# How to compare product ratings

Discussion in 'Physics & Math' started by Jennifer Murphy, Mar 30, 2019.

1. ### Jennifer MurphyRegistered Senior Member

Messages:
231
Is there a statistical tool that I can use to compare ratings, such as from Amazon, for similar products from different manufacturers?

Suppose I find three similar products with these ratings and the number of reviews that led to those ratings.
Code:
Product Rating Reviews
A     4.9      3
B     4.7     48
C     4.7    628

How much confidence can I have that Product A is really better than B and, especially, C? Intuitively, the rating for C would seem to be more reliable than the one for B, which would seem to be more reliable than the one for A.

Is there a statistical tool that will allow me to compute the "most likely" rating for these three products?

Is there a tool that will allow me to calculate the probability that the "true" rating for A is higher than that for B or C?

Is there some other tool that I don't know to ask about that will help me compare these ratings realistically?

3. ### RainbowSingularityValued Senior Member

Messages:
3,542
none

logarithmic algorithms do such things(to compute values on the internet accross platforms & websites of various data)

however(if you are just going to use 1 single site with 1 single set of data all using the same known value & statistical base lines), it is fairly easy(not very sexy for most people which turns them off) to do in long hand manual simple mathematics
i have met people who can do this type of thing in their head as you give them the data.

based on 'what' as a "truth" ?

emotional value
= subjective

5. ### Jennifer MurphyRegistered Senior Member

Messages:
231
Hmmm... That is exactly the value I place on your reply. 100% substance free. I guess I came to the wrong place. I thought this was the "Science" forum. I guess I went to the "Adolescent Sarcasm" forum by mistake.

7. ### RainbowSingularityValued Senior Member

Messages:
3,542
lol
your overt emotional response to 1 or 2 words that just happen to be triggers words is of no use to your desire to apply statistics.

and ...
"please give me plastic wrapped single use commercialized answers that i can stick into any slot"
"pass me the hammer i want to adjust something technical"

... you cant see the forest for the trees

8. ### QuarkHeadRemedial Math StudentValued Senior Member

Messages:
1,608
Jennifer, ignore Rainbow Singularity.

In statistics, sample size is the key to assessing confidence. So the larger the sample, the greater the confidence you can have in the statistic.

So, in your example, set C is by far the most reliable.

As for tools, there are many, but all require a knowledge of statistical math

Jennifer Murphy and sculptor like this.
9. ### RainbowSingularityValued Senior Member

Messages:
3,542
how do you know that when you do not know if the question is being asked to purchase an item
or
the question is being asked to formulate a statistics student essay question
or
if the question is being asked to evaluate quality of cross web-site ratings systems

it seems obvious to me that the ratings systems could be 3rd party hacks/incorrect or slanted
using web-site data to then give back a result of the star rating.

if you just want to have confidence in buying a product
the complex statistics of the over all question is just a long walk around the city to get to your front door step.
99% of all people will take the shortest route and then complain about being given the info to make up their own mind.

the quality of the product has no bearing what so ever to the ratings of the comments section
to apply a product value to the ratings values would be a 3rd or 4th step process and be wholly contingent on the quality and precise nature and origin of the previous 3 sets of data.

you could pretend its all very easy and obvious but by default that would be calling jeniffer an idiot.
which i am not doing.

all on amazon ?
only using amazon star ratings ?

all on amazon ?
comparing product star ratings of different products on amazon ?

you have a question about amazon star rating data accuracy to product value ?
product value to what ends ?
customer satisfaction as a non scientific not time rated 1st purchase only ?
ease of purchase ?
site use likability converted to star rating assigned to non real product value ?
10 year product quality to amazon star rating system accuracy value ?
default average responses outside the bell curve ?

that collects and quantifys for accuracy amazon star rating data to feed back quality response data value ratings that are then compared against the 10 year quality reports for products from different suppliers ?

Last edited: Mar 31, 2019
10. ### Michael 345New year. PRESENT is 69 years oldValued Senior Member

Messages:
7,905
Ouch

No you did get a SCIENCE forum with a penchant for religion

What I suspect you really really want and I will tell you what you really really want is PRODUCT REVIEW

Two blocks down, turn left, can't miss it

RainbowSingularity likes this.
11. ### RainbowSingularityValued Senior Member

Messages:
3,542
amazon employee phishing for someone else to do your work for you ?

12. ### QuarkHeadRemedial Math StudentValued Senior Member

Messages:
1,608
Why the use to which it is put affects the reliability of a statistic, as Rainbow Singularity seems to believe, is a complete mystery to me.

