Mo's moderately interesting sci/tech thread

Music, books, movies, TV, games, hobbies, food, and potent potables. And forum games! Pour a drink, put on your smoking jacket, light a pipe (of whatever), and settle in.
User avatar
JasonL
Posts: 22619
Joined: 05 May 2010, 17:22

Re: Mo's moderately interesting sci/tech thread

Post by JasonL » 05 Jul 2016, 16:41

Hugh Akston wrote:
JasonL wrote:
thoreau wrote:Somebody looked for the placebo effect in "brain training" exercises and found it. And got it published in one of the top science journals.

http://www.pnas.org/content/113/27/7470.abstract

We need more work like this. Lots more.
As of a few years ago anyway, casual googling about what placebo means results in more confusion than clarity. You'd think we'd have a better understanding about objectively measurable placebo improvements over subjective "I feel better" type things where the underlying measurables are unchanged.

That is, for some/many people, referencing the placebo effect in this context won't put a dent in their confidence in outcomes because for the type of measurements they are doing the whole thing is subjective anyway. Placebo improvement is indistinguishable from real improvement.
Placebo only really comes into play in cases where subjective reporting is the metric though, doesn't it?
That's what I originally thought before I went down the rabbit hole a few years ago. There are confounding studies and meta studies, some of which say there are placebo receptive conditions which will create persistent improvements in objective measurements. There's a distinction then made between physical parameters such as blood pressure and biochemical parameters such as cortisol. I harbor suspicions that subjective, pain level type improvements and even improvements in physical parameters are mostly bunk and on the whole could not be demonstrated as distinguishable from non treatment noise in the right kind of study, but it seems irritatingly difficult to prove that out. So you are left with a bunch of mystics hanging their hats on the proven power of placebo to cure whatever so break out the quantum harmonizer already.

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 05 Jul 2016, 16:46

JasonL wrote:
Hugh Akston wrote:
JasonL wrote:
thoreau wrote:Somebody looked for the placebo effect in "brain training" exercises and found it. And got it published in one of the top science journals.

http://www.pnas.org/content/113/27/7470.abstract

We need more work like this. Lots more.
As of a few years ago anyway, casual googling about what placebo means results in more confusion than clarity. You'd think we'd have a better understanding about objectively measurable placebo improvements over subjective "I feel better" type things where the underlying measurables are unchanged.

That is, for some/many people, referencing the placebo effect in this context won't put a dent in their confidence in outcomes because for the type of measurements they are doing the whole thing is subjective anyway. Placebo improvement is indistinguishable from real improvement.
Placebo only really comes into play in cases where subjective reporting is the metric though, doesn't it?
That's what I originally thought before I went down the rabbit hole a few years ago. There are confounding studies and meta studies, some of which say there are placebo receptive conditions which will create persistent improvements in objective measurements. There's a distinction then made between physical parameters such as blood pressure and biochemical parameters such as cortisol. I harbor suspicions that subjective, pain level type improvements and even improvements in physical parameters are mostly bunk and on the whole could not be demonstrated as distinguishable from non treatment noise in the right kind of study, but it seems irritatingly difficult to prove that out. So you are left with a bunch of mystics hanging their hats on the proven power of placebo to cure whatever so break out the quantum harmonizer already.
Help me parse this. Are you saying that you don't like it when researchers report that interventions (even interventions with mechanisms that are much more plausible than the quantum harmonizer) reduced pain more than a placebo?
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
JD
Posts: 10394
Joined: 05 May 2010, 15:26

Re: Mo's moderately interesting sci/tech thread

Post by JD » 05 Jul 2016, 17:24

Everyone's seen QR codes with embedded logos or whatever, like this:
Image

thanks to the QR code's error-detecting capabilities, but one guy points out that technically, you can manipulate the values to make (almost) the whole thing an image with no errors, like this:

Image
"Millennials are lazy. They'd rather have avocado toast than cave in a man's skull with a tire iron!" -FFF

User avatar
JasonL
Posts: 22619
Joined: 05 May 2010, 17:22

Re: Mo's moderately interesting sci/tech thread

Post by JasonL » 06 Jul 2016, 10:09

thoreau wrote:
JasonL wrote:
Hugh Akston wrote:
JasonL wrote:
thoreau wrote:Somebody looked for the placebo effect in "brain training" exercises and found it. And got it published in one of the top science journals.

http://www.pnas.org/content/113/27/7470.abstract

We need more work like this. Lots more.
As of a few years ago anyway, casual googling about what placebo means results in more confusion than clarity. You'd think we'd have a better understanding about objectively measurable placebo improvements over subjective "I feel better" type things where the underlying measurables are unchanged.

