I believe you are mistaken about only one covariate. See the attached and its references for full details. Best wishes, David Booth

BTW multivariate means more than one DV. I assumed you meant multivariable..Best wishes, David Booth

Mahmoud Elmaghraby, in SPSS terminology, covariate typically refers to a quantitative explanatory variable, and factor refers to a categorical explanatory variable. You are right in saying that the KM command allows only one explanatory variable, although it is a factor variable, not a covariate. The following is from the Command Syntax Reference manual for IBM SPSS Statistics, v 26.

Survival and Factor Variables (KM command)

You must identify the survival and factor variables for the analysis.

• The minimum specification is one, and only one, survival variable.
• Only one factor variable can be specified using the BY keyword. If you have more than one factor, create a new variable combining all factors. There is no limit to the factor levels.

Source: https://www.ibm.com/docs/en/spss-statistics/26.0.0?topic=km-survival-factor-variables-command#syn_km_survival_and_factor_variables

However, one can also include a stratification variable. The next bit is from another section of the Command Syntax Reference Manual.

STRATA Subcommand (KM command)

STRATA identifies a stratification variable—that is, a variable whose values are used to form subgroups (strata) within the categories of the factor variable. Analysis is done within each level of the strata variable for each factor level, and estimates are pooled over strata for an overall comparison of factor levels.

• The minimum specification is the subcommand keyword with one, and only one, variable name.
• If you have more than one strata variable, create a new variable to combine the levels on separate variables before invoking the KM command.
• There is no limit to the number of levels for the strata variable.

Source: https://www.ibm.com/docs/en/spss-statistics/26.0.0?topic=km-strata-subcommand-command#syn_km_strata

David Eugene Booth, I did not understand your statement that Mahmoud Elmaghraby was mistaken about only one covariate. Did you think he was saying that the SPSS COXREG command allows only one explanatory variable? Or was your objection that software other than SPSS allows more than one explanatory variable for KM survival analysis? I do agree with you that he most likely meant multivariable, not multivariate.

Cheers,

Bruce

I would be most happy to supply any individual with a simple R program that does all that was requested and uses modern Biostatistics terminology. In addition the program does modern adaptive lasso variable selection suitable for predictive models and associated Kaplan Meier plots. if variable selection is not necessary a very simple R program that does all but variable selection can be found in Jared Lander's R for everyone available for download in the z-library. As is well-known R can be run on virtually any device including a smart phone. I am attaching two papers that demonstrate the versatility of R, the single most important package in modern Biostatistics. R A FISHER was a great statistician and geneticist but he died in 1962. There's no reason to continue to use desk calculator language and techniques when much more efficient powerful and less confusing languages that correspond to modern Biostatistics terminology, needs AND provides for all current Biostatistics needs for free. Let's move into the modern era as all else in Biostatistics has done. With best wishes, David Booth

David Eugene Booth

Bruce Weaver

It seems that I need to adjust my confusing terminology. Yes, it is a multivariable analysis (7 explanatory variables, 7 X’s, and one dependent variable Y). All variables are categorical.

Thank you for help.

Again the R programs are available and questions are always welcome. Good luck with your research. David Booth