13. ### Jennifer MurphyRegistered Senior Member

Messages:
231
Emotional response? Pot? Kettle? Next I expect you to ask me if it's the wrong time of the month.

Please point out even the tiniest scintilla of actual statistical information in your response?

14. ### Jennifer MurphyRegistered Senior Member

Messages:
231
That's the plan.

That much I knew

That's why I came here. I was hoping that someone could show me how to compare these samples statistically. Can you?

15. ### sideshowbobSorry, wrong number.Valued Senior Member

Messages:
5,683
Science is like gold. It takes some effort to get it out. You can't expect us to serve it to you on a silver platter.

16. ### Jennifer MurphyRegistered Senior Member

Messages:
231
(sigh) I do expect relevant responses, especially on a science forum. Obviously, you do not know the answer. The juvenile trolls have taken over here. I was wrong. You and rainbutt would actually have to grow up to make it to "adolescence sarcasm". You are a few grades short now.

PS: Your subtitle is spot on.

17. ### Michael 345New year. PRESENT is 69 years oldValued Senior Member

Messages:
7,905
OK
From the reviews it appears
3 people reviewed can opener A
for average rating of 4.9
48 people can opener B
average rating 4.7
628 people can opener C
average rating 4.7

Using A and C

Suppose the average rating for A (4.9) is correct (few reasons why not, but stick with true)
628 reviewed WOULD THE RATING
GO UP? - DOWN? - STAY SAME?

Does not compute Will Robinson

Can opener A cost $1,092 Can opener C cost$48

A buyers expected it to perform better
C buyers happy if lasted for the camping trip

Go check out the items

Use it not other people's reviews you don't appear to trust anyway

Last edited: Mar 31, 2019
18. ### RainbowSingularityValued Senior Member

Messages:
3,542
appeal to maturity to try and get someone else to do your school work for you ?
lol

you still have not answered my original question requesting you define the nature to which your seek so as to remove vast quantities of required statistical analysis.

i wonder why that is ?
because of ... "emotional response" ?

wood for the trees or deliberate ignorance to avoid being caught out trying to ask others to do your school work for you ?

appeal to maturity...
do you think i did not notice you trying to curtail perceived responses to give you exactly what you wanted instead of help.
that's a bit "i want i want"
rather than "can i have some help doing something please ?"

you are asking someone else to not only give you the result but also do all the thinking and math for you.

meanwhile, you have not stated why you want it, which clearly defines you as potential amazon employee trying to get someone else to make them money for nothing.

"appeal to maturity of emotion?"

you posted under physics and maths.

19. ### SeattleValued Senior Member

Messages:
4,624
You don't need any statistical tools for this example. Most people rate it 4.7. Is there really much difference (for you) between a 4.7 rating and a 4.9 rating?

To need statistical tools you generally need more complex problems that you can't just look at and process without those tools.

When you do have more complex problems, you need to understand statistics before the "tools" are of any use to you.

20. ### RainbowSingularityValued Senior Member

Messages:
3,542
GEEKWIRE
Amazon changes its key formula for calculating product ratings and displaying reviews

by Todd Bishop on June 20, 2015 at 10:08 am

NBC NEWS
https://www.nbcnews.com/better/busi...ne-review-really-mean-product-good-ncna870901
Online reviews: Here's what's behind all those 5 star ratings
Research suggests the average person reads about seven reviews before making a decision. Sort through the swamp of ratings and reviews with these seven easy tips.

May 8, 2018, 9:14 AM GMT+12
By Stephanie Thurrott

21. ### RainbowSingularityValued Senior Member

Messages:
3,542
Here's Why You Shouldn't Trust Amazon Reviews
Mike Prospero ·
Senior Editor, Tom's Guide
Updated Sep 17, 2018

https://www.tomsguide.com/us/dont-trust-amazon-ratings,news-27936.html

where does it say on the product reviews how the product has performed over the last 3 years after it was purchased ?

...
?
?
Realistically you need to find product review non profit orgs and private citizens who do their own product reviews.
you then need to cross read with consumer organizations who study long term consumers and how the products perform over several years.

ignore all the star rating BS because its like some type of facebook-for young teens meets hunger games teenage popularity contest.
look for "product review" web sites and organizations with content that talks about the products performance over several years.

Last edited: Apr 1, 2019
22. ### sideshowbobSorry, wrong number.Valued Senior Member

Messages:
5,683
You might get some if you're patient. You are also likely to get some reality checks.

23. ### QuarkHeadRemedial Math StudentValued Senior Member

Messages:
1,608
She won't get any more from me.

She has been dismissive of those that tried to help, and downright discourteous over all.

This is no way to make friends

sideshowbob likes this.