Hypothesis Testing - Likelihood Ratio Test

The Likelihood Ratio Test (LRT) is a general, powerful method of testing two models when one model is a nested or special case of the other. In the illustration below, "CommonS" is a special case of "SPerPop" and therefore is displayed in the Testable Model List when "SPerPop" is selected in the Master Model List. If "CommonS" is then selected in the "Testable Model List", a Likelihood Ratio Test is performed, with the test results appearing in the lower left of the SURPH Dialog Window.

LRT.gif (25K)

Top of page