Why do Asian establishments seem to have lower Google reviews than in the West?

[webny99]

Someone who gives mostly 5s will have each individual 5 count less ...

This devalues the reviews of people who only choose to post reviews when the establishment truly did far exceed all expectations.

However, I think part of what this thread might be pointing up is that such might already be the case, at least here in the US. A lot of people seem to give 5-star reviews for what might actually be just an average experience.

[kphoger]

Imagine I'm a business traveler and eat out at hundreds of restaurants every year.  I only ever write an online review if the service was stupendous, far exceeding the expectations that I've gained over years of eating out.  I never leave scathing reviews, because I understand that everyone has an off day, and I also know how bad for business a single bad review can be.  So, over the last four years, I've reviewed seven restaurants–all of them at least 4.5 stars, and five of them were 5-star reviews.

Now imagine instead that I'm an average joe who never really ventures out of my home city.  I write an online review of practically every place I go.  Over the last four years, I've reviewed 31 restaurants just in my home city.  Because I'm kind of a prick, 22 of those were 1-star reviews.  Five of them, on the other hand, were 5-star reviews.

Under |1|'s proposed system, the 5-star reviews in the first case would be docked by a greater factor than in the second case.

[hotdogPi]

Imagine I'm a business traveler and eat out at hundreds of restaurants every year.  I only ever write an online review if the service was stupendous, far exceeding the expectations that I've gained over years of eating out.  I never leave scathing reviews, because I understand that everyone has an off day, and I also know how bad for business a single bad review can be.  So, over the last four years, I've reviewed seven restaurants–all of them at least 4.5 stars, and five of them were 5-star reviews.

Now imagine instead that I'm an average joe who never really ventures out of my home city.  I write an online review of practically every place I go.  Over the last four years, I've reviewed 31 restaurants just in my home city.  Because I'm kind of a prick, 22 of those were 1-star reviews.  Five of them, on the other hand, were 5-star reviews.

Under |1|'s proposed system, the 5-star reviews in the first case would be docked by a greater factor than in the second case.

The standard deviation is much less in the first one than in the second, so it's not by as much as you would think. There is, however, the problem that it would see the two 4.5s as low, since they're below the mean.

In your second example, your 22 1-star reviews would be devalued for both the mean being low and the standard deviation being high, which is a good thing because most of those restaurants probably don't deserve 1 star.

[formulanone]

Imagine I'm a business traveler and eat out at hundreds of restaurants every year.  I only ever write an online review if the service was stupendous, far exceeding the expectations that I've gained over years of eating out.  I never leave scathing reviews, because I understand that everyone has an off day, and I also know how bad for business a single bad review can be.  So, over the last four years, I've reviewed seven restaurants–all of them at least 4.5 stars, and five of them were 5-star reviews.

Now imagine instead that I'm an average joe who never really ventures out of my home city.  I write an online review of practically every place I go.  Over the last four years, I've reviewed 31 restaurants just in my home city.  Because I'm kind of a prick, 22 of those were 1-star reviews.  Five of them, on the other hand, were 5-star reviews.

Under |1|'s proposed system, the 5-star reviews in the first case would be docked by a greater factor than in the second case.

The standard deviation is much less in the first one than in the second, so it's not by as much as you would think. There is, however, the problem that it would see the two 4.5s as low, since they're below the mean.

In your second example, your 22 1-star reviews would be devalued for both the mean being low and the standard deviation being high, which is a good thing because most of those restaurants probably don't deserve 1 star.

Or instead of trying some complicated formula that doesn't promote honest reviews, just dock them by the number of times they've used "rude" in their descriptions.

(Hint: maybe by the third or fourth place one has been treated rudely, it's not the establishment's attitude that needs adjustment...)

In any case, Google ratings are not a good diagnostic tool. Like much of the survey system, it's geared to make people ignore flaws and promote dishonesty by omission.