Staredit Network Forums Lite Discussion Topic: Pseudoscience
Pseudoscience
Mar 28 2012, 2:04 am
By: Fire_Kame
Pages: < 1 3 4 5
 

Aug 1 2012, 3:29 am Sacrieur Post #81



Quote from Azrael
Because they are human beings and that is what they do. A better question is, why would you possibly think they wouldn't?

What do you think this is like scientology? Things made up for the hell of it don't make it anywhere in peer reviewed science.

The more likely explanation (far more likely, like, by a lot) is that it's going to affect someone's pocketbook somewhere. I'm not saying it's definitive, I'm saying it's really quite absurd to be a neutral skeptical observer and side against global warming. I think most of the scrutiny should be placed on those who have the motive. And if you don't think you should, you have some seriously screwed up detective skills.

So is it fusion power and it's renewable energy buddies trying to increase funding with their extremely limited budgets and lobbying ability, or is it giant oil conglomerates with oodles of cash and enough lobbying power to convince the world's number one military power to invade a country to get oil?



None.

Aug 1 2012, 3:32 am Aristocrat Post #82



Doesn't matter if global mean temperatures aren't rising. The Arctic ice cap is gone (shipping goes through there now), and the Antarctic is shedding ice shelves like dandruff. CO2 levels are rising, and if you don't recall, CO2 is not a good thing. Obviously, something is wrong here.



None.

Aug 1 2012, 3:57 am Lanthanide Post #83



Quote from Sacrieur
So is it fusion power and it's renewable energy buddies trying to increase funding with their extremely limited budgets and lobbying ability, or is it giant oil conglomerates with oodles of cash and enough lobbying power to convince the world's number one military power to invade a country to get oil?
From earlier in the thread:
Quote from Lanthanide




None.

Aug 1 2012, 6:50 am Jack Post #84

jackAndMooseDay[32] = true;

Lanthanide, have you ever read The Structure of Scientific Revolutions, by T. Kuhn?

The reason I ask is because it covers how scientific communities of a particular time generally have the same opinion on a subject simply because that is what the opinion of the scientific community is. Until a paradigm comes along to upset that particular cycle of scientific thought, the status quo will remain largely unchanged. The fact that there are many scientists who do not have a consensus on global change and the effects of humans on it is not surprising, in the light of Kuhn's book; they are not the majority of the scientific community but are to be expected. But just because there is a scientific consensus on a particular subject at a particular time does not make it the truth.

As for the actual subject at hand, I am not studied enough on it to be able to form a conclusive opinion on the matter, although my general leaning is that it is good to take care of the environment and that often "hippies" go too far in restricting other humans in their desire to take more care of the environment than necessary. I doubt anyone here is well read enough to have enough information to form a non-biased conclusion on the subject.



Red classic.

Aug 1 2012, 8:11 am Lanthanide Post #85



That's why the most recent thing I linked to is important: because the guy who undertook the study was a climate change skeptic, and he was funded by numerous groups that are skeptics of climate change, because they wanted a credible scientific study done that they could point to that would back up their claims.

What we ended up with, however, is a skeptic changing his mind after doing the study, as well as a study that was done using different methodology and rigours compared to other studies, yet came out with the same results. And then they published all of the data online so anyone can download it and look at it themselves.



None.

Aug 1 2012, 9:25 am LoveLess Post #86

Let me show you how to hump without making love.

Whenever I think of pseudoscience, I think more along the lines of philosophy built on top of facts, stacked so far that it can become either ridiculous or very meaningful. Nothing more, nothing less.

Such as watch any of the Spirit Science videos, they are pseudoscience, yet are far more philosophical and rather than confuse you or bring up more questions, it just makes you want to know more.



None.

Aug 3 2012, 5:45 am Oh_Man Post #87

Emperor of The First

Quote from Vrael
Fits the carbon dioxide data well? I am much confused:


Well now, I wouldn't want you to be confused!
There are a number of misleading elements of this graph, I'll examine each of them in turn.

