CR's ratings are far from consistant. I can point out numerous examples where they predict a 100% brand-new car will be just as reliable as its predecessor only because it shares the same nameplate and then turn around with another 100% new vehicle and say its "too early to tell," then slaps it with a mediocre rating based on the vehicle it replaces, though they have nothing in common. In addition, CR - for all its claims of scientific valuation - puts way too much emphasis on subjective fields, like seat comfort.
One test I recall often for unfairness is a fullsize sedan comparison between a marginally-equipped Grand Marquis LS, LeSabre Custom, Bonneville SE, and old Avalon XLS. Avalon won (of course), but mainly because of standard equipment and ride quality. CR doesn't bother to take into account a comperably-equipped LeSabre Limited with Gran Touring or Bonneville SLE with sport package not only solve the ride and handling issues, but cost the same as an Avalon XLS. CR also knocked the Buick and Pontiac for trunk hinges, yet did not for the Avalon.
How many people subscribe to CR? Of those, how many people return the evaluation cards? Of those, how many are accurate? What is your opinion of road noise? What is your opinion of material texture? And furthermore, the sample pool is very limited. For example, how many Infiniti Q45s were sold in the past xx years? How many are owned by CR subscribers? How many are owned by CR subscribers who returned those surveys? The real number is likely unscientifically low.