• Welcome, guest!

    This is a forum devoted to discussion of Wolverhampton Wanderers.
    Why not sign up and contribute? Registered members get a fully ad-free experience!

Climate Change Debate

The basic science is very compelling; it relies on four basic facts. Firstly global CO2 levels have been increasing steadily since the start of the industrial revolution. Secondly we can analyse atmospheric CO2 and determine that its from burning fossil fuels (basically carbon dating tells us that its 'old' carbon, rather than new carbon from things like plant respiration), so we know its man-made. Thirdly we know that CO2 is a strong absorber of IR radiation, which is the radiation that is emitted from the earth. Fourthly we know that the heat content of the earth is rising continuously. Its getting warmer.

(I dont want to overload this post, so wont cite the raw materials here, but can do this later if you want)

The problem with the denialist POV is that you need to essentially refute at least one of these tenets. And no-one has managed that yet.

So we have a cause, an effect, and a mechanism to tie the two together. The data that is altered by scientists is really the data that we use to quantify how much warming there is likely to be. In other words the basic premise that mankinds activities are warming the planet is not predicated on the adjustment of the data. The data is a tool to help us quantify how much, and its this area that is the focus for current research - using more accurate measurements, more powerful simulations, models that incorporate a wider range of potential effects etc.

If scientists just used unadjusted data then the accuracy of their conclusions would be detrimentally affected.
 
Visage said:
If scientists just used unadjusted data then the accuracy of their conclusions would be detrimentally affected.
Click to expand...

You might not believe this, but I'm honestly interested and not just trying to be arguamentative here - but are you saying you can only prove the views if you alter the data?
 
No, im saying that if you have inaccurate data then you get inaccurate results. Let me give you an example. The output from the Sun varies constantly, and that causes the atmosphere to change size - it warms up, and expands, or cools down and contracts. So if you've got a satellite thats in orbit and looking down measuring surface temperatures then the number it reads will change over time, even if the actual surface temperature is constant.

But if you know that this expanding and contracting is happening then you can adjust the data - you can seperate the signal from the noise.

Theres a discussion here: http://arstechnica.com/staff/2015/02/temperature-data-is-not-the-biggest-scientific-scandal-ever/
 
To use a social science example, think about a study examining socioeconomic status. On the face of the study it will appear clear that there is great variability of income no matter what sort of person you are. This is what we would call "unadjusted" data. It tells us much, but it isn't the whole story.

Now, when you adjust the data to control for variables such as race, a much more accurate portrait of socioeconomic disparity emerges. But no scientist in their right mind would look at such variable control and say "you've adjusted this data and therefore it is unreliable!" Quite the contrary is true; by adjusting the data you've gleaned an entirely different and valid perspective on class disparity that ultimately results in new (more accurate) results.
 
Sorry, I don't get it - surely you would take the data by race to give you your racial evidence, not adjust the data to reflect what you think is happening. You have actual data, Amending it means it's not actual data any more, it's interpreted data, through your lens of how it should look.

There's a difference between adjusting data, and splitting the data into smaller segments.
 
But that's just it, you're not changing the data. You're simply controlling for what we call confounding variables.
 
Fair enough, it's clearly normal practice in your circles - I don't work in a scientific environment but do work with data - in my world changing the numbers isn't interpretive, it's falsifying. In my world you would leave the numbers but explain the disparity.

I do get that these things are treated differently in different fields though, that's not meant disbelievingly.
 
Premier Quality said:
Sorry, I don't get it - surely you would take the data by race to give you your racial evidence, not adjust the data to reflect what you think is happening. You have actual data, Amending it means it's not actual data any more, it's interpreted data, through your lens of how it should look.

There's a difference between adjusting data, and splitting the data into smaller segments.
Click to expand...

It depends on how you do it. If you say 'We think that the data is skewed because of X, so we 'un-X'd' it, and the resultant data looked like this then thats fine. If you say 'The raw data doesnt match what we expected so we altered it until it fit' then thats wrong.

Its vitally important to explain what you did and why you did it. Any reputable scientist will do that.
 
Premier Quality said:
Fair enough, it's clearly normal practice in your circles - I don't work in a scientific environment but do work with data - in my world changing the numbers isn't interpretive, it's falsifying. In my world you would leave the numbers but explain the disparity.

I do get that these things are treated differently in different fields though, that's not meant disbelievingly.
Click to expand...

Well, again, the numbers aren't being changed. Just looked at differently.

A football analogy: you can make judgments about a performance based on the oppositions number of shots, but it is probably better to look at the number of shots on target instead.
 
So in the article that started this, the raw data shows that area has been getting colder over that time period. The revised data shows it's getting warmer over that time period. I don't see how that can happen without changing the numbers.
 
Alan USA said:
Well, again, the numbers aren't being changed. Just looked at differently.

A football analogy: you can make judgments about a performance based on the oppositions number of shots, but it is probably better to look at the number of shots on target instead.
Click to expand...

I think you're talking about something different here Alan.

