(1) are not always pretty which ones "the formation of judgments, the expression of preferences, the mechanics of taste" because they are always perpetually elusive dynamic === So must be Useless

(2) If someone review something essentially with talk—the aggregate level through sheer number ===Must be fastidious

(3) The result is achieved by the outliers marginalized, and statistical ==== so must be consensus

Can someone explain the answer please ?

This is a crazy tough one and really a good question for the GRE

The trove of data generated from online reviews may offer quantitative insight into a perpetually elusive dynamic: the formation of judgments, the expression of preferences, the mechanics of taste.

Basically this sentence tells us that the data might give us insights and clues in what apparently is clueless and then we do have some example in which it applies


The results, filled with subtle biases and conformity effects, are not always pretty.

here we do have a shift: the results which could be biased NOT always though are good or useful


While anyone review is essentially (i) ____ — the low transaction cost tags it with the cheap talk — the aggregate level is where, through sheer numbers, the noise can be (ii)___, the outliers marginalized, and statistical (iii) ____ achieved.


This is the key sentence. Notice the initial word while which preludes to us a contrast.

We said before that the date reviews often are biased so what this first part of the sentence tells us is that the reviews essentially are useless. They do not help us any either

However, when these reviews are COMBINED through the glass of clear data (sheer) their noise i.e. their role - becomes useful. As such, we have a good filter from the noise and the biased data to come up with a much more clear data and reviews - see the examples where the outlines are put in a corner and these do not create interference - a consensus (which means a vast agreement among the parts=) might be achieved.

hope this helps

Hi couldn't understand the explanation. I chose useless for blank i and amplified for blank ii. my reason- as it mentions that anyone review is essentially useless, when its aggregated, through sheer(large) number it amplifies the noise and hence, statistical extreme is achieved. I might be wrong but this is what i fathomed from the sentence.

Please Sir, read my explanation above. What you inferred from the sentence is a bit out of balance

Noise in this sentence, i guess, is considered a disturbance or something unwanted that should be removed or reduced. So, i think so it should be filtered for the second blank. I think so a technical definition of noise is considered in this sentence.