Statistics quiz!

[Super Macho Man]

I'm posting this because I'm astonished at how many otherwise smart people I'm seeing on the forums make a very, very basic statistical mistake - one of the first things one learns in any stats class - and I'm interested in seeing the outcome.

Answer according to your "gut feeling," don't overthink it - answer the way you would instantaneously feel if it really happened. Here's the scenario...

Your Brand X computer, the sixth Brand X computer you've owned in the past 10 years, is starting to act up. Your history with Brand X computers has been nothing but awful. Multiple failures on every one. But they're always so inexpensive, you just can't pass them up. Consumer Reports says there is nothing awry with Brand X - its reliability ratings are fine.

Meanwhile you have a friend who just bought a Brand Y computer. It's his sixth Brand Y computer in the past 10 years, and it's been nothing but trouble-free for him so far. In fact, every Brand Y computer he's ever owned has been wonderful to him! You checked Consumer Reports and there is nothing special about Brand Y - its reliability ratings are right in the middle - in fact a perfect tie with what Brand X's are.

You're so pissed at your Brand X that you think it's time to buy a new computer. You have store credit at Mom N Pop's Computer Warehouse, so you stop in and there's a Brand X computer and a Brand Y computer sitting on the shelf right next to each other. These two are all they have in stock. Their specs are both identical, similar styling, same price. If you were to buy one, which one do you think would end up being more reliable for you? Again... gut feeling.

Brand X
Brand Y

 

[Super Macho Man]

Answer!

Don't read below until you've read the quiz question and answered it........
..
..
..
..
..
..
..
..

If you answered Brand X, your intuition would be...
....
...
..
..
.
..
..
..
..
..
..
..
.
mistaken.

Both computers have a virtually equal chance of being reliable for you. Discuss.

 

[Killyp]

Brand Y, because reviews are completely irrelivant. They state someone else's opinion, not your own...

And I would much more rather prefer a Machine which is average, but doesn't give me loads of problems, than a Machine which gives me loads of problems, but gets good reviews...

 

[someguy]

Discuss.

Simple.

For me, "Brand X" played the role of Windows-based PC's. I've used them all my life and always had one problem or another with them (which is why I'm as tech savvy as I am). I've owned at least 6 or 7 "Brand X" computers in my life.

Then, a friend of mine was going on about how he doesn't understand why anyone would use anything buy "Brand Y" (Mac in this case).

I switched and haven't looked back since. Definitely the right move for me, regardless of whether or not the decision was supported by statistics.

 

[miloblithe]

IT doesn't matter. The reactions to computer X and Y are based on the personality of the consumer, in this scenario; there is no difference between them. If the "I" switches to Brand Y, he'll be just as likely to complain as the friend would to not complain if he switched to brand X.

 

[Super Macho Man]

Brand Y, because reviews are completely irrelivant. They state someone else's opinion, not your own...

Well, I used Consumer Reports because it is a survey of a large sample of consumers who have reported their reactions, whether positive, negative, or neutral. It doesn't take into account individual reviews or reactions, which is an important point. I don't know if you knew that or not (or maybe I have the wrong impression of CR), but it might change the way you're thinking about the situtation...

 

[Silentwave]

I'd buy brand Y just to spite the gremlins that inhabit Brand X's computers, which are always eating away at the cables and hitting the Hard Drives to mess up my computers!

 

[Dros]

The reviews are an aggregate. I would hypothesize that the failures in brand x are due to a wiring defect in the house (something not experienced by many people in the aggregate review) and would try brand y to see if their components are more robust in that aspect.

 

[mkrishnan]

Your poll would've been better if there was actually a technically correct option for which one of your voters could vote. It's easy to say that Americans don't understand statistics when you only offer them incorrect options and then scoff at the fact that they chose incorrectly!

There's also another problem inherent in this. Your basic point that people tend to overweight less powerful anecdotal evidence (based on small sample sizes) and underweight more powerful empirical evidence (based on large sample sizes) is correct.

