Norm --
one more thing...
so, full disclosure here, I've bought two of my three bikes without meaningful test-rides. (The first one, I wouldn't have known the difference, the second I did choose through test rides and I think I chose well, and the new one, well, that was a lark and I think it's going to work out, but I knew a fair amount about what I was looking for...).

BUT --- you learn so much about bikes, and about your preferences, from test rides. It really makes a HUGE difference. My boyfriend bought a new bike last year, and we probably demo'd at least 10 bikes in the processs. Roubaix, Cannondale System Six and Six13, Cervelo R3, Trek Madone, Trek 5200, others... And my take-away lesson is, that while fit and on-the-page geometry ARE important, there's something elusive the actual riding experience that cannot be gleaned from geometry charts. Our mutually favorite bike, the Cervelo R3, we ultimately decided against because it was so much more expensive than the Madone for the same components, and the pleasure of riding it fell into the frosting category, not the "really makes a difference for what we'll be doing" category -- but it was truly special. The Cannondale System Six was a very nice and expensive bike -- but I thought the handling was sluggish and inferior to the less expensive Six13, and the acceleration wasn't as macho either. The Trek 5200 we looked at was spec'd fairly similarly, component-wise, to the 2007 (pre-compact geometry) Madone 5.2 he bought -- but the carbon was very different -- no way to understand this without riding.

So I think she really should be involved in the test-riding process. Partly for fit reasons, but also because ride feel and handling characteristics vary widely between bikes in the same price bracket, and preferences regarding these variables can be personal -- they can fall into the "is chocolate better or vanilla?" category.

That's my rant for now. I totally support what you're doing though.