By the sounds of things from Vis the temperature readings and what not are actually altered but done so to counteract natural variation that disrupts proper recording whereas you seem to be talking about filtering certain results out of a sample size in order to display a trend in a sub-section of the original sample.

I guess to translate it across to your social science example if you were studying say income relative to age. In Vis's method you might want to alter the actual income of people based in areas like London which get a fair uplift compared to areas further North so that everyone is on a level playing field, then you can see if there is a disparity between income across different age groups, in your method you'd just ignore people from extreme ends of the spectrum and focus on one geographical area.
 
Just because it made me smile Alan and to do my own football analogy - I see this as wolves got
Relegated under dean Saunders - his interpretation of the data is that we were fucking awesome and unlucky not to win the league!
 
Premier Quality said:
So in the article that started this, the raw data shows that area has been getting colder over that time period. The revised data shows it's getting warmer over that time period. I don't see how that can happen without changing the numbers.
Click to expand...

The numbers are changed.

Put it this way, if every monitoring station in Peru was showing warming, but one station is showing cooling, would you trust that data?
 
MARKakaJIM said:
I think you're talking about something different here Alan.

By the sounds of things from Vis the temperature readings and what not are actually altered but done so to counteract natural variation that disrupts proper recording whereas you seem to be talking about filtering certain results out of a sample size in order to display a trend in a sub-section of the original sample.

I guess to translate it across to your social science example if you were studying say income relative to age. In Vis's method you might want to alter the actual income of people based in areas like London which get a fair uplift compared to areas further North so that everyone is on a level playing field, then you can see if there is a disparity between income across different age groups, in your method you'd just ignore people from extreme ends of the spectrum and focus on one geographical area.
Click to expand...

Not exactly; as a scientist (even a social scientist) you never ignore any data. Controlling for certain variables doesn't mean throwing any data out, it just means you're looking at the data in slightly more specific ways. This doesn't necessarily disagree with Vis' method: if you're looking at income levels in different geographic areas then it makes sense to adjust for things such as cost of living. In that example, cost of living is the confounding variable that must be controlled for.
 
Alan USA said:
Not exactly; as a scientist (even a social scientist) you never ignore any data. Controlling for certain variables doesn't mean throwing any data out, it just means you're looking at the data in slightly more specific ways. This doesn't necessarily disagree with Vis' method: if you're looking at income levels in different geographic areas then it makes sense to adjust for things such as cost of living. In that example, cost of living is the confounding variable that must be controlled for.
Click to expand...

So how exactly would you change individual data in such an experiment to give a clearer indication of disparity between races, as per your example?
 
Visage said:
The numbers are changed.

Put it this way, if every monitoring station in Peru was showing warming, but one station is showing cooling, would you trust that data?
Click to expand...

Of course not, but by the same token, he's saying all the stations in that area are
Showing the same thing, but they've all been 'adjusted' to say something different. If that's the case, surely we would both say the changes are wrong?
 
MARKakaJIM said:
So how exactly would you change individual data in such an experiment to give a clearer indication of disparity between races, as per your example?
Click to expand...

The simplest way would likely be to determine how far a dollar (or pound) goes in one city and then adjust accordingly in the other cities. So, let's assume that we're using London as our baseline, i.e., £1 in London is our constant. We can then look at such variables as average cost of goods and services in other cities (such as Birmingham) and come to the theoretical conclusion that one pound in London equals two pounds in Birmingham. You can then adjust all incomes in the sample by this rate in order to control for the cost of living.

So, in this example, if you make £50,000 annually in London, you're equal to someone making £25,000 annually in Birmingham. Does that make sense?
 
Alan USA said:
The simplest way would likely be to determine how far a dollar (or pound) goes in one city and then adjust accordingly in the other cities. So, let's assume that we're using London as our baseline, i.e., £1 in London is our constant. We can then look at such variables as average cost of goods and services in other cities (such as Birmingham) and come to the theoretical conclusion that one pound in London equals two pounds in Birmingham. You can then adjust all incomes in the sample by this rate in order to control for the cost of living.

So, in this example, if you make £50,000 annually in London, you're equal to someone making £25,000 annually in Birmingham. Does that make sense?
Click to expand...

I think i've misunderstood your first post.

That's exactly what i'd do in my hypothetical income/age experiment, for some reason anticipated different shit going down for your race/income scenario whereas you can easily just substitute race for age as the yardstick that you're measuring against and you've got the same geographical swing to try and eliminate.

Now i remember how much i hated my dissertation, trying to collect appropriate data and justify my bullshit theory, worst thing i've ever done.
 
Premier Quality said:
Of course not, but by the same token, he's saying all the stations in that area are
Showing the same thing, but they've all been 'adjusted' to say something different. If that's the case, surely we would both say the changes are wrong?
Click to expand...

I go back to my original point. If the numbers are skewed to meet an agenda then that would show up in the publication. The authors will always say the raw data was blah, and the processed data was blah, and the way the processing that was done is blah.

And if the processing was unjustified, or done incorrectly, then the paper would fail peer review.
 
You must log in or register to reply here.
Back
Top