In an earlier POST I discussed the broad dimensions of the lie that “this month was the 8th hottest ever”. I discussed this earlier from a top-down perspective where I showed that the data is expected to show exactly that and that this data is obviously worthless in determining a trend. In fact, as I showed, we would be SURPRISED to NOT find the hottest decade ever in the last decade. Now I tackle the question from a statistical point of view. Why is this a common and rather stupid mistake to make?
First off, we know that one month of data can not create a trend. Global warming true believers LOVE to point out that 15 years is too short of a time period to use as a snap-shot of the climate and they dismiss the fact that temperatures stopped going up 15 years ago due to this. Whether or not they have a good argument is not the point here though. The point is that they state emphatically that 15 years is a too short of a time period. So what do they do all the time to just make this claim comical? Why, they say we must be warming because the last month was the 8th hottest and the year to date is the third hottest. So, basically in their minds 15 years is too short of a time period for a trend, but one month is long enough. Logically, it’s a huge disconnect and only the most brain-dead true believer would actually fall for that.
To make matters worse, the problem is that using statistics we have several techniques to determine trends in the data. The most commonly used is known as linear regression where you fit the data to a simple line. This is common when discussing temperature because the trend is more than likely a line either heading upward (warmer) or going down (colder). The other option is basically stagnant or no change. There are several other methods that make sense when discussing trends that I will not mention here for simplicity sake, but rest assured these methods ALSO use at least 10 years of data in determining a trend. Because taken by itself, one simple statistic such as “hottest decade ever” tells you just one thing that is useful, “in the time we have measured temperature, the last decade was hotter than the rest.”
But anyone who uses tests such as “the decade was the warmest on record” or other nonsense are simply not well-versed in statistics and probably have no understanding of the climate in general. Scientists have agreed rather strongly that we have warmed since coming out of the Little Ice Age. This is without doubt, the problem is that global warming believers want to make the claim that the warming is worse than ever without using the primary and even secondary statistical methods for determining trends. They want to use random useless facts to make their point. And this is the problem. There are tons of odd statistics that show nothing in terms of where we are heading. To put this into perspective, claiming that this is the ninth hottest month simply tells you that for month X, the temperature was a little on the high side. Was this just a freak occurrence due to hot weather where we measure temperature? Or was it due to a trend? Or was it just noise? And that therein lies the problem that good statistics tries to answer. Bad usage of statistics attempts to get a point across to mislead people into believing you are telling the truth. And this is what is happening. Either the people who parrot this message are stupid (which I suspect is always the case) or they are spreading a lie for political gain. Either way, these people deserve nothing but derision for being dunces. They are making a simple mistake and they should be treated as such. One of the worst people in this regard is a certain cartoonist who runs the site “sceptical science.” This guy who is a cartoonist by trade makes this mistake at least 5 times a day. I tend to think this is an indictment on the Australian education system to produce a dunce that bad who can not even figure out simple things on statistics.
Good statistical usage gets rid of noise and creates a better picture of what the data is showing. Bad statistics will intentionally mislead and confuse people. And that is the problem. Statistics is hard to explain and even harder to apply correctly to every situation. The number of scientists that incorrectly uses these is larger than the number who uses them correctly. Perhaps the world will never use statistics correctly all the time, but the hope is that hopefully people will be chastened into at least not making these simple mistakes and just making themselves look like huge dunces.