Please enable JavaScript to view this site.

The data summary section of the tabular results sheet generated for Cox regression is fairly straightforward. This section contains information on the following values:

Rows in table - the total number of rows in the data table containing at least one value for variables included in the analysis

Rows skipped (missing data) - rows that contain at least one value for variables in the analysis, but are missing a value for the variable assigned to elapsed time, the variable assigned to the outcome (event/censor), or variables assigned as predictor variables

Rows analyzed (#observations) - equal to the difference between "Rows in table" and "Rows skipped (missing data)", this represents the number of rows in the data that contain values for all variables included in the analysis

 

 

Number of ties - the number of observations with events that occur at non-unique elapsed times (in other words, if any row contains an event that occurred at the same elapsed time as another row, this is considered to be a "tie"). Note that censored observations and their elapsed times are not considered when calculating this value

 

 

Number of censored - the number of rows included in the analysis with a "censored" value in the outcome (event/censor) variable

Number of deaths/events - the number of rows included in the analysis with an "event" value in the outcome (event/censor) variable

Ratio - equal to "Number of censored" divided by "Number of deaths/events"

 

 

Number of censored - same as reported above, the number of rows included in the analysis with a "censored" value in the outcome (event/censor) variable

Number of observations - same as reported above, the number of rows in the data that contain values for all variables included in the analysis

Ratio - equal to "Number of censored" divided by "Number of observations"

 

 

Number of observations - same as reported above, the number of rows in the data that contain values for all variables included in the analysis

Number of parameter estimates - the number of beta coefficients fit by the model

Ratio - equal to "Number of observations" divided by "Number of parameter estimates"

© 1995-2019 GraphPad Software, LLC. All rights reserved.