Does exposure affect sensor dynamic range?

Let's say that I want to test Camera A vs Camera B in terms of sensor DR.  Assume that both cameras have the same mount, sensor size, and that the same lens is being used to test both cameras.  F-stop, subject distance, and lighting conditions are fixed.  The test scene is a very high DR scene.  Each camera is used at it's respective base ISO value.

I shoot Camera A at a variety of shutter speeds until I find the one with highlight detail that is barely recoverable, just at the point of clipping in the RAW file.  This happens to be a 1/4s exposure.

I shoot Camera B at a variety of shutter speeds until I find the one with highlight detail that is barely recoverable, just at the point of clipping in the RAW file.  This happens to be a 1/2s exposure.

Now I push both of those RAW files to see which one has more recoverable shadow detail relative to noise.  Is this a "fair" DR test, or does the difference in exposure (Sensor A getting only half as much light as Sensor B) affect the number of stops of sensor DR which will be measured?  Assume that we're not talking about enormous differences in exposure (eg, 1/10s vs 30 min), where I'm sure many other factors come into play.

Whether fair or unfair, it seems a reasonable and practical approach to DR testing since many landscape photographs are made that way: setting base ISO and exposing for the highlights without respect to shutter speed.

A related question: Why would DxOmark find a higher DR for an ISO which is below the base ISO for a camera?  For example, they found that for the Sony RX100, the highest DR is achieved at manufacturer ISO 80.  Yet it seems that ISO 125 is the base ISO for that camera, as evidenced by the fact that the camera will not choose an ISO lower than 125 in auto ISO as well as the RX100 manual explicitly indicating that [JPEG] DR may be less when choosing an ISO less than 125.

--
http://aminsabet.com