‹‹ Back to SVS Home

14.2 P-Value plot types

14.2 P-Value plot types

14.2.1 P-Value Plots sorted by Var #

For P-Value plots which are sorted by variable number, you can select which level of multiplicity adjustment for the p-values you would like to plot: P (raw P), aP (adjusted P), or bP (bonferroni-adjusted P).

In addition, you can consolidate either the (raw) p-values or the adjusted p-values using Simes’ method over a moving window of tests (potential splits). A slider allows you to select the size of this moving window. The result for each position of the window is shown for the test at the window’s center. (NOTE: Zero is shown for tests near the edge of the plot which are not at the center of any window.)

With Simes’ method, a process resembling finding the False Discovery Rate is used. If there is a significant effect within the window influencing more than one p-value, Simes’ method will show this. On the other hand, Simes’ method will “correct” single p-values that are much more significant than their neighbors. See 17.15.

14.2.2 P-Value Plots sorted by adjusted P-value

For plots sorted by adjusted P-value, by default, the FDR plot is shown. This plot shows the values of the False Discovery Rate (FDR) corresponding to the adjusted P-values (aP) of the respective tests (potential splits). The interpretation is, “What would the rate of false discoveries (false positives) be if I accepted ALL of the tests at or below the adjusted P-value of this test?” See 17.15.

You may also select a p-value plot using one of the multiplicity adjustments – P (raw P), aP (adjusted P), or bP (bonferroni-adjusted P).