Select the graphs to display for Fit Cox Model with Fixed Predictors only

Stat > Reliability/Survival > Cox Regression > Fit Cox Model with Fixed Predictors only > Graphs

Display survival plot for predictor values

Displays the survival plot. Use a plot with a single curve to assess the probability of survival over time for the function on the plot. Use a plot with multiple curves to compare the probability of survival for different settings of the predictors. By default, Minitab uses the mean of each predictor for continuous predictors and the reference levels for categorical predictors. You can select a different option in the dropdown to specify new predictor values for the survival plot. For a categorical predictor, a new value must be one of the levels of the categorical predictor.

Enter individual values: Specify individual predictor values that Minitab uses for the survival function. If you do not specify a value for a continuous predictor, Minitab uses the mean of that predictor. If you do not specify a value for a categorical predictor, Minitab uses the reference level.
Enter columns of values: Specify a column of predictor values that Minitab uses for the survival function. The column type must match the variables in the model.

If the model is stratified, you can use Display plots for each stratum to decide how Minitab displays each stratum.

In separate panels of the same graph: Displays the survival plots for each stratum in a separate panel of the same graph. This setting is the default setting.
Overlaid on the same graph: Overlay the survival plots for each stratum on the same graph.
On separate graphs: Displays the survival plots for each stratum in a separate graph in the output pane.

Deviance residuals versus risk scores

Displays a graph of the deviance residuals versus the risk scores. The graph measures the effect of a given subject on the model. Use this plot to detect outliers.

Andersen plots for the stratification variables

Displays the Andersen plot for the stratification variables. Use this graph to assess the proportional hazards assumption for the strata. If the assumption holds, the curves are straight lines through the origin. If the variables do not need to be in the model, the curves roughly follow the 45° line. Minitab displays this plot only if the model is stratified.

Arjas plots for the variables

Displays the Arjas plots for the variables that you specify. Use this plot to assess the proportional hazards assumption for a categorical predictor. Also use this plot to assess whether a predictor makes a useful contribution to the model. You must specify at least one column to create a plot.

You must discretize a continuous variable before you display the Arjas plot for the variable. For example, if researchers have biological knowledge that some values of a predictor are low and the remaining values are high, then one technique is to discretize the predictor into two categories to create an Arjas plot. The analysis includes a model that uses the original predictor and an Arjas plot of the discretized predictor, even though the discretized predictor is not in the model.

Martingale residuals versus the variables

Displays graphs of the Martingale residuals versus the variables that you specify. Use this plot to assess whether you should add a predictor to the model, or whether you should use a different functional form for an existing predictor. For example, use a plot of the Martingale residuals with a variable to assess whether a square term for a continuous predictor would improve the fit of the model. You must specify at least one column. The column must be numeric or date/time data and must have the same number of rows as the response column.