Yes, perhaps I phrased my comment poorly. Havenwar also interpreted my post as such (see the other thread/reply). All I meant was to draw attention to the distinction between significance and the size of the effect. With a large sample size, you can find significance, but it might not be a large/important effect.
I'm also not familiar enough with the US elections, but even so, this still seems like a minimal effect to me. Firstly, there is no partisan information gathered, merely the result that *more* people voted. (Presumably this would mean more Democrat votes, but how much? Perhaps no more than 0.1% total??) Secondly, the implication of the article (to me at least) is a more generalised view, that one can influence one's internet friends' habits. If I spent time haranguing my friends about something, I'd hope to affect more than one friend for every 250 that I harass!
As for my critique of statistical significance versus effect size, you may have a point. Perhaps I should have been more explicit, but I still stand by the sentiment. I think that "statistical significance" means something to most people that is agnostic of large sample sizes. That is, if I read that something is "statistically significant", with no more information given, I presume that it's using an "average" amount of samples. I'm not sure what your background is, but if a study does not achieve statistical significance, a common way to achieve this is by increasing the sample size. If that is the case, then it's useful to also quantify the size of the effect, in order to give context to this "significance".
But you are right. In my original comment, I meant to suggest that a large sample size could reveal a *true* difference, even if it were quite minor in the magnitude of effect. I did not mean to suggest that a large sample size could create false artefacts.
It was not meant to be a dig at statistics per se, but more a critique at differentiating between statistical significance and size of the effect. What I meant to imply was that while I agree the effect was significant, the size of the effect was relatively small. Also, I'm not 100% sure what you mean. I agree that having a small sample size results in a non-accurate result (technically a non-precise result). Hence, it results in less statistical significance than large sample sizes. At least for a simple hypothesis test of a proportion, as n increases, the Z-test statistic also increases in magnitude, corresponding with a drop in the p-value.
It's not even a very large effect. From TFA "People who received messages alerting them that their friends had voted were 0.39% more likely to vote than those who received messages with no social information". Get a sample group large enough (61 million users), and you'll find many things to be statistically significant.
Nowhere does it say they will listen in to your conversations. From TFA, "We may use non-personally identifiable demographic information (e.g. location, gender and age) to target ads."
If you rarely use Windows, you could partition your disk, and format the Windows part to FAT. Then, on the odd occasion where you want to transfer files, you could just mount its partition in Linux (that's what I do), or OS X. I like the idea of sandboxing Windows, and not letting it touch/corrupt my main (ext4) file system.
Yeah, I hadn't checked the original paper either, but (as you might know) often these commentaries try to emphasise the exciting part of the paper anyway!
What a ridiculously sensationalist article! From TFA: "the researchers could not tell which traits were being selected for". So basically the results are that some people reproduced and some did not. Hence we infer that some were genetically predisposed to be fitter, whether through resistance to disease or sexual attractiveness or something else. Well, isn't that obvious? I'm sure this still happens today!
I haven't used Tasker, but the free "Llama - Location Profiles" can detect location by cell towers (hence it doesn't need GPS or Wifi to be on), then can turn off wifi, etc. It can also turn volume down at night, etc. (similar to Tasker, I gather).
Despite the claims on their website that "Transparency Life Sciences is the world’s first drug development company based on open innovation", the Open Source Drug Discovery group (based in India) has been doing this for over three years now, focusing on tuberculosis. I'm also aware of "Open Source Drug Discovery for Malaria".
From TFA, the reason why the gas was leaving the planet was because of it's small size. Presumably, the (similary gaseous) atmosphere of Earth doesn't float off into space because the gravity here is much higher.
The idea that an ancient cephalopod would arrange objects into patterns — even unintentionally — is highly speculative, says Roger Hanlon, a marine biologist at the Marine Biological laboratory in Woods Hole, Massachusetts. "There's nothing in the scientific literature that suggests that modern-day cephalopods do anything like this," he notes.
In their natural environment, octopuses do sometimes grab shiny objects, says James Wood, an expert in cephalopod behaviour and associate director at the Waikiki Aquarium in Hawaii. "They're curious creatures, and they certainly do manipulate their environment," he notes. But while octopuses may pile up rocks outside the mouth of their sea floor dens, for example, they are not known to bring animal remains home. "I've yet to find a vertebrate bone in a cephalopod midden."
