Publication date: 11/10/2021

The Effect Summary report appears if your model contains more than one effect and if it can be calculated quickly. (If the report does not appear, select Likelihood Ratio Tests from the red triangle menu to make both reports appear.) It lists the effects estimated by the model and gives a plot of the LogWorth (or FDR LogWorth) values for these effects. The report also provides controls that enable you to add or remove effects from the model. The model fit report updates automatically based on the changes made in the Effects Summary report. See Effect Summary Report in Fitting Linear Models.

Note: The Effect Summary report is not applicable to models fit with Hierarchical Bayes.

The Effect Summary table contains the following columns:

Source

Lists the model effects, sorted by ascending p-values.

LogWorth

Shows the LogWorth for each model effect, defined as -log10(p-value). This transformation adjusts p-values to provide an appropriate scale for graphing. A value that exceeds 2 is significant at the 0.01 level (because -log10(0.01) = 2).

FDR LogWorth

Shows the False Discovery Rate LogWorth for each model effect, defined as -log10(FDR PValue). This is the best statistic for plotting and assessing significance. Select the FDR check box to replace the LogWorth column with the FDR LogWorth column.

Bar Chart

Shows a bar chart of the LogWorth (or FDR LogWorth) values. The graph has dashed vertical lines at integer values and a blue reference line at 2.

PValue

Shows the p-value for each model effect. This is the p-value corresponding to the significance test displayed in the Likelihood Ratio Tests report.

FDR PValue

Shows the False Discovery Rate p-value for each model effect calculated using the Benjamini-Hochberg technique. This technique adjusts the p-values to control the false discovery rate for multiple tests. Select the FDR check box to replace the PValue column with the FDR PValue column.

For more information about the FDR correction, see Benjamini and Hochberg (1995). For more information about the false discovery rate, see Statistical Details for the Response Screening Platform in Predictive and Specialized Modeling or Westfall et al. (2011).

The options below the summary table enable you to add and remove effects:

Remove

Removes the selected effects from the model. To remove one or more effects, select the rows corresponding to the effects and click the Remove button.

Add Profile Effect

Opens a panel that contains a list of all columns in the data table for the OneTable, Stacked data format, and for the columns in the Profile Data table for the Multiple Tables, Cross-Referenced data format. Select columns that you want to add to the model, and then click Add below the column selection list to add the columns to the model. Click Close to close the panel.

Add Subject Effect

Opens a panel that contains a list of all columns in the data table for the OneTable, Stacked data format, and for the columns in the Subject Data table for the Multiple Tables, Cross-Referenced data format. Select columns that you want to add to the model, and then click Add below the column selection list to add the columns to the model. Click Close to close the panel.

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