# Thread: How to rationalize mainstreams obsession with statistics?

1. Originally Posted by Fluffywolf
These days, you can't turn your head or you'll hear it on the radio, see it on tv, or read about it in the newspaper.

Statistics.
Numbers are easy to understand. Takes away a lot of the complexity in decision-making for the consumer -- and that's what companies want.

When people read that a generic cologne is chemically indistinguishable from its overpriced, name brand counterpart - so much so that '75%% of consumers can't tell the difference!' - they are banking that you'll want their cheaper product instead. Or that the \$39.99 purse you see at a mid-tier clothier has '99%' similarity to the \$500 Coach alternative, they anticipate that the comparative breakdown will drive home the point that it just doesn't make sense to buy high.

It's a one-two punch. Appeal to what you want by showing how much you save.

2. Originally Posted by Fluffywolf
So what is it about statistics that peaks peoples interests?
Statistical effects:

interest is peaked in 15% of the subjects
interest in piqued in 40% of the subjects
25% aren't paying attention anyway
20% call BS
15% of the subjects are non-existent and probably fabricated, you'll often find this if you really analyze the numbers

Admit it. You read this post because it appeared to contain an objective answer to a subjective question.

3. I have to agree with Night that one of the reasons the general public likes stats is that it's a form of short-hand. Instead of having to work with multiple parameters, you get it condensed into one (or a few) simple number. Then, you can use a simple "good if it's above, bad if it's below" sort of decision. Example, the movie got a score above 9.0 on Rotten Tomatoes, so I must go watch it.

The problem arises of course when people don't realize that summarizing stuff into a single number causes a loss of information and accuracy, or when they take that statistic as an undeniable truth and ignore all other pieces of information.

Statistics can also be used as a common language for discussion, especially in sports. Taking the baseball example, two fans can quickly strike up a conversation if they're both speaking the same "stats".

Then, there's the human need to rank and compare stuff, and statistics is also good for that. Gives a stronger form of legitimacy because saying pitcher X is good because he has a WAR of 5.0 sounds more factual than "he looks like he knows how to get hitters out and help his team win"

4. I'm turning upside down today..

5. I'm in a Stats class right now and I'm pretty convinced that the stats themselves don't lie. But, there are a few problems:
1. Type I error, which is where a researcher tries to make the research support his/her hypothesis. This is immoral, and is not uncommon because people are biased and we don't like being wrong
2. Stupid/unethical reporters/journalists/people-in-general miss-interpreting statistics. I think this happens A LOT. In fact, I would say this is where MOST statistics research goes wrong; because a journalist opens a peer review journal and reads a correlational study and is too dim or sensation seeking to realize that, as my professors have driven into my head, CORRELATION DOES NOT MEAN CAUSATION, or they don't consider the "third variable problem," and then of course they're pulling an study from a peer review so that means it probably hasn't yet had time to provide sufficient evidence of re-testablity.
3. People don't know what questions to ask. Let's say that a study says "there is a statistically significant increase in the possibility of contracting rabies in people who travel cross country versus people who do not." And people think, 'oh, I shouldn't take that trip to Yosemite'. When instead they should be asking, what was the original likelihood? Because, just maybe, the original likelihood (of the non-traveling group) was 1 in 100,000 and a statistically significant increase, (we'll say 2.5% for kicks and giggles) now makes the likelihood 1 in 97,500. Or, another important question would be "what's the variant?" Or "was it a unimodal study?" These are BIG questions that entirely change the outcomes, and how people should respond, and, I don't see them asked anywhere.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
Single Sign On provided by vBSSO