I have a scientific background, I just don't want to wave it around.
Waving it around or not, you should at least try and seem like you have a scientific background if you're going to pontificate so clumsily on scientific matters. The result of our conversation went like this:
Pollutant: A substance or condition that contaminates air, water, or soil. Pollutants can be artificial substances, such as pesticides and PCBs, or naturally occurring substances, such as oil or carbon dioxide, that occur in harmful concentrations in a given environment. Heat transmitted to natural waterways through warm-water discharge from power plants and uncontained radioactivity from nuclear wastes are also considered pollutants.
I don't recognize the free dictionary as an authoritative source. However, your statements are still logically disconnected. We pollute -> AGW is real does not follow.
It's irrelevant what you recognize as an authoritative source if you can't identify accurate definitions. Try the publication known as
Nature. Not
Atlas Shrugs, Richard Lindzen, or other such libertarian sources whose agenda is suspect with regard to scientific study.
Once again, I cant say for sure about climate science. But for all things biomedical (cancer, alzheimer, etc research), nature is actually one of the worst journals out there. You can never even tell what the hell the people did to get their results, let alone assess how valid their conclusions are.
I want to emphasize this a bit more. It is clear to me Nature does not care about showing your data (they accept dynamite plots), they dont care if you report the sample size, they allow using SEM rather than standard deviation in showing how uncertain the results are (SEM gets smaller if your N is bigger, ie spent more money). This is basic rational person stuff that is not enforced.
I could go on and on, that journal is crap and only lives on due to inertia of consensus and authority. I have really come to the conclusion that the last fifty years of science was just "experts" measuring how sure they are of their opinions. I am not alone, look it up.
Do this:
https://bitcointalksearch.org/topic/someone-needed-to-search-old-academic-papers-and-textbooks-127448The reporting of statistics of biomedical science involves three steps.
1. Open SAS.
2. Do as many statistical tests as SAS allows.
3. Publish
In fact most biological sciences are like above, except for those in ecology, evolution, or computational biology.
I think that this is just a symptom of the disease. The disease itself is the "null ritual" meme that pervades all fields of science. The only exceptions are those that actually try to model and predict things (like your examples).
In the context of big business vs science, my point is that scientific consensus means nothing. Most scientists aren't even capable of reasoning properly since their minds have been clouded by indoctrination with nonsense "statistics". This happened to me as well, I only happened to ask why we do things the way we do them. Anyone who actually takes the time to look into what has been going on will come to the same conclusions I have. P-values and statistical significance as widely used is measuring how much effort you put into generating your data. Effect size is measuring the amount of bias in the field.
There has been some progress but it has been more accidental than anything, there is not much low hanging fruit left that can be detected if we continue relying on these methods. Climate scientists do create models and try to predict things, so I suspect that field may be better than most.
Here is the basic error:
The p value is calculated as the probability of the data (or more extreme) occurring by chance if we assume a strawman null hypothesis is true (every null hypothesis is false, because there are always differences between any two groups/things). This number is then erroneously interpreted as an error rate or worse, probability the researcher's pet hypothesis is true. Because the likelihood of "more extreme" and less likely results are averaged in (which even the guy who came up with it, Ronald Fisher, said was indefensible save as an approximation), the error rate seems much lower than it actually is. You can also use bayes' theorem to prove that interpreting the face value of a p-value as a probability requires you set a prior probability the null is false of 75-90%.
Even then, just because the null (strawman) hypothesis is false, it does not make the researcher's pet hypothesis any more likely to be true. The entire thing is a waste of time but it is the foundation of pretty much all modern scientific reasoning, funding, and publishing. Worse it discourages actually looking at data and trying to figure out what is going on, since researchers believe they are using an objective method backed by mathematics and logic, which they are not. They using methods of reasoning that (it looks like so far) were invented by the guy who created the ACT in an effort to provide people with "non-controversial" statistical methods at the behest of his university and publisher.
Edit: and oh yea. Comparing the probability scientific consensus is right with the chance the arguments of politicians and propagandists are right isn't even possible, since you get a divide by zero error. So lets stop doing that.
