are compared to background.

Remarks:

ChemStat's implementation of the parametric t-test can perform individual well comparisons at either the 90%, 95%, 97.5%, or 99% confidence levels. If you are unsure of which level to use, the 99% probably preferable.

As an individual well comparison, the test is always a two-way comparison, but compliance locations are not flagged as statistically significant unless they are elevated compared to background. The exception is for pH, where compliance locations are flagged as statistically significant if they are either above or below background.

As a group comparison, the test only indicates statistical significance if the compliance group is elevated when compared to background, as per U.S. Navy (1999).