First, let's look at the temperature variation on the y axis. It goes from 4C all the way up to 18C, even though the lowest point the data ever gets is 7C, and the highest 11C. In effect, the graph has been 'zoomed out' to give the misleading appearance of having very little correlation to the CO2 emissions (I'll talk about that later). ANY set of data plotted on a graph will look like a fuzzy straight line if zoomed out enough! (Imagine zooming it out to 0-100C, it would look like a solid line, whereas the C02 emission graph would still look the same.) That's what the author of this graph was going for when feeding eager climate change deniers such as yourself.

Here is what it looks like zoomed in to its proper dimensions (7-11C):

SOURCE: http://en.wikipedia.org/wiki/Central_England_temperature

Next point, the annual CO2 emissions on the y axis. Couple of problems, first - you are comparing CENTRAL ENGLAND temperatures with GLOBAL CO2 emissions. Did you really expect a 1 to 1 correlation? \:lol\: Looking at one specific location to disprove global warming is a pointless venture. Each specific location has its own weather anomalies and yearly highs and lows that could obscure the temperature increase. For example, the Gulf Stream.
Further on that point, the graph is showing annual CO2 emissions, year-by-year whereas it is only accumulated CO2 emissions, (the amount of CO2 currently accumulated in the atmosphere) that contribute to global warming. And don't forget, the ocean currently absorbs approximately 60% of that emission (this is a whole OTHER problem, but not one that contributes to global warming at this time).

Next up, that green line through the data is pretty misleading. Of course, as I already mentioned, it looks pretty flat when zoomed out. In the zoomed in graph, you can see the slant more clearly. But really, you shouldn't be using linear regression to represent data with significant outliers (like, the massive temperature increase occurring recently!!). That line is devised from averages based on the entire data set, so of course it is not going to be representative as it ignores outliers. So of course, he states the trend is '0.26C per century". Well OF COURSE, for like three of those centuries there was zero emissions and therefore no significant temperature variation!!
You can make the trend show whatever you want if you pick different time ranges. The author of this graph included the entire time range and used a linear regression to make the temperature increase in the 1900s seem minimal. If you start the graph from when CO2 emissions first start accelerating, you'll see a very different and frightening trend-line, believe me.

Here is another graph with the exact same data, only instead of using linear regression it uses 11 year averages.

Now you're starting to see a correlation, though of course, as I stated before, the correlation between emissions and temperature is going to be a lot weaker because it isn't the emissions that are directly affecting temperature, it is accumulated C02 in the atmosphere!

Last point, that dotted red line at the top. "Climate model prediction". That isn't what scientists are predicting the temperature should be TODAY, it is what they are predicting the temperature will be 100 YEARS FROM NOW. It even says so in the yellow box: "The red dashed line represents recent climate model predictions of expected temperatures 100 years from now, supposedly due to CO2 emissions." Maybe you knew that, maybe you didn't. I just wanted to make sure.

That about wraps it up, here is some further reading for you:
http://www.metoffice.gov.uk/hadobs/hadcet/
http://en.wikipedia.org/wiki/Central_England_temperature
How to Lie with Statistics, Darrell Huff.


Now I don't expect you to stop being a climate change denier after this post. However, I DO expect you to stop using this graph as 'evidence' for you're erroneous beliefs.




Aug 3 2012, 6:39 am Vrael Post #88



I'm not sure you realize this, Oh_Man, but I am pro-climate change.

However, as your quarrel is with the reliability of the graph and not my belief, I suppose I should support it:

Quote
In effect, the graph has been 'zoomed out' to give the misleading appearance of having very little correlation to the CO2 emissions
The graph I posted has a range of 14 degrees celcius on the Y axis, your graph has 5. The difference is approximately 3 times as 'zoomed out' as you like, but since this doesn't affect the linear regression model of the actual data I think your point about the 'zoomedness' isn't very valuable. It clearly isn't flat.

