Social Question

nikipedia's avatar

Why are you skeptical of statistics, and what would it take to convince you they have value?

Asked by nikipedia (27669points) November 22nd, 2010

Mark Twain had the line about lies, damned lies, and statistics. I’m sure he’s right that people invent and misuse statistics regularly—but this seems to have generated a sense in the general population that no statistic can be trusted.

Do you mistrust all statistics by their very nature? If so, why? And what would it take to convince you that statistics are vital to understanding causal relationships in this here universe we live in?

Observing members: 0 Composing members: 0

14 Answers

jca's avatar

i do trust statistics if i know the source and if there is no financial benefit for a certain result. For example, I read once that chocolate does not cause acne. That came from the makers of M&M’s.

Simone_De_Beauvoir's avatar

I am not skeptical of statistics but am skeptical of how they are sometimes used out of context or out of the original article and how they’re then presented and perpetuated via the media to the public. I think most laypeople have no background in critically assessing the meaning behind statistics or the methods used to obtain statistics and are therefore unable to either believe them outright or question them. In order for me to be convinced by a statistic, I would need to know exactly what the experiment was about, every point about its design, methodology and how statistically significant the results are. Furthermore, what matters more than a statistic is what researchers are recommending or implying about them. Sometimes, there is a statistic and it gets shared and I read the original article and the research team is a lot less willing (that has been publicly shared by others) to say something outright – they’re a lot more cautions, saying it has limited applicability or something like that and that more research is needed (and often there is oppositional research, NOT being paraded by the media) but people don’t see those cautionary phrases. Sometimes, statistics aldo don’t tell the whole story (but it’s up to all of us to not assume so, regardless) – two different people can use the same number and lead a person to believe two different conclusions. I am also aware of scientists who crunch their numbers a certain way to get the desired result they want.

ucme's avatar

Oh there’s no doubting some stats have inherent value. If they’ve been thoroughly rersearched then that’s fine. It’s when they get random & pointless I take issue. For instance…...37% of left handed Mexican men say they mount their donkey from the right!! Okay I just made that up, but kind of illustrates my point.

iamthemob's avatar

Basically, what @Simone_De_Beauvoir said. Statistics that are often presented in the media aren’t really statistics, which is the application of mathematical techniques on groups chosen according to relatively strict experimental standards in order to show whether certain variables cause significant differences between those groups, and therefore causation between the variable and the result or at least something suggestive of causation. They are more often than not percentages, or polls. Also, they are often presented without the details necessary for us to make an assessment of whether or not the techniques involved were solid.

Basically, the media is ruining any usefulness statistics have. My skepticism of statistics is more along the lines of the fact that most studies are published by sources in the media I have learned to be skeptical of.

crisw's avatar

“Do you mistrust all statistics by their very nature?”

No, not at all. I love statistics. But I also took stats classes, and have read a lot about statistics, so I have a fair understanding of how they should and should not be used.

I think that a lot of the public misunderstanding of statistics comes from a deep lack of knowledge of how statistics work. People know very little about how to judge whether a given set of data and its analysis is valid or not.

iamthemob's avatar

Exactly. People see a ”%” and think, “Statistics!” I don’t think statistics until I see the terms “standard deviation,” “control sample,” or things of the like.

wundayatta's avatar

I don’t mistrust statistics, but I prefer to know how the study was done. What was the sample size? How was the sample selected? What was the participation rate? What were the questions? How were they tested? etc, etc.

Failing that, I have to decide if I trust the organization that is producing the statistics.

The innumeracy in this country is criminal. I think most people have no idea of what statistics tell us. Certainly, no one ever reports the variance, which I think is an important number that would help those people who say, “but I had this experience and it was different, so the study must be wrong.” There’s a lot of variability in responses, so it doesn’t mean the prediction will be accurate for any one individual.

Summum's avatar

I mistrust most of the stats given over the news media. Because most everything is one sided or the group of people giving the results are ones that are one sided. I just don’t trust our government nor the news media.

rts486's avatar

Statistics don’t lie, but liars use statistics. I had a semester of statistics in school. I learned a person can use statistics to support anything they want. The only way I would use them is if I have the basic raw data, but then it wouldn’t be statistics.

jca's avatar

Here’s another statistic that is misleading. “Autism cases have increased (blank)-fold% since (insert date here)” or “Autism cases are on the rise.” If you read the recent history of autism, other disorders, such as PDD (Pervasive Developmental Disorder) have recently reclassified as autism. So yes, autism cases are on the rise, but the reason why is hidden in the fact that the other disorders that have been placed under the umbrella of Autism have added to the total number of autistic people. That statistic is misleading.

funkdaddy's avatar

The problem is statistics are used to summarize complex issues in a way that can be quickly digested. The tendency for most people who present “the numbers” is to further prune it down to just the juicy bits or the parts that prove a given point.

That’s when I get skeptical, when someone drops one number out of a whole array of options as proof of their point. It doesn’t make sense to say 40% of people love cats without either relaying the question or relaying the other data. It just doesn’t mean anything.

If 40% of people love cats and the question was “Do you love cats, true or false?” that’s far different than someone asking “What animal would you most like in your home?” with 40% saying cats, 40% saying none and 20% saying dogs.

It’s a benign simplified example, but if you replace “cats” with a political candidate, you have the makings for what passes as a typical use of statistics in the news.

I trust statistics when I can see enough of the information to feel informed or understand the material they’re based off of. And that’s the problem, the way their typically used outside of the studies and papers their based off of is meant to summarize quickly so presenting the full data kind of defeats the point.

Footnotes are a great solution that are underused and don’t seem to have an equivalent outside of text. (video, tv, etc. rarely cite sources outside of “A new study from the Federal Bureau of Cats says…”)

Economic statistics do a great job I think because they’re standardized and reported the same each and every time. The problem again is you have to have a certain understanding of the material and what the numbers actually measure is rarely covered when they’re presented.

DerangedSpaceMonkey's avatar

It would take 73% of all people who make up 15% of all statistics to stop making them up at least 10% less than they are now.

Berserker's avatar

I see them as the means to get a good idea on a situation or whatever, but I certainly don’t conclude anything concrete by hem, since stuff changes all the time.

mattbrowne's avatar

When they are created by people without sufficient academic training in math.

Answer this question




to answer.
Your answer will be saved while you login or join.

Have a question? Ask Fluther!

What do you know more about?
Knowledge Networking @ Fluther