Good point! I'm not sure if you are talking about black smokers? If so, I'm pretty sure that these ecosystems still rely on oxygen (along with carbon dioxide and hydrogen sulfide) to fix carbon. Again, this oxygen originally derived from photosynthesis, requiring sunlight. Nevertheless, I seem to recall some underwater ecosystems that rely on the weak geothermal light for photosynthesis, which may be more relevant.
Ultimately, these animals are still dependent on a source of energy such as a sun. These animals are not capable of capturing such energy themselves, but rely on photosynthetic organisms such as plants or algae at the bottom of the food chain. As another poster pointed out, carbon dioxide is necessary for this photosynthesis, not oxygen. This is converted into sugar, which animals subsequently consume. Hydrogenosomes function similarly to mitochondria in converting sugar into ATP (more accessible energy), but differ in that they produce (surprise, surprise) hydrogen as a byproduct.
From the article: "He also found that slum dwellers there collected their excrement in a plastic bag and disposed of it by flinging it He plans to sell it for about 2 or 3 cents — comparable to the cost of an ordinary plastic bag."
Unintuitively, the accuracy of the estimate of a proportion has little to do with the size of the population. The accuracy relies more on the size of the sample. I'm sure there are more sophisticated ways to do this, but from my first year statistics course, the standard error is ((p(1-p))/n), where p (estimated proportion) = 87/200 and n (sample size) = 200. Hence, the standard error = 0.035 and the 95% confidence interval is 0.435 ± 0.069.
i.e. we are 95% confidence that the true proportion lies between 36.7% and 50.4%. That's still quite high.
I think the "easygoing" nature of Aussies is the problem. Our newspapers and other media are a disgrace. We are totally clueless as to international affairs. Sport is always the number one story. Conversely, we don't suffer from as much social unrest as other countries, but sometimes this is frustrating, as I feel this may be helpful.
The green colour of chlorophyll relates to the historical precursors to the first photosynthetic organisms. Originally (a few billion years ago), early bacteria were non-photosynthetic, fermenting carbon anaerobically. This rapidly depleted the primordial "soup." The first organisms to utilise light (something akin to Halobacterium halobium) used a pigment called bacteriorhodopsin to help its metabolism. Bacteriorhodopsin absorbs a central band of visible light.
The evolution of chlorophyll followed (perhaps in Cyanobacteria) in organisms at the bottom of the sea. These were the first organisms to fix carbon dioxide. Being at the bottom of the ocean, only the far bands of visible light were available to them (blue and red), and hence green chlorophyll evolved.
Since then, accessory pigments have also evolved (e.g. phycobiliproteins), which have reclaimed other parts of the visible spectrum, and changed the colour of the plants or algae.
Slashdot Mirror
Yes, perhaps I phrased my comment poorly. Havenwar also interpreted my post as such (see the other thread/reply). All I meant was to draw attention to the distinction between significance and the size of the effect. With a large sample size, you can find significance, but it might not be a large/important effect.
I'm also not familiar enough with the US elections, but even so, this still seems like a minimal effect to me. Firstly, there is no partisan information gathered, merely the result that *more* people voted. (Presumably this would mean more Democrat votes, but how much? Perhaps no more than 0.1% total??) Secondly, the implication of the article (to me at least) is a more generalised view, that one can influence one's internet friends' habits. If I spent time haranguing my friends about something, I'd hope to affect more than one friend for every 250 that I harass!
As for my critique of statistical significance versus effect size, you may have a point. Perhaps I should have been more explicit, but I still stand by the sentiment. I think that "statistical significance" means something to most people that is agnostic of large sample sizes. That is, if I read that something is "statistically significant", with no more information given, I presume that it's using an "average" amount of samples. I'm not sure what your background is, but if a study does not achieve statistical significance, a common way to achieve this is by increasing the sample size. If that is the case, then it's useful to also quantify the size of the effect, in order to give context to this "significance".
But you are right. In my original comment, I meant to suggest that a large sample size could reveal a *true* difference, even if it were quite minor in the magnitude of effect. I did not mean to suggest that a large sample size could create false artefacts.
