Math

[skierinvermont]

I vote positive.

[BlizzardWx]

Ehh..Even after 3 decades we would just be close to .1C which is very small. But a trend would be any type of consistent change, even if its insignificant.

[Jonger]

Are you asking if its a "good thing", or is this sarcasm?

[valkhorn]

One thing I learned in lab experiments is there's always a margin of error. If you end up with zero or a perfect match, your methods are wrong or your measurements not precise enough. so the one person who voted zero, and I can guess who you are, you're not very scientifically literate.

[FPizz]

With a margin of error Val, that .03 can easily be negative?

[Vergent]

#9

http://www.skepticalscience.com/global-warming-stopped-in-1998.htm

[Snow_Miser]

With a margin of error Val, that .03 can easily be negative?

This. Skier gave a very biased poll, not letting the pollers know what the error margins of that 0.03 Degrees C/Decade trend were.

[valkhorn]

With a margin of error Val, that .03 can easily be negative?

Math doesn't work that way. A range of values doesn't exactly equal a value within that range. There is a probability, but it is not equal.

You cannot say .2 = .1 +/- .2

[valkhorn]

This. Skier gave a very biased poll, not letting the pollers know what the error margins of that 0.03 Degrees C/Decade trend were.

The error margins don't matter. You could have error margins of plus or minus 100 and it still would not equal 0. Equals means it is exactly that value all of the time.

If you were using probabilities, then the only time you could use an equal sign is if the probability is exactly 1. When you have error margins, or a range, you will never have that probability.

[Snow_Miser]

The error margins don't matter. You could have error margins of plus or minus 100 and it still would not equal 0. Equals means it is exactly that value all of the time.

If you were using probabilities, then the only time you could use an equal sign is if the probability is exactly 1. When you have error margins, or a range, you will never have that probability.

Error margins don't matter? Wouldn't a 0.03 Degree C trend with a +/- 0.05 Degree C error margin have a possibility to be a -0.02 Degree C trend?

[valkhorn]

Error margins don't matter? Wouldn't a 0.03 Degree C trend with a +/- 0.05 Degree C error margin have a possibility to be a -0.02 Degree C trend?

Having a possibility - however remote is not equal to having a probability of 1.

[skierinvermont]

The poll question was not whether .03C/decade was a statistically significant positive trend, the question was simply whether it is positive.

[FPizz]

The poll question was not whether .03C/decade was a statistically significant positive trend, the question was simply whether it is positive.

Yes, close the thread.

[Snow_Miser]

Having a possibility - however remote is not equal to having a probability of 1.

Yes, but a significant trend is a trend that deviates significantly from zero. Since this trend does not deviate significantly from zero, we can say that a statistically insignificant trend can be refered to as having zero trend.

[valkhorn]

Yes, but a significant trend is a trend that deviates significantly from zero. Since this trend does not deviate significantly from zero, we can say that a statistically insignificant trend can be refered to as having zero trend.

No. It is measurable, it is significant. Even if it was insignificant, then a positive trend with error bars is NOT a zero trend.

positive trend != negative trend != zero trend

You're seriously painting yourself in a corner. Also, please answer my question to you on the other thread you started.

[skierinvermont]

A 4 year old could answer this question better than SL can.

[Snow_Miser]

This is the definition of a statistically significant trend from Wikipedia:

Once we know the "noise" of the series, we can then assess the significance of the trend by making the null hypothesis that the trend, , is not significantly different from 0. From the above discussion of trends in random data with known variance, we know the distribution of trends to be expected from random (trendless) data. If the calculated trend, , is larger than the value, , then the trend is deemed significantly different from zero at significance level .

According to this definition, the significance of a insignificant trend is not different from zero, thus the 0.03 C/Decade trend has the same significance as a zero trend.

Not sure what's so hard to understand about this.

[Snow_Miser]

A 4 year old could answer this question better than SL can.

Thank you for yet another one of your amazingly insightful comments.

[valkhorn]

Thank you for yet another one of your amazingly insightful comments.

Care to answer my question on your 'petition' thread? Or are you just going to ignore that.

Again, a trend if positive is positive. There's no two ways about it.

[valkhorn]

This is the definition of a statistically significant trend from Wikipedia:

Once we know the "noise" of the series, we can then assess the significance of the trend by making the null hypothesis that the trend, , is not significantly different from 0. From the above discussion of trends in random data with known variance, we know the distribution of trends to be expected from random (trendless) data. If the calculated trend, , is larger than the value, , then the trend is deemed significantly different from zero at significance level .

According to this definition, the significance of a insignificant trend is not different from zero, thus the 0.03 C/Decade trend has the same significance as a zero trend.

Not sure what's so hard to understand about this.

You looked up a definition but you didn't do the math. Show your work.

[Snow_Miser]

Care to answer my question on your 'petition' thread? Or are you just going to ignore that.

Again, a trend if positive is positive. There's no two ways about it.

What we're talking about is if the trend significantly deviates from zero, which it doesn't. There's no way around that.

[skierinvermont]

What we're talking about is if the trend significantly deviates from zero, which it doesn't. There's no way around that.

No, we are talking about whether an observed trend is positive or not. Not whether it is statistically significantly positive. The observed trend of +.03C/decade is positive. Notice the + sign.

In 95% of simulations the trend is positive (IE >0) after 15 years.

