15 Comments

I'm another person who seeks out well written negative reviews when buying products, but I think there might a hidden irrationality there as well,- that of believing every story must have two sides and that nothing could be that perfect. Reading a trustworthy negative review dispels the niggling feeling that there must be some hidden flaw in the product, since unlike a well written positive review they aren't trying to justify their purchase. So it's a kind of "better the devil you know" phenomenon, only even more absurd than usual.

Expand full comment

Negative reviews are very useful (in helping me make a favorable choice) if the review is detailed but picky on dimensions I don't care about.

For example, someone who isn't much of an ethnic foodie might go to a restaurant which said:

"Elegant presentation, ok prices, and good service, but shockingly banal food. I thought the attempt to serve spicy kidneys on toast with ice cream was a flawed afterthought but I suppose it's ok if you just want a good hamburger and a somewhat pretentious milkshake. Don't try to order wine."

Or conversely, I've seen reviews of very good but unusual Chinese restaurants panned by people looking for moo goo gai pan, chow mein, or sweet & sour pork who aren't happy. In those cases, the negatives are clear positives.

Expand full comment

if I am looking at a negative review of a resort and it says it is has to many kids running around it I might like that if I have kids myself.

Expand full comment

A prior is a piece of information used to make deductions or Bayesian inferences from. They are the pieces of data that a Bayesian analysis operates on to produce a conclusion or posterior, the product of the Bayesian inference. Priors could also be considered the “facts” that are used to draw a conclusion from.

If two individuals had the same priors and applied an honest analysis, then they would reach the same conclusions. This is the way that arithmetic works. If you have the same numbers and apply the same arithmetic operations to them, you get the same answer.

If two individuals with the same priors don't agree, it can't be because they are applying an honest analysis to their priors. If two honest individuals with different priors communicate, eventually they will have the same priors because they will keep communicating until their priors coincide. If the priors don't eventually coincide, then one of them is not being honest and rational.

There are a number of people I have essentially no disagreements with, as far as I know. I suspect that is because I hang out with skeptics and the skeptic mindset is to try and be meta-rational. A skeptic can only argue from facts using logic. A skeptic has to have priors that are based in facts and needs to be able to articulate them and then argue with them using logic. If you do that, then you are a skeptic, and meta-rational. Robin's observation that meta-rational people are very rare must be due to the people he is mostly hanging out with.

It has been said that reality has a liberal bias. I think that is correct, and the essence of conservatism is to adopt the priors that have always been adopted whether they are right or wrong and never change them. It is my perception that the major disagreements in politics are dishonest.

Expand full comment

I briefly skimmed the executive summary but didn't read the full article. I'm low status, I couldn't read it! I was reading a foreign language!

What is a prior? Is it the same concept as tracking or is it more like batting average?

Expand full comment

Tyler Cowen & Robin Hanson built off Aumann's result in Are Disagreements Honest?

Hanson also says that Uncommon Priors Require Origin Disputes

Expand full comment

The presence of any negative reviews likely means that the reviews are honest and not false advertising by the hotel.

Well written negative reviews likely mean that the person writing them expected their negative review to have some effect, or they wouldn't have bothered.

On the other hand, negative reviews could also be false advertising by the hotel's competitors. The use of negative reviews by ideologues opposed to a certain point of view is extremely common.

For example anti-vaxers trashing Paul Offits book.

http://www.amazon.com/Vacci...

Global warming deniers trashing Al Gore's book.

http://www.amazon.com/Incon...

A strongly bimodal distribution of reviews shows bias on the part of at least one group of reviewers. I recently read Robert J. Aumann's paper “Agreeing to disagree”. Two individuals can't have the same priors as common knowledge and then not agree on posteriors. At least one of them has to be biased, ignorant or dishonest. .

Expand full comment

Like paul said - I'd discount a review that was well writen and entirely positive. The thing most likely to make me by is a negitive (harsh) review that pointed out all the things I dont care about.

Expand full comment

It's axiomatic "chicks dig jerks" and maybe the lack of niceness is a better indicator than men are willing to admit.

I often wonder is it more Christian to be nice, or brutally honest. I find that beneath niceness is often neediness caused by overvaluing some thing or outcome.

If overvaluing earthly things is not idolatry then I don't know what is.

Expand full comment

Also a "hard to please" high status person is superior to an "easy to please" high status person. A negative review suggests the reviewer experienced superior service elsewhere, meaning his status is high enough to warrant superior service.

Expand full comment

demand for a hotel increases if the reviews on TripAdvisor and Travelocity are well-written, without spelling errors; this holds no matter if the review is positive or negative.

That they are reviewing the hotel is strong evidence that they chose the hotel in the first place. That they chose the hotel is evidence that the hotel is good, and the more competent they are the stronger that evidence is. That they write well and don't commit spelling errors is evidence that they are competent. Moreover, that they write well and don't commit spelling errors is also evidence that they have high standards. A negative review by someone with high standards is weaker evidence that the product is bad than a negative review by someone with low standards.

But if you thought that a negative review makes you like a product less, well then you are more deluded than you realize.

Statistical trends displayed by large numbers of people do not automatically extend to an arbitrary "you" selected by the filter of "do you read this blog".

Expand full comment

I'm not completely sure about this one, Dr. Hanson. As reviews exist mainly to inform about the product, a well-written review probably informs more effectively than a badly-written one. So if "good reviews" lead to less uncertainty, and less uncertainty leads to more purchases, then high-quality reviews can increase sales (without status-associations).

Moreover, I personally focus mostly on negative reviews (a sort of "maximin" "worst-case scenario" focus), and when I find a negative review that I trust it frequently makes me more likely to purchase the product (if I decide that my downside is capped at some acceptable level).

Possibly I "associate" with the review-author in some way ("Those other reviewers were clearly to incompetent to use the product the right way/to seek help effectively, etc, but this articulate reviewer is more like me.") but I am really not sure.

Expand full comment

('Is', not 'it.)

Expand full comment

"But if you thought that a negative review makes you like a product less, well then you are more deluded than you realize." This may apply to many other people, but *not to me*: I am too sophisticated and self-aware for that. And I'll bet Robin Hanson it, too!

Expand full comment

So if I receive bad service at a hotel, I should post a poorly written positive review?

Expand full comment