User blog comment:LilyheartsLightning/Metacritic Effect/@comment-4121017-20130104012928/@comment-3325589-20130105042048

On the other hand, I think Atlas Shrugged: Part 1 is an excellent example of why audience reactions are not a definitive guide to the quality of a work. As long as we're talking stats, another useful concept to consider here is that of sampling bias and the unrepresentative sample. You point out the disparity between audience and critical reactions, but there's another factor to consider: it did about as badly in box office take as it did with the critics.

There are a few possible explanations for this, but the one that strikes me as most plausible is as follows: everything to do with Ayn Rand is pretty controversial. There is a large base of devoted fans of her work, and also a large group of people who think her work is utterly godawful. Because of this, many audience members who would have given the film negative reviews if they had seen it simply chose not to see it in the first place because they were already sure they wouldn't like it, meaning that a disproportionate number of those who did score it were die-hard Rand fans who were inclined to enjoy any adaptation of her work that was reasonably faithful, even if it wasn't very good by conventional standards. The film got out the faithful but failed to attract much of anyone else, which resulted in its notable failure to turn a profit.

In this respect the average critic (including some less-mainstream conservative/libertarian critics in our sample, since a lot of film critics tend to be left-leaning), who is expected to review on standard filmic merits (plotting, characterization, visuals and sound, etc.) regardless of their own opinions on Objectivism, is more likely to be representative of the average audience member than the skewed sample of potential audience that actually went to see Atlas Shrugged: Part 1.