HOWEVER.

However, in order to rigorously conclude that the Consumer Reports data is more reliable than the anecdotal data, one must add the assumption that defect rate is not related to usage profile. Which is a fair assumption sometimes. But probably not a fair assumption in computers. If you and your friend have vastly different usage profiles, there is no data here to negate the claim that it is your usage profile and not your brand that lead to the failures. Nor is there any data to negate the potentially compelling claim that usage profile and brand interact.

You're still left with the unavailable best answer -- that there is no clear evidence that either brand is a better pick than the other in terms of reliability.

But that is not the same as concluding that the anecdotal evidence in this scenario is not potentially powerful, and not potentially more powerful than the statistical data.

Again... if your vested interest is in proving how smart you are and how dumb others are, you can certainly create a "test" that appears to prove this. But whether it actually does so...

 

[Super Macho Man]

IT doesn't matter. The reactions to computer X and Y are based on the personality of the consumer, in this scenario; there is no difference between them. If the "I" switches to Brand Y, he'll be just as likely to complain as the friend would to not complain if he switched to brand X.

You don't mean to tell Poor Brand X Schmuck that those six consecutive failures are all in his head, and vice versa, do you?

 

[KingYaba]

I went with my gut and clicked on option Y. Barely read the post too

 

[Super Macho Man]

Your poll would've been better if there was actually a technically correct option for which one of your voters could vote. It's easy to say that Americans don't understand statistics when you only offer them incorrect options and then scoff at the fact that they chose incorrectly!

There's also another problem inherent in this. Your basic point that people tend to overweight less powerful anecdotal evidence (based on small sample sizes) and underweight more powerful empirical evidence (based on large sample sizes) is correct.

HOWEVER.

However, in order to rigorously conclude that the Consumer Reports data is more reliable than the anecdotal data, one must add the assumption that defect rate is not related to usage profile. Which is a fair assumption sometimes. But probably not a fair assumption in computers. If you and your friend have vastly different usage profiles, there is no data here to negate the claim that it is your usage profile and not your brand that lead to the failures. Nor is there any data to negate the potentially compelling claim that usage profile and brand interact.

Thank you for your input, Professor Nitpick. Of course, there are a lot of weaknesses with the quiz. I guess I didn't spell it out in fifteen pages of detail because I meant for it to be a simple little thing intended to be thought about and not intricately picked apart under a microscope. It will not be showing up in the Journal of Cognitive Psychology anytime soon, or ever, but that's ok.

My point is seeing so many posts along the lines of

"I'll never buy Brand ___, I've had bad experiences with it"
"Buy Brand ___, I have something of theirs and I love it!"
"Brand ___ hasn't failed me yet"
"Don't buy Brand ___, it's crap"

 

[mkrishnan]

My point is seeing so many posts along the lines of

Like I said...I totally agree with that point. But then we all know that Dell is Brand X!

 

[Loge]

I would buy Brand Y. I take the stats into account, but I also take into account my personal experiences. If nothing else, I would get the satisfaction of moving away from a company that has given nothing but trouble.

 

[Super Macho Man]

Like I said...I totally agree with that point. But then we all know that Dell is Brand X!

Ok... yes, I admit it, Dell is brand X.

Don't buy Dell, it REALLY IS crap!

 

[ejb190]

The hidden truth - the friend was a IT specialist by profession...

 

[63dot]

Brand Y, because reviews are completely irrelivant. They state someone else's opinion, not your own...

And I would much more rather prefer a Machine which is average, but doesn't give me loads of problems, than a Machine which gives me loads of problems, but gets good reviews...

i agree

and reviews look at brand new machines when some problems don't show up...but within a month, as a computer tech for 7 years, i have seen almost all problems present themselves up front

it's rare, actually very rare, to see problems come up from the first year to about the third year...but after that, stuff just wears out

 

[crazycat]

If a brand is giving me lots of problems i would dump it so fast that i would be on my way to a store to by brand y.