More about observational and experimental science

Looking at the stats pages of my WordPress account, I found quite some views suddenly coming from reddit, more specifically from the thread The Observational science that is mistaken for an Experimental science, where some quotes from my previous post with the same title were discussed. I don’t see how to react there (and I don’t even know if I want to, looking at the style of reactions), but the reactions are rather typical, so some more context could give better understanding.

[Quoted from my post]
The problem is that it takes experimental research to determine cause and effect.

[Comment on this quote]
This is simply not true. Cause and effect can be determined from first principles (Such as increasing radiative forcing will make it warmer), inferred from observations (Thunder is caused by lightning) or form part of a theory that is proven by other means (gravitation affects light because it is a curved space-time phenomenon)

The first example is an interesting one as it takes us in the domain of global warming: “increasing radiative forcing will make it warmer”. I can agree with that one to some extent. In a simple linear system this will be a no brainer, but the problem is that we are observing this increase in a complex, coupled, chaotic system. There are looooaaaads of interactions/feedbacks in this system. This will make this (at first sight) simple equation a lot less clear. CO₂ can increase temperatures when there is more of it in the atmosphere, but what most people forget is that other things will change too… For example, the increased temperatures can mean more evaporation (which needs energy), more clouds (which could block sunlight) and so on and so on. Such a system could dampen or counter the effect. One of the big questions in global warming is not whether there is a warming, but how much of that effect is due to human CO₂ emissions. The fact that it warms doesn’t necessarily prove that humans are the cause.

[Quoted from my post]
Experiments that can’t be created in climate science because we only have one Earth and therefor don’t have the luxury to create an experiment to pinpoint the exact cause of this temperature increase.

[Comment on this quote]
Nevertheless the earths climate can be studied, and knowledge gained.

Nobody said the Earth’s climate could not be studied or knowledge could not be gained in an observational science. The point was that it is necessary to have an experimental approach to pinpoint a cause. One needs to manipulate only one variable, keeping all (or at least most) other variables constant. In such an approach the outcome will be from the variable that was controlled, avoiding possible interference from other variables.

In Experimental research the investigator controls the variables, while in Observational research the investigator observes nature taking its course. It is a set of observations without controlling variables, therefor the risk that some not observed variable is the real cause and wrongly assuming that (one of) the measured variable(s) is the cause. Only by controlling the variables and by varying one of them, it is possible to pinpoint what exactly is the cause. So I stand by my statement that cause and effect can only be determined by experimental research.

[Quoted from my post]
Despite that, climate scientists declare in no uncertain terms that they do know unequivocally the cause of the increasing temperatures since the 1950s: CO2.

[Comment on this quote]
Certainly all the scholarly papers that look at it put CO2 as a major contributor, and anthropogenic forcing is larger than natural forcing.

But I’d like to see an example of this “unequivocal declaration” that you’re talking about. Unequivocal declarations smack more of denialist propaganda than science.

I would have thought that Russell’s observation that “the stupid are cocksure while the intelligent are full of doubt” still holds.

Can you point me to some examples of these “climate scientists declaring in no uncertain terms that they do know unequivocally the cause of the increasing temperatures since the 1950s”?

Or do you suspect, as I do, that this is simply made up for the rhetoric?

It only took me a couple minutes to find this little gem: Katherine Hayhoe: what we know. This is certainly not the first (climate) scientist who states absolutely no doubt whatsoever, probably not the last either.

I have no doubt that Russell’s observation still holds, but as far as I can see the quotes pointed out there was reason for doubt, while the climate scientists are really cocksure. Just saying. 😉

This was the comment that was linked to, but the rest of the thread wasn’t much better. The most confused comment was this (emphasis by the commenter him/herself):

[Quoted from my post]
In an observational study it was found that women who were prescribed estrogen had less heart attacks, yet in follow up clinical studies it was found that estrogen was actually increasing the risk on heart attacks.

[Comment on this quote]
What’s the difference here?

There’s this “observational study” that had a control group and a treatment group that were compared, and then there were these “clinical studies” that had a control group and a treatment group.

Clinical, of course, means observational:

based on observation and treatment of patients as opposed to theory or basic science.- Free Dictionary

Is this the best denialist propaganda can come up with now? A consensus doesn’t count in climate science, because of a difference I made up, that doesn’t even have a meaning?

Huh?

First, as far as I know, the observational study had no “control” group at all. They followed a huge group of post-menopausal nurses taking estrogen and the result was that the nurses had less heart attacks than should be expected from their age and other risk factors. The HERS study on the other hand was a double-blinded, randomized, controlled study. Variables were kept as controlled as possible, something the observational study did not do.

Second, the statement that “Clinical, of course, means observational” makes no sense at all. I didn’t make up the difference, because I didn’t even make the difference. Both studies were “clinical” (also from the Free Dictionary: the first meaning of “Clinical” is “Of, relating to, or connected with a clinic”).

The difference was that the HERS study was a randomized, controlled study (this is experimental) and the Harvard’s Nurse study was not (this is observational). The distinction was therefor not between “observational” and “clinical” as seems to be suggested, but between “observational” and “experimental”. By the way, that was also a giveaway in the title.

The reason I gave this example was to explain WHY there was such a difference in both types of studies. The observational study came to a (wrong) conclusion because the outcome of the study was caused by something else that was not subject to their study. This is the risk of all “observational” studies. The result can come from another factor that was not observed during the study (a “confounding factor”). In this case it was the subjects having a reduced risk of heart disease in the first place, therefor creating an unduly positive picture of taking estrogen.

The central point of the post was that climate science is painted as a science that has it all figured out. My perception is that this is not warranted because of it being an observational science with all its limitations. Yet this important point is not being told…

… and should be.

Advertisement