Quote
Couple of problems, first - you are comparing CENTRAL ENGLAND temperatures with GLOBAL CO2 emissions. Did you really expect a 1 to 1 correlation?
Quote
Further on that point, the graph is showing annual CO2 emissions, year-by-year whereas it is only accumulated CO2 emissions, (the amount of CO2 currently accumulated in the atmosphere) that contribute to global warming. And don't forget, the ocean currently absorbs approximately 60% of that emission (this is a whole OTHER problem, but not one that contributes to global warming at this time).
These are of course, completely valid and worthwhile points. Sufficient sample size is necessary to make any sort of reliable assumptions about the larger population, and the data from even one reliable source could result in models which have little to no accuracy at predicting the global scale. As for your point about the total CO2 emissions, the integral of an exponential function is an exponential function, if my aim in providing the graph had been to 'disprove' global warming, naturally I should take the time to show a better model.

If you take notice however, the aim of my post was not to disprove global warming, but to show a source with actual data. The graph has the source of its data printed right on it, which anyone can go and download.

Quote
Next up, that green line through the data is pretty misleading. Of course, as I already mentioned, it looks pretty flat when zoomed out.
It says clearly on the graph that the slope is .26 degrees celcius per century.

Quote
But really, you shouldn't be using linear regression to represent data with significant outliers (like, the massive temperature increase occurring recently!!). That line is devised from averages based on the entire data set, so of course it is not going to be representative as it ignores outliers.
Linear regression is used to model linear data, regardless of outliers. If you like, I could run the linear regression myself and give you the correlation coefficient. Just looking at the graph, I'd say it'll be pretty close to 1. Though, you appear to be capable of it, perhaps you should do it yourself.

Quote
Here is another graph with the exact same data, only instead of using linear regression it uses 11 year averages.
This of course, is the key to establishing a true correlation. Is 11 years enough? The radical global warming claims involving temperature increases like the 100-year prediction on the graph in question are all based on trends like in your graphs containing data in only a decade or two. Consider if we had lived in 1730, where for the past 40 years the 11 year moving average had been steadily increasing. The model would be predicting approximately the same radical temperature increases as the 11-year moving average does now; but we can see in hindsight that the temperature returned to normal levels. I am not convinced at all by an 11 year moving average of anything which may have a typical period of hundred or thousands of years, as it is possible and perhaps even likely that the earth's climate change may be. 11 years is mathematically insignificant. Now, if 40 or 60 years from now, the temperature still continues to steadily rise at an accelerated pace, key term there being accelerated, consistent with measured CO2 levels, I'd say the model would definitely appear to have some credibility since the typical oscillation in the past 400 years has occured (approximately) somewhere between 2 and 4 decades.

Post has been edited 1 time(s), last time on Aug 3 2012, 7:25 am by Vrael.



None.

Aug 3 2012, 9:30 am Oh_Man Post #89

Emperor of The First

Quote from Vrael
The graph I posted has a range of 14 degrees celcius on the Y axis, your graph has 5. The difference is approximately 3 times as 'zoomed out' as you like, but since this doesn't affect the linear regression model of the actual data I think your point about the 'zoomedness' isn't very valuable. It clearly isn't flat.
Obviously it doesn't affect the data. I'm saying it is misleading. Graphs are used to eyeball data and examine trends. There is NO REASON for the temperature to be 'zoomed out' other than to misrepresent the data. The author is trying to make it look flatter by making it 3 times as smaller. As you can see by the graph I posted, there is a distinct increase in temperature. Just look at the centuries BEFORE increased CO2 emissions, then compare to AFTER.


Quote from Vrael
These are of course, completely valid and worthwhile points. Sufficient sample size is necessary to make any sort of reliable assumptions about the larger population, and the data from even one reliable source could result in models which have little to no accuracy at predicting the global scale. As for your point about the total CO2 emissions, the integral of an exponential function is an exponential function, if my aim in providing the graph had been to 'disprove' global warming, naturally I should take the time to show a better model.

If you take notice however, the aim of my post was not to disprove global warming, but to show a source with actual data. The graph has the source of its data printed right on it, which anyone can go and download.
Sure, far be it from me to stop you posting data about global warming. But use a proper graph like the ones I gave, not one drawn up by a climate change skeptic with the intent to deceive.