It was not meant to be a dig at statistics per se, but more a critique at differentiating between statistical significance and size of the effect. What I meant to imply was that while I agree the effect was significant, the size of the effect was relatively small. Also, I'm not 100% sure what you mean. I agree that having a small sample size results in a non-accurate result (technically a non-precise result). Hence, it results in less statistical significance than large sample sizes. At least for a simple hypothesis test of a proportion, as n increases, the Z-test statistic also increases in magnitude, corresponding with a drop in the p-value.
It's not even a very large effect. From TFA "People who received messages alerting them that their friends had voted were 0.39% more likely to vote than those who received messages with no social information". Get a sample group large enough (61 million users), and you'll find many things to be statistically significant.
Nowhere does it say they will listen in to your conversations. From TFA, "We may use non-personally identifiable demographic information (e.g. location, gender and age) to target ads."
No. But I think that's the whole point of Godwin's law.
If you rarely use Windows, you could partition your disk, and format the Windows part to FAT. Then, on the odd occasion where you want to transfer files, you could just mount its partition in Linux (that's what I do), or OS X. I like the idea of sandboxing Windows, and not letting it touch/corrupt my main (ext4) file system.
Yeah, I hadn't checked the original paper either, but (as you might know) often these commentaries try to emphasise the exciting part of the paper anyway!
What a ridiculously sensationalist article! From TFA: "the researchers could not tell which traits were being selected for". So basically the results are that some people reproduced and some did not. Hence we infer that some were genetically predisposed to be fitter, whether through resistance to disease or sexual attractiveness or something else. Well, isn't that obvious? I'm sure this still happens today!
I haven't used Tasker, but the free "Llama - Location Profiles" can detect location by cell towers (hence it doesn't need GPS or Wifi to be on), then can turn off wifi, etc. It can also turn volume down at night, etc. (similar to Tasker, I gather).
Despite the claims on their website that "Transparency Life Sciences is the world’s first drug development company based on open innovation", the Open Source Drug Discovery group (based in India) has been doing this for over three years now, focusing on tuberculosis. I'm also aware of "Open Source Drug Discovery for Malaria".
From TFA, the reason why the gas was leaving the planet was because of it's small size. Presumably, the (similary gaseous) atmosphere of Earth doesn't float off into space because the gravity here is much higher.
Good point! I'm not sure if you are talking about black smokers? If so, I'm pretty sure that these ecosystems still rely on oxygen (along with carbon dioxide and hydrogen sulfide) to fix carbon. Again, this oxygen originally derived from photosynthesis, requiring sunlight. Nevertheless, I seem to recall some underwater ecosystems that rely on the weak geothermal light for photosynthesis, which may be more relevant.
Ultimately, these animals are still dependent on a source of energy such as a sun. These animals are not capable of capturing such energy themselves, but rely on photosynthetic organisms such as plants or algae at the bottom of the food chain. As another poster pointed out, carbon dioxide is necessary for this photosynthesis, not oxygen. This is converted into sugar, which animals subsequently consume. Hydrogenosomes function similarly to mitochondria in converting sugar into ATP (more accessible energy), but differ in that they produce (surprise, surprise) hydrogen as a byproduct.
From the article: "He also found that slum dwellers there collected their excrement in a plastic bag and disposed of it by flinging it He plans to sell it for about 2 or 3 cents — comparable to the cost of an ordinary plastic bag."
i.e. we are 95% confidence that the true proportion lies between 36.7% and 50.4%. That's still quite high.
I think the "easygoing" nature of Aussies is the problem. Our newspapers and other media are a disgrace. We are totally clueless as to international affairs. Sport is always the number one story. Conversely, we don't suffer from as much social unrest as other countries, but sometimes this is frustrating, as I feel this may be helpful.
The evolution of chlorophyll followed (perhaps in Cyanobacteria) in organisms at the bottom of the sea. These were the first organisms to fix carbon dioxide. Being at the bottom of the ocean, only the far bands of visible light were available to them (blue and red), and hence green chlorophyll evolved.
Since then, accessory pigments have also evolved (e.g. phycobiliproteins), which have reclaimed other parts of the visible spectrum, and changed the colour of the plants or algae.