Based on what you're saying, it seems very possible to me that there are ranges at which the transponder signal can be receive but the skin paint has not.
Signals in both directions will attenuate according to the normal square-of-distance law. Suppose that the aircraft is at a distance from the radar such that the signal is attenuated to just below the detection threshold. That is the radar can not get a skin paint -- but barely. However, the strength of the radar signal that strikes the aircraft at that range is 4X (in an ideal world; in reality its probably even higher) the strength of the signal that arrives back at the radar receiver. This is because doubling the distance (which assumes perfect reflection; in reality it's not quite that good) will quarter the power.
Therefore, the transponder can very likely detect the incoming radar signal -- and respond to it -- at ranges beyond which the radar can get a skin paint, since it receives 4X as much power.
The question, then, is whether the transponder replies with greater energy than the reflection at those long ranges. If so, then there is a zone in which the transponder signal can be detected but the skin paint cannot. Turning off the transponder in that zone would make the plane instantly and completely disappear from radar.