Quote from Vrael
Quote
Next up, that green line through the data is pretty misleading. Of course, as I already mentioned, it looks pretty flat when zoomed out.
It says clearly on the graph that the slope is .26 degrees celcius per century.
As I said in the previous post, you can make it say whatever you want when arbitrarily choosing your own time frame. The further back in time you go, the lower that slope would get. If you're trying to pinpoint a correlation between CO2 emissions and temperature variation, you don't tack on hundreds of years of data where there were ZERO CO2 emissions!

Quote from Vrael
This of course, is the key to establishing a true correlation. Is 11 years enough? The radical global warming claims involving temperature increases like the 100-year prediction on the graph in question are all based on trends like in your graphs containing data in only a decade or two. Consider if we had lived in 1730, where for the past 40 years the 11 year moving average had been steadily increasing. The model would be predicting approximately the same radical temperature increases as the 11-year moving average does now; but we can see in hindsight that the temperature returned to normal levels. I am not convinced at all by an 11 year moving average of anything which may have a typical period of hundred or thousands of years, as it is possible and perhaps even likely that the earth's climate change may be. 11 years is mathematically insignificant. Now, if 40 or 60 years from now, the temperature still continues to steadily rise at an accelerated pace, key term there being accelerated, consistent with measured CO2 levels, I'd say the model would definitely appear to have some credibility since the typical oscillation in the past 400 years has occured (approximately) somewhere between 2 and 4 decades.
In 1730 it went back down in about a decade. LOOK at these graphs, this has been going on for decades. The average hasn't dropped below 9 in AGES. You don't need another 40 or 60 years to finally make up your mind:




Stop digging your trench. Just never use that graph again.

Post has been edited 1 time(s), last time on Aug 3 2012, 9:37 am by Oh_Man.




Aug 4 2012, 4:15 am Vrael Post #90



Quote
There is NO REASON for the temperature to be 'zoomed out' other than to misrepresent the data.
Yeah! And they definitely shouldn't have used green for the arrow, I mean so many people associate "green" with "good" or "go" that they're definitely playing psychological tricks on you to get you to believe everything is "good."!!!!! These fiends!! They're evil manipulators!!

Joking, obviously, but I'm merely saying it isn't zoomed out to the point where its in any way misrepresentative. I've seen cases where graphs are horribly distorted like you describe, and I just don't think this is one of them. Maybe it is, but unless you have something revolutionary to say about this point I'm done with it.

Quote
But use a proper graph like the ones I gave, not one drawn up by a climate change skeptic with the intent to deceive.
What's the old saying? If you go looking for worms you'll find worms, or something like that. Point being, you started this conversation based on the premise I was a "climate change denier" using "misrepresented" data, and with the goal to disprove me, rather than analyze what I put forth for its own merits.

Quote
The further back in time you go, the lower that slope would get. If you're trying to pinpoint a correlation between CO2 emissions and temperature variation, you don't tack on hundreds of years of data where there were ZERO CO2 emissions!
The slope could just as easily grow much larger, with a few hundred more years of data. The prior hundreds of years of data is absolutely necessary in determining the effect of CO2 on the global climate. I would have thought you'd know this, but the statistical correlation is derived as follows:
1). We have a set of data pertaining to a situation (climate change in this case)
2). From this data, we can establish normal parameters like variance, covariance, mean, etc, about this data.
3). We have a new set of data pertaining to the situation which falls outside the established parameters
=> Since everything except CO2 levels has remained constant, the CO2 is responsible for the variation from established trends.

The strength of this inference is based upon the strength of the data, 1) and 3). The stronger and more normalized your initial data set, 1), and the more variant your 3) data set is, the stronger the implication that your hypothesis was correct (unless you're really wrong and theres some other force that is responsible, or your model is incorrect or something like that). Perhaps I have assumed you had some basic knowledge of statistical properties, but if not, I encourage you to read about the Law of Large Numbers, which essentially forms the basic principle on which these kinds of inferences are made. (Which in essence I demonstrate below).