Only in 80% of simulations is the trend large enough such that if observed, it would be statistically significant (IE >.04C/decade).

Thus the observed trend of +.03C/decade falls outside the 80% confidence interval, but not the 95% confidence interval.

If you want to talk about the actual theoretical trend (not the observed trend), then if the observed trend is +.03C/decade there is maybe a 10% chance that the actual trend is zero or less (and thus outside the 95% confidence interval. But there is also a 10% chance that the actual 15 year trend is .06C/decade and even inside the 80% confidence interval.

We can now make 3 claims:

1. The observed trend falls within the 95% CI, but outside the 80% CI.

2. There is a 10% chance that the actual trend falls outside both the 95% and 80% CI.

3. There is a 10% chance that the actual trend falls inside both the 95% and 80% CI.

[valkhorn]

What we're talking about is if the trend significantly deviates from zero, which it doesn't. There's no way around that.

Please answer my question in your other thread. I don't want to have to post it here to show to everyone that you're ignoring it.

Skiier answered you correctly. You are incorrect.

EDIT: I'll ask it here anyway:

If the error margins are equal on both sides, the trend is STILL positive. Error margins usually are equal on both sides. It amazes me the mental gymnastics you are willing to do to cling to your position in spite of the fact that the Earth has still been warming since 1998. This is a fact. You have been refuted.

I think we could save a lot of hot air by just asking you one question - could you be wrong about global warming, and what evidence would you need to be shown that you are wrong?

For me it's easy to answer this question. One needs to find another natural mechanism to account for the current warming we are seeing globally. Plenty of people are looking for it but it hasn't been found yet.

If you ignore this, or can't answer it, then it shows you are dishonest. Surely, you don't want to be thought of as dishonest, do you?

[TerryM]

What an amazing thread!

I suppose the Flat Earthers will be relieved to know that mathematics, logic and even common sense can be overcome by simply repeating the same flawed mantra ad infinitum.

Reductio absurdum comes full circle when one refuses to admit the absurdity, but instead continues trying to defend it

Terry

[skierinvermont]

What an amazing thread!

I suppose the Flat Earthers will be relieved to know that mathematics, logic and even common sense can be overcome by simply repeating the same flawed mantra ad infinitum.

Reductio absurdum comes full circle when one refuses to admit the absurdity, but instead continues trying to defend it

Terry

An excellent demonstration of why the term denier is still applicable to many.

[Snow_Miser]

No, we are talking about whether an observed trend is positive or not. Not whether it is statistically significantly positive. The observed trend of +.03C/decade is positive. Notice the + sign.

In 95% of simulations the trend is positive (IE >0) after 15 years.

Only in 80% of simulations is the trend large enough such that if observed, it would be statistically significant (IE >.04C/decade).

Thus the observed trend of +.03C/decade falls outside the 80% confidence interval, but not the 95% confidence interval.

If you want to talk about the actual theoretical trend (not the observed trend), then if the observed trend is +.03C/decade there is maybe a 10% chance that the actual trend is zero or less (and thus outside the 95% confidence interval. But there is also a 10% chance that the actual 15 year trend is .06C/decade and even inside the 80% confidence interval.

We can now make 3 claims:

1. The observed trend falls within the 95% CI, but outside the 80% CI.

2. There is a 10% chance that the actual trend falls outside both the 95% and 80% CI.

3. There is a 10% chance that the actual trend falls inside both the 95% and 80% CI.

If 80% of climate models predicted a consistent statistically significant warming during this timeframe while that has been absent, that represents a pretty substantial discrepency between the modeled and the observed temperature trends. I suspect that over the next decade or so, this discrepency will continue to grow. There is no sign of warming resuming anytime soon in the troposphere or the surface, and this is because of multidecadal oceanic oscillations and the sun.

[Snow_Miser]

Please answer my question in your other thread. I don't want to have to post it here to show to everyone that you're ignoring it.

Skiier answered you correctly. You are incorrect.

EDIT: I'll ask it here anyway:

If you ignore this, or can't answer it, then it shows you are dishonest. Surely, you don't want to be thought of as dishonest, do you?

This is a very easy question to answer.

If the decade from 2010-2019 is not cooler than 2000-2009, I'll reverse my views about the natural aspect of global warming over the past 100 years.

Now I'll ask that question right back at you. What will it take for you to change your views about the anthropogenic aspect of climate change?

[Snow_Miser]

Reductio absurdum comes full circle when one refuses to admit the absurdity, but instead continues trying to defend it

So saying that because a trend is statistically insignificant, claiming that it does not deviate significantly from zero because of this is absurd?

[valkhorn]

This is a very easy question to answer.

If the decade from 2010-2019 is not cooler than 2000-2009, I'll reverse my views about the natural aspect of global warming over the past 100 years.

Now I'll ask that question right back at you. What will it take for you to change your views about the anthropogenic aspect of climate change?

You didn't answer the question. All you did was shift the goalposts.

I already answered your question. Please re-read.

[Snow_Miser]

You didn't answer the question. All you did was shift the goalposts.

I already answered your question. Please re-read.

Okay, let me rephrase it differently for you.

If the 2010-2019 decade is warmer than the 2000-2009 decade, then I will be "proven wrong" about the magnitude of the natural contribution to Global Warming.

And by the way, there are plenty of natural mechanisms to explain Global Warming over the last 100-150 years that are in the scientific literature.