Robert McCrum and his literati-colleague/acquaintances at the Grauniad demonstrated that whatever the center-right literati affiliated with the Times (either one) can fuck up, they can fuck up too. And then draw equally invalid conclusions from their mistakes. Today's Guardian/Observer includes the results of a poll of "150 literary luminaries" asking them to name "the best novel (in English, excluding America) for the years 1980–2005," allowing that "how you define 'best' is up to you." I find it hard to think of a more-useless exercise in name-dropping.
To begin with, how can one possibly expect to get a coherent list by asking for the "best" without any guidance whatsoever? Looking at the list of "winners" and "other nominees", one finds some truly peculiar entries that were almost certainly encouraged by that (absence of) basis. For example, someone nominated Harry Potter and the Half-Blood Prince; to anyone who has actually read the Harry Potter books with any critical acumen, if one volume (and only one) was to be nominated one would expect it to be Harry Potter and the Prisoner of Azkaban (which can, with some difficulty, stand alone... while HBP cannot). Many of the nominees are thought by even the ardent advocates of those authors to be among the weaker books by that author; nominations for The Handmaid's Tale1 in place of the vastly more-accomplished The Blind Assassin, or two different novels by Angela Carter instead of her masterful collections of short fiction, or Earthly Powers over virtually any of Burgess's other "eligible" novels (not to mention that Earthly Powers falls outside the period, because it was first published in 1979!), or anything at all by "John le Carré," reflect a certain mob mentality and absence of awareness in the survey population.
But that's not as much fun to criticize as is the very survey method itself. After all, a substantial part of the flaw in asking for a single "best" work from voluntary responses can be statistically ameliorated by asking for multiple nominations of three or five best, whether or not ranked within that group. Instead, one must ask why this particular population of literati has any more credibility than does any random sampling of citizens. The key question is this: Just how widely read are these literati? For example, only four of them write speculative fiction, one of whom writes "childrens' books," while two others have archly abandoned speculative fiction; a disturbingly high proportion of the "panel" consists of novelists first published after 1998; several of the "panelists" are essentially celebrity-authors who have written highly popular utter crap; and so on.
And people wonder why the publishing industry is so screwed up.