Step One:

The qualitative exclusion of 94 sites was accomplished with this original, generalized rubric:

Any site that showed one or more of the following would be crossed off our list:
-Surface at site is largely hardened lava, making anything fluvial difficult to get to.
-Terrain is unusually rocky or relieved, making landing or exploration dangerous.

Any site worth investigating would show evidence of ancient hydrological activity.

Any site that showed hydrological activity AND showed one or more of the following would be kept on the list:
-A surface albedo indicating that water-deposited sediments are present.
-Curved flow channels are present.
-Ejecta from nearby craters are visible.
-Sediment layers are exposed in nearby cliffs or canyon walls.
-Flat terrain appears safe to land on.
-Ancient rocks there date from the Hesperian or Noachian ages.

Step Two:

Mapping the remaining sites on the surface gave the following two pictures. Images courtesy of the Center for Mars Exploration Landing Site Catalog, modified to show only the sites remaining after Step One.

After circling sites that appeared to be close, we had 11 clusters of sites that offered good potential. This was before the rover range was shortened.

Steps Three and Four:

At this point, we made a large chart listing the characteristics that each site had. We worked using our cluster method, but when we had to abandon the cluster idea, our chart was still useful.

Step Five:

Step Five was to take our diversity chart and eliminate all sites that did not have four or more diversity checks. We kept Gusev because we knew how much information was available about that site. Although Gusev may not be as diverse as other sites, it began to look like a safe back-up at this point.

Step Six:

The "Pro-vs.-Con" charts that we used to order our thoughts about the six best sites appear below.