Quote
The average hasn't dropped below 9 in AGES. You don't need another 40 or 60 years to finally make up your mind
The 11-year simple moving average hasn't dropped below 9 degrees celcius in approximately 1.2 centuries, correct, but I am baffled as to why you're choosing an 11-year-moving-average as the basis for all your statistical claims. I'm also not sure why you are assuming 1.2 centuries qualifies as "AGES", since we're talking about temperature cycles, and not how long its been since you watched your favorite movie. Back to the 11 year moving average, Let's consider by analogy: Suppose I had a bucket, and I told you that if you reached into it, one of two things would happen: Either you would pull out a 10-lb ball of pure gold (I'm sorry, 44.48 Newtons for you metric users \:P ) or your arm would get chopped off. Now I have my friend joe reach in first, and he pulls out a 10-lb ball of gold. Are you going to want to stick your arm in there? No! (Unless you're not a risk-adverse person at all maybe). There hasn't been sufficient sampling done yet to establish any sort of reliable trend or probability for either of the two options. Maybe joe got lucky and got the .0001% chance who would pull out gold. Maybe you say to me "No way! what if the chance is .0001% of pulling gold out?" So I have my friend Ted come over, and he pulls out a ball of gold too. Now, this doesn't change the probability of pulling gold out, but it does change the probability that the unknown gold-probability is extremely low. The more times we can reach in the bucket and record the outcomes, the more sure we can be about the probabilities of those outcomes.

My point, is that even your 11-year-moving average has a period (by which I mean period of periodic functions, i.e. period of sin(x) = 2*pi) of approximately 4 decades (I'm estimating purely by eye, but I think you get my drift). So in that 350 year time frame, that means we have roughly 9 data points. One of those points is what we are trying to come to a conclusion about, so we are comparing our data to 8 other points. This is not sufficient, in a rigorous sense, to establish a trend. However, its all we have, so we may as well use it all in making our predictions. This is why you add those extra 2 centuries of data on to the analysis, because without them, the evidence for "CO2 is causing global warming" is half gone. In essence, Oh_Man, this:
Quote
If you're trying to pinpoint a correlation between CO2 emissions and temperature variation, you don't tack on hundreds of years of data where there were ZERO CO2 emissions!
Is dead wrong from a statistical standpoint.

That concludes my defense of the graph, but returning to the year 1730, I wonder if the volcanic activity had not happened, would the trend from 1690-1730 have continued? If we had another 400 years of data I think that would be a very significant piece of the puzzle to analyzing our own climate change problem.

Also, what I wrote above is one reason the people from Lanthanide's link are using 1.6 billion data sets.



None.

Aug 4 2012, 5:11 am Lanthanide Post #91



Quote from Vrael
but I am baffled as to why you're choosing an 11-year-moving-average as the basis for all your statistical claims
For starters, I don't think Oh-Man created any of the graphs he presented, in fact it looks like he got the content of that post from a website somewhere that directly rebuts the graph you posted.

Anyway, the 11-year moving average is because the sunspot cycle is 11 years. By having an 11 year moving average, it ensures that a full sun-spot cycle, from any arbitrary start to end point, is included in each graphed data point.


All of this would be better placed in the Environmental Issues thread in Serious Discussion, here: http://www.staredit.net/topic/14446/1/ Unfortunately though Nuderaider locked it under the mistaken impression that "noone cares anymore", so we'll need a mod to unlock it so we can discuss this stuff in the much more appropriate thread.



None.

Aug 5 2012, 8:12 am Oh_Man Post #92

Emperor of The First

You did not need to write so much in response to so little, Vrael. I admit I could have phrased it better, but this sentence:
Quote
The further back in time you go, the lower that slope would get. If you're trying to pinpoint a correlation between CO2 emissions and temperature variation, you don't tack on hundreds of years of data where there were ZERO CO2 emissions!

was supposed to be read as a follow on from the previous sentence:
Quote
As I said in the previous post, you can make it say whatever you want when arbitrarily choosing your own time frame. The further back in time you go, the lower that slope would get.
not as its own one-sentence statement.

