Publication date: 07/08/2024

Response Screening Plots

The default plots included in the Response Screening report window depend on the selections made in the launch window. The plots focus on False Discovery Rate (FDR) statistics. See Statistical Details for the Response Screening Platform. The following list describes all of the possible plots and when they are available.

FDR PValue Plot

Plots the FDR p-values and the standard p-values on the vertical axis and the Rank Fraction on the horizontal axis. This plot indicates how much adjustment is being applied by showing how many of the tests are significant with and without adjustment for the false discovery rate. The Rank Fraction ranks the PValues in order of decreasing significance. FDR PValues are plotted in blue and PValues in red. A blue horizontal line shows the 0.05 significance level. A red increasing line provides an FDR threshold for unadjusted p-values. A p-value falls below the red line precisely when the FDR-adjusted p-value falls below the blue line. This enables you to read significance relative to the FDR from either the adjusted or unadjusted p-values. There is also a legend next to the plot that indicates the number of tests and the portion of significant tests at both the 0.05 and 0.01 alpha levels.

FDR Logworth by Effect Size

Plots the FDR logworth on the vertical axis and the effect size on the horizontal axis. This plot helps graphically visualize very small p-values, which are associated with large effects. When transformed to the logworth (-log10(p-value)) scale, highly significant p-values have large logworth values and nonsignificant p-values have low logworth values. A logworth value of zero corresponds to a nonsignificant p-value of 1. Any logworth value above 2 corresponds to a p-value below 0.01. There is a reference line on the plot at logworth = 2 for quick visualization.

Generally, larger effects lead to more significant p-values and larger logworth values. However, this relationship is not necessarily strong because significance also depends on the error variance. In fact, large logworth values can be associated with small effects, and small logworth values can be associated with large effects, because of the size of the error variance. The FDR Logworth by Effect Size plot enables you to explore this relationship.

FDR Logworth by RSquare

(Available only for continuous Y variables.) Plots the FDR logworth on the vertical axis and RSquare values on the horizontal axis. Larger logworth values tend to be associated with larger RSquare values, but this relationship also depends on the number of observations.

FDR Logworth by Kappa Plot

(Available only if the Kappa option is selected in the launch window and there are X and Y columns that are categorical with the same number of levels.) Plots the FDR logworth on the vertical axis and Kappa on the horizontal axis.

FDR Logworth by Corr Plot

(Available only if the Corr option is selected in the launch window and there are categorical X and Y columns.) Plots the FDR logworth on the vertical axis and the correlation on the horizontal axis.

Logworth by Difference

(Available by default only if there is at least one continuous response, at least one categorical X variable, and the Common Y Scale option is selected in the launch window.) Plots the logworth on the vertical axis and the mean difference of the horizontal axis. If the Volcano Plots Use FDR Axis option is selected in the launch window, the FDR logworth is plotted on the vertical axis. Means differences are calculated for each level versus a control level. The default control level is the first level. If there is only one variable and it has more than two levels, there are check boxes that enable you to select which levels to show on the plot.

If a point is not significant (FDR p-value 0.01) it is colored gray. If a point is significant, FDR p-value < 0.01, it is colored red if the mean difference is positive and blue if the mean difference is negative.

Note: This plot is also called a Volcano plot.

Logworth by Slope

(Available by default only if there is at least one continuous response, at least one continuous X variable, and both the Common Y Scale and Common X scale options are selected in the launch window.) Plots the logworth on the vertical axis and the slope on the horizontal axis. If the Volcano Plots Use FDR Axis option is selected in the launch window, the FDR logworth is plotted on the vertical axis. The slopes are calculated from the simple linear regression between each response and each continuous variable.

There is a reference line to indicate significance. The reference line is located where the FDR logworth is 2, which corresponds to an FDR p-value of 0.01. Note that any logworth value above 2 corresponds to a p-value below 0.01. If a point is not significant (FDR p-value 0.01) it is colored gray. If a point is significant (FDR p-value < 0.01) it is colored red if the slope is positive and blue if the slope is negative.

Note: This plot is also called a Volcano plot.

Practical Difference

Plots the maximum of the FDR logworth difference and the FDR logworth equivalence on the vertical axis. The horizontal axis depends on specifications made in the Practical Differences and Equivalences window. If a portion is specified, the horizontal axis is the difference divided by the difference to detect. If a specific difference is specified, the horizontal axis is the difference.

Robust FDR PValue Plot

(Available only if the Robust option is selected in the launch window.) Plots the robust FDR p-values and the standard p-values on the vertical axis and the Robust Rank Fraction on the horizontal axis. This is similar to the FDR PValue Plot, but uses the robust values instead of the standard values.

Robust FDR Logworth by FDR Logworth Plot

(Available only if the Robust option is selected in the launch window.) Plots the robust FDR logworth on the vertical axis and the FDR logworth on the horizontal axis.

Robust FDR Logworth by Effect Size

(Available only if the Robust option is selected in the launch window.) Plots the robust FDR logworth on the vertical axis and the effect size on the horizontal axis. This is similar to the FDR Logworth by Effect Size plot, but uses the robust FDR logworth instead of the standard FDR logworth.

Cauchy FDR PValue Plot

(Available only if the Cauchy option is selected in the launch window.) Plots the Cauchy FDR p-values and the standard p-values on the vertical axis and the Cauchy Rank Fraction on the horizontal axis. This is similar to the FDR PValue Plot, but uses the Cauchy values instead of the standard values.

Cauchy FDR Logworth by FDR Logworth Plot

(Available only if the Cauchy option is selected in the launch window.) Plots the Cauchy FDR logworth on the vertical axis and the FDR logworth on the horizontal axis.

Cauchy FDR Logworth by Effect Size

(Available only if the Cauchy option is selected in the launch window.) Plots the Cauchy FDR logworth on the vertical axis and the effect size on the horizontal axis. This is similar to the Cauchy Logworth by Effect Size plot, but uses the Cauchy FDR logworth instead of the standard FDR logworth.

Want more information? Have questions? Get answers in the JMP User Community (community.jmp.com).