In case you know people who prefer donations to charities over gifts at Christmas, check out GiveWell’s new charity ranking. They only give their top three charities–the Against Malaria Foundation, GiveDirectly and the Schistosomiasis Control Initiative–which Andrew Gelman finds a little strange but GiveWell were nice enough to give him an answer as why this is the case.

I started giving monthly to Village Reach a couple years ago thanks to the recommendation on GiveWell. I was kind of curious as to why they weren’t on the top three list anymore and I found it was for a good reason:

VillageReach was our top-rated organization for 2009, 2010 and much of 2011 and it has received over $2 million due to GiveWell’s recommendation. In late 2011, we removed VillageReach from our top-rated list because we felt its project had limited room for more funding.

It’s encouraging to know that GiveWell is not only concerned with how effectively money is used by these charities but how much they actually need the money. I’m going to continue giving to Village Reach but I’m going to consider one of the charities on the new list to give to as well (since I took another one of the charities I give to monthly off my list).

In case you’re curious, here’s the list of criteria GiveWell uses to evaluate charities:

• Strong evidence of positive impact on people’s lives. More on this criterion here.
• Highly cost-effective activities. We seek charities that provide high “bang for the buck,” in terms of changing many lives (significantly) for relatively little money. Available cost-effectiveness estimates involve a great deal of uncertainty and approximation; we place limited weight on estimated cost-effectiveness, but we are mindful of extremely large differences. More on this criterion here.
• Room for more funding. It isn’t enough to identify a strong program; we seek to identify strong programs that can productively use more donor funding. More on this criterion here.
• Transparency and accountability to donors. Recommended charities must be willing to share enough in-depth information about their work that we can assess them on the above criteria.

No Comments »

Yesterday, John Ibbitson wrote an article in the Globe and Mail disparaging, “ivory-tower economists.” This morning, ivory-tower economist Stephen Gordon took exception with this characterization tweeting: ”Forget ivory tower types. Real insight comes from getting spun six ways from Sunday by a source you can’t name!”

Now, I have no idea what the hell these guys are talking about, investment or some such thing. That’s why I follow Gordon, really, because I like getting his take on a subject I have some interest in, but don’t have the background to understand completely . That said, I do have experience in the ivory tower and I think both of these guys have a point, although, not necessarily the one they think they’re making.

Gordon’s right that the ivory-tower has some great insight because that’s what they get paid to do: understand and study complex issues. On the other hand, and this is based on my experience in ecology and epidemiology (and, actually, some of the economics papers I’ve read related to my research), ivory-tower research has a tendency to oversell the precision of its findings. In fact, epidemiologist John Ioannidis has made an entire career out pointing out where ivory-tower epidemiologists (and, by extension, ivory-tower economists who often use similar methods) overstate the degree of certainty of their findings.

As a quick, more technical example, most studies do not take into account measurement error. Measurement occurs because of imprecision in our measurements and it doesn’t have to be because somebody made a mistake. Blood pressure, for instance, can have important variation over the course of a day or week, but when we measure it for studies, we take one time point (or sometimes a couple) and assume that it is a perfect representation of that person’s “overall blood pressure”, whatever that might actually be. Once research starts correcting for these types of measurements error, error bars can start getting wider and wider until we are a lot less certain about the range of an effect than we we were previously.

Bottom line: often there’s no better place to go for information than the ivory-tower (if you can actually read what primary literature has to offer) but beware, sometimes–and here I’m being constructively vague about what that sometimes might mean–the ivory-tower can claim more certainty than is warranted by the data.

A nice complementary discussion to this is Andrew Gelman commenting on the degree of certainty in Nate Silver’s work trying to predict the upcoming American election.

No Comments »

Neil DeGrasse Tyson with some Carl-Sagan-esque thoughts (h/t Ben Goldacre). The first half is good but if you’ve only got 5 minutes, start listening at 7:35:

No Comments »

A post came over my twitter stream yesterday (thanks @sara_johnson!) and piqued my attention enough that it actually distracted me from the completely insane Penguins-Flyers NHL playoff series. That automatically makes it worth writing about.

I hate writing a post only to critique another post but sometimes you learn more from ideas that are wrong than from those that are right. My problems, however, start with the title: “Why medicine is not a science and health care is not health.” Definitely eye-catching but a title that doesn’t match the text very closely (particularly when the first two sub-headlines seem to confirm that medicine is, in fact, a science as does a portion of the text). A minor issue maybe, but the point of the article, as far as I can tell, is that medicine doesn’t focus enough on the difficult to measure aspects of health, which I agree is an important point, and that might be, in part, due to the fact that the powers that be have no motivation to dredge up these complex issues. Bringing the discussion of medicine as a science is neither informative or, I think, accurate.

Here’s an example of this confusion:

My real objection to medicine as a science is that by focusing on what can readily be quantified, it ignores what cannot, such as the social determinants of health and disease.

