LATimes: Berkeley Earth Surface Temperature team's preliminary findings

[Winehole23]

A team of UC Berkeley physicists and statisticians that set out to challenge the scientific consensus on global warming is finding that its data-crunching effort is producing results nearly identical to those underlying the prevailing view.

http://www.latimes.com/news/local/la...0,772697.story

[LnGrrrR]

I'll just answer the question that I know WC will bring up:

Over the years, Muller has praised Watts' efforts to show that weather station data in official studies are untrustworthy because of the urban heat island effect, which boosts temperature readings in areas that have been encroached on by cities and suburbs.

But leading climatologists said the previous studies accounted for the effect, and the Berkeley analysis is confirming that, Muller acknowledged. "Did such poor station quality exaggerate the estimates of global warming?" he asked in his written testimony. "We've studied this issue, and our preliminary answer is no."

[DarrinS]

This is important work; however, I don't think they should draw any conclusions with only 2% of the station data.

[LnGrrrR]

This is important work; however, I don't think they should draw any conclusions with only 2% of the station data.

They haven't technically drawn any "conclusions"... just that the data they've seen so far matches up with the general consensus. They stated in the article that their opinion might change when more data comes in.

*cue boutons saying their opinion will change once the Koch brothers get wind of this and threaten to cut off funding*

[DarrinS]

They haven't technically drawn any "conclusions"... just that the data they've seen so far matches up with the general consensus. They stated in the article that their opinion might change when more data comes in.

*cue boutons saying their opinion will change once the Koch brothers get wind of this and threaten to cut off funding*

"findings"

http://berkeleyearth.org/findings

[Winehole23]

That's why they're called preliminary findings. Maybe if the CA GOP hadn't been so eager to hear them they wouldn't have gotten so much press.

[ChumpDumper]

"findings"

http://berkeleyearth.org/findings

The hearing was called by GOP leaders of the House Science & Technology committee

[TeyshaBlue]

lolz. pqnd.

[DarrinS]

That's why they're called preliminary findings. Maybe if the CA GOP hadn't been so eager to hear them they wouldn't have gotten so much press.

Don't get me wrong, it's a very important project, I just think their findings need to be given weight commensurate with 2% of station data.

[ChumpDumper]

Don't get me wrong, it's a very important project, I just think their findings need to be given weight commensurate with 2% of station data.

So how many points of raw data are contained in 2% of station reports, Darrin?

Give me a rough estimate. You're claiming some specific statistical knowledge here.

[DarrinS]

So how many points of raw data are contained in 2% of station reports, Darrin?

Give me a rough estimate. You're claiming some specific statistical knowledge here.

Do you think 2% is large?

[ChumpDumper]

Do you think 2% is large?

You didn't answer the question.

The answer was in the article you didn't read.

Again.

Damn, you're stupid.

[ChumpDumper]

Tell us all, Darrin: would you consider the corroboration of a million measurements significant?

[DarrinS]

Tell us all, Darrin: would you consider the corroboration of a million measurements significant?

You should contact these people and tell them how dumn they are.

A preliminary analysis of 2% of the Berkeley Earth dataset shows a global temperature trend that goes up and down with global cycles, and does so broadly in sync with the temperature records from other groups such as NOAA, NASA, and Hadley CRU. However, the preliminary analysis includes only a very small subset (2%) of randomly chosen data, and does not include any method for correcting for biases such as the urban heat island effect, the time of observation bias, etc.

[MannyIsGod]

It depends on where the 2% is. If the 2% if all in one place, obviously its not indicative. If its spread out properly, then its not terrible information but it is incomplete.

Using 2% out of context is perfectly your style, Darrin.

[ChumpDumper]

Tell us all, Darrin: would you consider the corroboration of a million measurements significant?

[DarrinS]

It depends on where the 2% is. If the 2% if all in one place, obviously its not indicative. If its spread out properly, then its not terrible information but it is incomplete.

Using 2% out of context is perfectly your style, Darrin.

Post above yours is context.

[MannyIsGod]

Yeah - but why didn't you provide that on your first post?

[Spurminator]

2% = not a big enough sample

"It's cold today" = GLOBAL WARMING PWNED!!!

[MannyIsGod]

I'd also say its very telling that a randomly selected 2% is echoing the other complete datasets. It could definitely change, but whats the likelyhood of it randomly matching the curves of several other data sets? Its not very high of course.

In fact I would say those odds are far less than 50:1, or you know, 2%.

[CuckingFunt]

2% = not a big enough sample

"It's cold today" = GLOBAL WARMING PWNED!!!

I guffaw'd.

[ChumpDumper]

I'd also say its very telling that a randomly selected 2% is echoing the other complete datasets. It could definitely change, but whats the likelyhood of it randomly matching the curves of several other data sets? Its not very high of course.

In fact I would say those odds are far less than 50:1, or you know, 2%.

Had Darrin actually read the article, he'd know how many random data points were matched.

He continues to consider his ignorance to be a coat of armor.

[boutons_deux]

Repugs want to cut $500M from IRS budget, so one can expect the auditing rates to go even lower.

[DarrinS]

political pollsters survey far less than 2% of the population to determine, generally, very accurate actual results. if the sample pop is really random, then you don't need that many

what % of taxpayers does the IRS audit? http://www.wwwebtax.com/audits/audit_avoiding.htm

don't know the specifics of the UC-B study, but just because 2% sounds tiny doesn't mean it's utterly insignificant

In some cases, you might be justified in using small samples, e.g. Poisson sampling for pass/fail. Likewise, with polls where the outcomes are small, discrete sets. Temperature is different animal.

[DarrinS]

Repugs want to cut $500M from IRS budget, so one can expect the auditing rates to go even lower.

good