In survival analysis, what is a survival object? Relating to the Surv function in R. 
Could you please explain this to someone without an extensive statistics background?
 A: "Survival object" is not a statistical term. What is meant here in the R documentation is "object" in the programming sense of the word. The survival package uses this kind of object for things such as the response variable in survreg.
A: Survival analysis requires keeping track both of the follow-up time and whether the last follow-up time represented an "event" or not. (In the latter case it's called a "censored" observation.) The simplest survival object produced by Surv in R, for example, can be thought of as a matrix, with one row for each case, the first column representing the last follow-up time and the second indicating whether there was an event at that time (typically 0 for censored observations, 1 for events). For other uses (e.g., interval- censored data or counting-process analysis), the object may have 3 columns. Programs in R that do survival analysis typically expect an object produced by Surv, although sometimes you have to provide separate time and censoring/event variables.
Be sure to read the documentation for whichever statistical analysis program you are using, as the coding for "events" can differ from the above. R, for example, also allows 1/2 coding for censored/event, and MATLAB, as I recall, has 0 for event and 1 for censored as its default.