I can see that point here is that we don’t focus enough on the social determinants of health and being a social epidemiologist, I couldn’t agree more. But why confuse this important idea with the idea that medicine isn’t a science? Ignoring problems doesn’t make something not a science.

Further, is she talking about medical practice here or research? She can’t be talking about research because there is an entire field of research, social epidemiology, that addresses this. Is it difficult to study? Are social determinants of health difficult to quantify? Yes and yes. But what’s being done is still clearly science and, at least some of the time, of value.

If she’s talking about medical practice, are we now expecting our already overburdened front-line physicians to be addressing questions as complex as the social determinants of health? I very much wish we could do this. Actually, I’m told at least one physician has made a good start by inventing what is essentially social pediatrics (unfortunately the website is in French only). It would seem to me more effective to address the social determinants of health at the level of public health and policy, where science is both possible and being done, rather than at the point of care.

As a kind of technical aside, and as I’ve mentioned in a previous post, epidemiology in this case, will never be able to say why an individual gets a particular disease precisely because it is a science and inference on that level isn’t possible (without a time machine!). Epidemiology can answer the question, “why do people get sick?” but not, “why did this person get sick?”

Now into the health care is not health section:

Science prefers to isolate and understand one thing at a time, but the ideas I just mentioned are difficult to separate from their social context and cannot readily be studied in a controlled environment. The topics for medical research that seek and receive funding are those that investigate ‘reductionist’ theories of disease: things that fit neatly into categories sanctioned by the medical establishment and that can be readily measured, compared, and replicated. The cause and effect relationship between stressful living conditions and the health of individuals does not fit neatly into any simple mechanical model with the potential to make successful predictions.

First, science isn’t limited to ‘controlled environments.’ If that were true then, yes, clinical medicine or observational epidemiology wouldn’t be science but then neither would astronomy, climate science (which the writer admitted earlier was a science!) or a number of other alleged sciences. Science is more difficult without ‘controlled environments’ but not impossible.

Second, we can and do study all of the things mentioned. Is it difficult? Yes. Are we as confident in our results as we would be if they came from a trial? Not by a long shot. But these things are being studied to the best of our abilities (although at our social epidemiology journal club I’d probably be more likely to say ‘to the worst of our abilities’!).

I fully agree that the social determinants of health need higher priority in our health system both from the points of view of research and practice. I also agree that it’s disingenuous to say any kind of for-profit health system can have health as its true bottom line. But why drag whether or not medicine is a science into this discussion? Even if there were good arguments for this being true, which is not the case here, why distract from the main points of the post? Then again, if this article hadn’t had these particular quirks, I would have had nothing to write about! Actually, I have a couple partly finished posts that I should really get around to…

1 Comment »

Maybe I’m behind the times but I had no idea what I was getting into when I saw this. And neither should you. Click on the screenshot below on the right to check out this two minute German film project:

The “About us” section of the website only says: “We are three film students and we didn’t care about nature. This is our way to do something.”

As @bouchane said on twitter:

No Comments »

I hate it when people use a more complicated words when a simpler one will do the trick. An example I find particularly piercing is when people say ‘utilize’ when ‘use’ could easily do the trick. Often, it’s done just to sound more sophisticated though, once I realized this, the word ‘use’ started to sound more sophisticated!

Along these lines, a couple days ago, the Ottawa Citizen published an article referencing a Science editorial about how terrible jargon is. The first line of the Citizen article–the lede in journalist jargon–says it all:

One of the world’s top academic journals is begging scientists to speak plain English — or German, or Chinese — instead of the “insane newspeak” of jargon that’s only used for showing off.

So, you’d think I’d be totally on board with this. Jargon is used for showing off but it’s far cry from being “only used for showing off”. Both the writer of the Citizen article and the Science editorial recognize the existence of an important problem but the wrong one.

Jargon, acronyms and abbreviations all have a very important role to play in science writing and communication. Used properly, they can relate a complicated idea to someone in one word rather than in two paragraphs. The problem isn’t the existence of jargon but that some scientists’–or any class of specialists really–inability to recognize when their audience can’t match the jargon with the definition.

In a scientific journal–especially one with an expensive paywall–the target audience is other people who are specialists in the field. I don’t see any problem using jargon people with a similar background will understand. The problem starts when these scientists talk to the press of give a presentation to a general audience and use the same jargon.

I still think scientific articles should be accessible to the general public but rather than making the article itself more accessible to the public, it should be accompanied by a lay-summary detailing how the study was done and its main results and implications. That way, a more detailed understanding of the research is still available in the scientific article to those who want to do a bit of background reading in order to understand it.

I’m sure everyone reading this either works or is interested in a field where they use jargon that other people might not understand. If you’re good at communicating you recognize what jargon people will understand and which they won’t.

