Top 10 Tricks to Find That Temperature Trend You Always Dreamed Of

1) What is a trend?

A trend is a line you can cross through a graph to make the science go away. You can use a nice thick pen or an electronic paint brush to do this. Trend lines have been a key resource in the battle against data.

Cross out that data and it no longer counts

2) I saw a graph I don't like. Can I cross it out?

Yes, but you have to be careful. You can't just put any old line through it. It has to be a very specific and carefully chosen line. If you draw a trend line through data incorrectly you might inadvertently make the data appear alarming. Most trends in data these days are alarming, so if you don't know what you are doing you seriously risk making the data look worse. Trend divination is a skilled art. You need to be adequately trained in the ways of blog science and scientific obstructionism before you can begin crossing out graphs at will. When in doubt defer to the professionals, some of us are paid to do this kind of thing afterall.

3) How many possible trend lines are there?

A lot. Of different lengths and gradients through different types of data. Far too many in my opinion, reflecting the ridiculous size of the federal government.

4) Okay, so how do I carefully choose a trend line?

You need to be well aware of what makes a good trend and what makes a bad trend. Bearing this in mind the choice can nevertheless be quite fiddly at times; imagine if you will picking a small fruit off a tree. You don't want to pick any old fruit, you want to pick the best one you can.

5)  What makes a good trend line?

Long flat trends are good. Steep trends are always bad. If in doubt, remember the rythme: Long and low, taxes no. Steep and high, freedom bye. Imagine a trend line being a bit like a low bridge spanning an alarming ravine, which will in effect make it less alarming. You want to find the longest flattest line you can.

6) So I have my trend, what do I do now?

If you find a good trend you need to parade it before to your local media, while reciting its properties over and over. How long is it? I have a big one, mine is 18 years and 11 months long. This proves nothing has happened for almost 19 years.

7) What happens if your trend goes short

Even the best trend will on occasion suddenly go short. This seems to happen roughly every 5 years or so, meaning there is absolutely no trend in the breaking of trends. If you are inexperienced you might panic and question the validity of your method when you wake up to find your treasured 19 year trend fallen. But once you get good at this you will be able to simply advance the beginning of the trend forwards a few years and trawl the news for excuses. El Nino, communists, Obama. All these things are valid reasons to cite.

8) What if I can't find a trend that works?

Sometimes try as you might you just can't find a sufficiently long flat trend to make a scientific record go away. This is a sure sign that particular scientific record must be fraudulent, even if you swore by it before. True data will always conform with political reality. Perhaps it is time to check scientists emails and investigate their NetFlix accounts to see what movies they've been watching.

9) Where can I find good trends?

These days I find I am only satisfied with a good lower troposphere satellite trend. That said there are some fantastic up and coming trends on other planets where we have little to no data.

10) What does it mean for a trend line to be "not significantly different from zero"?

Statistical significance is a concept invented by statisticians in order to prove nothing alarming can ever happen. If a trend is not statistically significant from zero it means the trend is 100% confirmed to be exactly zero and therefore our taxes, if there must be taxes, would be better spent elsewhere.

As an example imagine Bob sets out from home one morning to go to work. He travels either East or West. Because we don't know in which direction he has gone, or how far he travelled, this proves he hasn't left the house.