An agency estimates how long job-training participants remain employed and will not assume that duration follows a named distribution such as the exponential or the Weibull. Which survival curve matches that stance?

A nonparametric survival curve estimates the chance of still being employed from the observed times and does not assume those times follow a named lifetime distribution. Refusing the exponential and the Weibull is exactly that stance, so the curve is nonparametric. It can follow the observed exit times without a rate, shape, or scale taken from a named family. An exponential curve assumes a constant rate of leaving and is a named parametric form the agency declined. A Weibull curve fits a named shape and scale and is also parametric, so it contradicts the refusal to lock in a named distribution. A single-day cross-section of who is employed on the survey date does not estimate a time-to-exit curve at all. Exam caveat: nonparametric does not mean free of assumptions, because the usual curve still treats censoring as unrelated to the remaining employment risk. Operational check: confirm the fitted employment curve is built from the observed times and that the specification has no exponential, Weibull, or other named lifetime parameter.