Now I’m just have to go back and reread this post to make sure I didn’t use any jargon…

(Via: @dgardner; comic used with permission from Dave’s Stupid Web Comic)

1 Comment »

2 Comments »

CBC’s The Sunday Edition had a parade of cancers researchers on the show through November and early December in an attempt to satisfy a continuous stream of outraged listeners. It started out as a conversation about why cancer screening recommendations have been reduced prompting irate listeners to wrote in saying their daughter was getting every sort of cancer screening regardless of what the guidelines say. Of course, they forget that screening has important risks associated with it as well. But that’s another story.

To quell that storm, the following week another guest came in to talk more in depth about cancer screening but this guest then made the mistake of mentioning how many cancers can be attributed to behaviours such as obesity and smoking. In fact, the guest said, behaviours cause more cancers than environmental toxins despite host Michael Enright prodding her to say otherwise. Cue another deluge of angry letters saying that ‘isn’t it obvious that the toxins in our environment are causing pretty much all our cancer and shame on this researcher for blaming the victims of cancer instead of the big corporations spewing all these toxins!’ Whoa. That was intense.

When it comes to blaming the victim, these people have it entirely backward. Epidemiology cannot, ever, point to a person and say with certainty that their cancer was caused by x, y or z. Even if a three-pack-a-day smoker got lung cancer, there’s always the chance that something else caused the lung cancer.

What epidemiology can do is to look at two groups or people who are as similar as possible in every way except one being obese and the other not (which, especially in the case of obesity, is very difficult to do). If it turns out one group has more cancers than the other, epidemiologists can attribute it to obesity. Epidemiology can only ever tell what might affect your risk of cancer, it can never say why a particular person actually got cancer.

The only way we could ever blame a specific person for their cancer is if we went back in time, made them not obese in this case, and ran the clock forward again to see if they still got cancer. If they didn’t, then you could say their obesity, in some way, caused their cancer. We’re a long way from time travel so we’ll have to settle for the first approach I described.

What’s more is it’s clear that the public, and more than likely a good chunk of those angry listeners, do blame victims of cancers in some cases. I can’t find the data right now but which cancer charity do you hear the least about: breast, prostate or lung? Why do you think that is?

No Comments »

You should watch me write these posts. I had no idea this post was going to end up talking about healthcare.

Jesse Galef of Measuring Doubt caught ESPN writer Pat Yasinkas making a classic mistake called outcome bias. Talking about the Atlanta Falcons who tried and failed on a fourth and inches attempts:

When Mike Smith first decided to go for it on fourth-and-inches in overtime, I liked the call. I thought it was gutsy and ambitious. After watching Michael Turner get stuffed, I changed my mind. Smith should have punted and taken his chances with his defense.

I hope it’s pretty obvious what’s wrong with . It’s easy to decide what the right call would have been after the fact. Next time Yasinkas wants to play the lottery, he should ask to see if he can see the numbers that will be drawn first. Read the rest of Galef’s post in which he demonstrates that the Falcon’s actually did make the right call.

I see sports fans making this mistake all the time and I’m sure I do it too. Why didn’t the Canadiens draft Jeff Carter, Ryan Kesler or Ryan Getzlaf with their first pick in 2003 instead of Andrei Kostitsyn (this might seem like an obscure complaint but Habs fans will never forget it)? There is definitely skill involved in assessing hockey talent in young players–some scouts are better than others–but there’s also a lot of unpredictable things that can happen to a young player that makes him a better or worse player than he was on draft day.

I also see people and journalists judging politicians in the same way. We expect our politicians to always make the right decisions and we judge them based on the outcomes of their decisions when we should really be judging them against what the opposition would have done in the same situation (ignoring the fact that the opposition will always say they would have done the right thing).

But there’s another more subtle way people make this mistake. When discussing Canadian healthcare with my more conservative friends I often hear, “why should I have to pay for someone else’s healthcare?” Oh, my dear conservative friend, you don’t get to see the outcome of the lottery before you decide whether you’re going to buy your ticket. In classic Rawlsian style, what if I asked you to choose before you were born, before you knew you were going to be born healthy or with a debilitating expensive-to-treat disease, whether you wanted single payer healthcare or not, would you choose to risk it? Maybe. You might decide you still want to roll the dice but the choice shouldn’t be as obvious anymore.

I would then go on to ask you if it’s fair that someone who has a genetic disease should pay for their disease given that are not to blame for their disease, but that’s for another post.

No Comments »

I can’t remember who I was talking to, but I had an interesting discussion the other day about how people might–I said might–think of humanity and the Earth as a whole once commercial flights to space became commonplace. Here’s a video from the International Space Station that might give us a glimpse into what that might look like. I always find that Sigur Ros is the best soundtrack for this kind of thing so I shut off the sound on the video and played the song Hafsol along with it instead. And make sure you’re watching it in HD.

(via Open Culture)

No Comments »