Feeds:
Posts
Comments

Archive for September, 2012

Urgent: action needed

In the past decade we have seen a huge increase in the amount of data that can be easily accessed by just about anyone.  This is in stark contrast to prior decades, where data were carefully guarded by whoever had collected it. While this change has many potential benefits, it is not without risks.

With so much data available on the web, any wannabe statistician can download data and start analyzing it all willy-nilly.  Many types of errors could be made, such as assuming correlation equals causation, getting the causal direction wrong (‘if I eat more ice cream it will become summer’), data snooping (high false positive rate), not understanding variability, randomness, or generalizability. For example, someone who was unaware of truncation by death (survivorship bias), might see that 80 year old smokers have a longer life expectancy than 80 year old non-smokers, and decide to start smoking.  This could have fatal consequences.

How do we protect the public from these kinds of errors, while still making data available to a much wider audience than we have in the past?  I propose the following.  Data should be available to anyone who wants it, but they first must receive written permission to use the data from a licensed PhD-level statistician.  It would be a simple process.  Let’s say Debbie wanted to analyze data from the National Health and Nutrition Examination Survey.  Debbie would make an appointment with a statistician who has a license to issue data.  Debbie would inform the statistician of her goals, and the statistician would help draft an analytic plan.  The statistician would then issue her a password that could be used to download the data.  The statistician would decide for how long she could use the data, before having to come back for a renewal.

For the good of the public, it is time for the American Statistical Association to lobby congress and make this happen.

Now, I’m sure some people will say “I should be able to have access to data without permission from a statistician.  If I make a mistake and it harms me, well that’s my problem.”  Okay, Brainiac.  I could make the same argument about medicine.  I have used an asthma inhaler for 10 years.  I know it works.  I know how to use it. I know I need it.  I know how to read the warnings.  Why should I have to go to an MD to get written permission for a refill?  You see, even though I know what I’m doing, not everyone does.  If an MD wasn’t overseeing the process, no doubt someone would crack open the inhaler canister and pour the fluid in their eyes.  It’s the same with data.

The cynical (i.e., stupid) people in the audience are probably thinking “this is just a way to drastically increase the demand for PhD level statisticians, which will lead to a huge increase in salary for statisticians, including the author of this blog.”  While no doubt that that would be a consequence, it by no means is a factor.  The only goal here is to protect the public.

Advertisements

Read Full Post »

Cultural universals

If you go looking for cultural universals, how do you decide what to look for?  Most likely, you will begin with a list of things that are common in the local culture that you are familiar with, and look for them elsewhere.  However, once you are looking for X, it’s a lot easier to see X (many cultural characteristics have subjectivity element to them).  For this reason, I suspect that the list of cultural universals is very dependent on the culture of those making the list.

Read Full Post »

The Healer

The most beloved person in town was a man named Gary Cooper.  Gary was a father of three.  His wife, who, as he put it, suffered from boarderline personality disorder, left him last year.   Gary had tried to help her with her problems, but this is one of the rare cases where he was unsuccessful.

You see, Gary was a great healer.   He helped friends, family and especially his children.   Everyone considered him one of the kindest people they’d ever met.

For example, when his daughter Hannah told him that she had two friends at school, he immediately realized what this meant — there were a lot of kids in her class that didn’t like her.   At first, Hannah denied it “I don’t mind having fewer friends.  I like to daydream in class, and then play with my best friends at recess.”   Gary said “Hannah, it’s okay to be sad.  You probably feel like most of the kids in the class don’t like you.  I understand why that would upset you, but pretending not to be bothered by it won’t help you.”    After that, she could finally see what her life was like, and cried about her unpopularity.   Gary went on to spend many hours a week helping her to deal with her pain, and teaching her how to make friends.   Everyone who heard about what he was doing for his daughter was impressed.   People would talk about how big his heart was.

Gary’s other children suffer from mental health problems as well (probably inherited from his wife) and he helped them find ways to live a normal life.  People would say to each other “I don’t know how Gary does it. ”

Gary was like that with everyone.  He helped his neighbor Doug deal with his dependent personality disorder, and his sister Tammy deal with her mother-in-law’s emotional coldness.

Gary would diagnose and then heal.  Or, in many cases, diagnose, convince and then heal.  The convincing part was an important step, because many people are resistant when it comes to seeing what is wrong with them.  Thankfully, Gary was such a kind and gentle person that people trusted him.

A few months ago, a nasty woman named Mindy moved into town.  She kept hearing about this wonderful person named Gary, but she wasn’t impressed.  In fact, she would accuse Gary of creating the problems that he was curing.  People hated Mindy. But Gary handled Mindy in his usual selfless way.  He explained to her that she was delusional – that she is someone who just can’t see the problems that are all around her.  He offered to help, but Mindy refused and moved away.

Some people just cannot be cured.

Read Full Post »

Arranged marriage

When we make a promise, we are committing our future selves to do something that they might not want to do.  For short-term promises, this is not necessarily a problem, because you will be a lot like your short term future self.  For example, you will have a lot in common with one-year-in-the-future-you (call it you+1).  You will share most of the same memories.  You will have had most of the same experiences.  The genes that are expressed will be quite similar.

However, you+30 might be a lot different from you.  Different sets of genes will be expressed.  You will have far fewer shared memories.  Etc.

Suppose you are 25 years old and decide to get married.  You are committing you+30 to this particular person.  You are essentially arranging a marriage for you+30.  Are you the best judge of what’s best for you+30?

Your parents, for example, each share roughly half of your genes with you.  And if they are currently the same age as you+30, and you take into account gene expression, they might be ‘more related’ to you+30 than you are (in some ways).  If someone is going to arrange a marriage for you+30, why not your parents?

note: I am not suggesting that what we typically think of as arranged marriage is better (or worse) than self-selection.  I am just thinking about the ethics of long-term commitments.   For example, if people lived for 5,000 years, on average, would we allow people 25 years of age to commit the 3,500 year old version of themselves to a partner?  And if not, what time span is too long?

Read Full Post »