I wasn't making so silly a point as the one you eagerly jumped on. I was stating that you can't pinpoint a correlation between CO2 emissions and temperature variation using a linear regression with hundreds of years of data where there was no CO2 emissions (and therefore no significant temperature variation). All that extra data pulls the line down and masks the outliers (the rise in temperature due to CO2 emission).

Furthermore, when I said
Quote
The further back in time you go, the lower that slope would get.
I meant lower as in closer to flat, (as per the law of large numbers) not lower as in past flat and into a negative slope. Again, I could have phrased this more clearly.

Quote from Vrael
What's the old saying? If you go looking for worms you'll find worms, or something like that. Point being, you started this conversation based on the premise I was a "climate change denier" using "misrepresented" data, and with the goal to disprove me, rather than analyze what I put forth for its own merits.
Hardly, you said so yourself in your previous post:
Quote from Vrael
I'm not sure you realize this, Oh_Man, but I am pro-climate change.

However, as your quarrel is with the reliability of the graph and not my belief, I suppose I should support it:
My initial and subsequent posts have all been on elucidating why that graph is deliberately misleading. I called you a climate change denier because that is what I thought you were at the time, but nothing I said really relied on that - it was all about demonstrating the problems with that graph.

I still can't really understand how you ARE pro climate change to be honest, you seem entirely the opposite to me, and that graph comes from a climate change denialist's website!

Quote from Vrael
Joking, obviously, but I'm merely saying it isn't zoomed out to the point where its in any way misrepresentative. I've seen cases where graphs are horribly distorted like you describe, and I just don't think this is one of them. Maybe it is, but unless you have something revolutionary to say about this point I'm done with it.
It is zoomed out three times larger then it needs to be. Obviously, this is done to attempt to mask temperature variation, ESPECIALLY when he has plotted the carbon emissions graph on the same y axis. As I've said, there is NO REASON the temperature range is set that high OTHER than to mask temperature variation. Are you going to defend against this point? Or accept it?

In addition, I also mentioned how it was further misleading to try and compare temperature variation to CO2 emissions, when it is accumulated CO2 in the atmosphere that is actually responsible for the correlation. You agreed with this.

Lastly, I explained how using a linear regression was further misleading as it ignores outliers by tacking on hundreds of years of data with no CO2 emissions, effectively pushing the slope of the linear regression down. The green line was further given the appearance of flatness by spreading out the y axis for temperature, compare the two graphs yourself to confirm. I substituted this with a graph using 11 year moving averages (Lanth has defended this for me) which gives a far better representation of possible correlation then a linear regression does.

If you are truly pro climate change, you are doing yourself and the movement a disservice by continually propagating this deliberately misleading graph, created by a known climate change denier.

I'd also like to get other people's thoughts on this. Surely you can see the graph was created to mislead?

Post has been edited 1 time(s), last time on Aug 6 2012, 10:53 am by Oh_Man. Reason: fixed typo




Options
Pages: < 1 3 4 5
  Back to forum
Please log in to reply to this topic or to report it.
Members in this topic: None.
[03:44 am]
jjf28 -- there we go
[03:43 am]
O)FaRTy1billion[MM] -- Let's just make AIs that integrate with humans to become indistinguishable
[03:35 am]
jjf28 -- you know, if they choose to
[03:34 am]
jjf28 -- how bout we build AIs with the goal of continually raising human intelligence (through means such as genetic engineering, advanced teaching techniques, brain-to-technology integration); thus keeping humans near-on-par with machine intelligence; keeping humans as relevant thinking agents
[03:31 am]
jjf28 -- hmm I didn't think of this earlier...
[02:54 am]
O)FaRTy1billion[MM] -- You're a Harry, wizard.
[02:38 am]
BloodyZombie117 -- You're a wizard Harry.
[02:37 am]
Dem0n -- Son my credit card is so good. Lemme get a loan. I'll pay it back.
Please log in to shout.


Members Online: jjf28, O)FaRTy1billion[MM], lil-Inferno, Roy