That is, for some/many people, referencing the placebo effect in this context won't put a dent in their confidence in outcomes because for the type of measurements they are doing the whole thing is subjective anyway. Placebo improvement is indistinguishable from real improvement.
Placebo only really comes into play in cases where subjective reporting is the metric though, doesn't it?
That's what I originally thought before I went down the rabbit hole a few years ago. There are confounding studies and meta studies, some of which say there are placebo receptive conditions which will create persistent improvements in objective measurements. There's a distinction then made between physical parameters such as blood pressure and biochemical parameters such as cortisol. I harbor suspicions that subjective, pain level type improvements and even improvements in physical parameters are mostly bunk and on the whole could not be demonstrated as distinguishable from non treatment noise in the right kind of study, but it seems irritatingly difficult to prove that out. So you are left with a bunch of mystics hanging their hats on the proven power of placebo to cure whatever so break out the quantum harmonizer already.
Help me parse this. Are you saying that you don't like it when researchers report that interventions (even interventions with mechanisms that are much more plausible than the quantum harmonizer) reduced pain more than a placebo?
No. I like it when they do that. I am talking more about this idea that floats around out there in harmonizer land that placebo is how magic healing works. There's this unclear area in subjective reporting of pain levels for example where some people think placebo is real persistent improvement over non treatment. The more you enter the land of self reporting the stronger the placebo effect seems to be.

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 06 Jul 2016, 10:59

Thing is, sometimes placebos really do work, even by very objective measures. Otherwise they wouldn't need to do placebo-controlled trials in studies that use measures other than self-reports. Nobody really knows why placebos sometimes work, or whether the effect comes from believing in the pill or the fact that participants in clinical trials also get regular check ups and attention to their overall health, or something else. But placebos sometimes work. And sometimes people have objective improvements simply from resting and taking good care of themselves. It's less about quantum harmony and more about the fact that the old medical advice to eat right, get moderate exercise, avoid stress, and get plenty of rest seems to be good for damn near everything. And people rest and destress better when they feel like a doctor is on top of things.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
JasonL
Posts: 22619
Joined: 05 May 2010, 17:22

Re: Mo's moderately interesting sci/tech thread

Post by JasonL » 06 Jul 2016, 11:08

My complaint was really that we don't seem to have a very clear consensus on if and under what circumstances placebo is better than nothing, and we really don't know what it means for self reported subjective measures. That makes it very unclear what it means to be placebo controlled for self reported measures. It is something we need to do, but it isn't great for clarity if the uncertainty in the control is large. A cynical part of me doubts self reporting of anything ever.

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 06 Jul 2016, 11:23

Placebo-controlled, double-blinded studies are designed around the very problem that you identify with self reports. Give 100 people placebo. Give 100 people a real medication. Ask both groups "Do you feel less pain/exhaustion/etc.?" Make sure that the person asking the question and interpreting the answer doesn't know what kind of pill they got. See if there's a difference. This seems reasonable to me, provided that one accepts the premise that doctors should take things like pain, fatigue, etc. seriously.

Clinical trials are often much more carefully conducted than, say, implicit bias studies reported in the NYT. Some clinical researchers go so far as to persuade the IRB to let them give "placebos" that will induce side effects, so that the patients can't use observations like "Hey, I'm not vomiting!" to deduce that they are in the control group.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
the innominate one
Posts: 12356
Joined: 17 May 2011, 16:17
Location: hypertime continuum

Re: Mo's moderately interesting sci/tech thread

Post by the innominate one » 06 Jul 2016, 13:49

I wonder how frequently that placebo effect is just type I error.
"Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." -E Benn

"No shit, Sherlock." -JsubD

"now is the time to go fuck yourself until you die." -dhex

User avatar
JasonL
Posts: 22619
Joined: 05 May 2010, 17:22

Re: Mo's moderately interesting sci/tech thread

Post by JasonL » 06 Jul 2016, 14:28

the innominate one wrote:I wonder how frequently that placebo effect is just type I error.
Right. There was a big controversial meta study basically saying you can't show any persistence to placebo period. http://www.nejm.org/doi/full/10.1056/NE ... 5243442106

Then counter studies saying the opposite with the same data set.

So ... I dunno, hence annoyed.

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 06 Jul 2016, 14:48

There's enough data suggestive of a real placebo effect that I would be much more likely to trust a study where the control group gets either placebos or the current standard of care (for a disease where we already have a good treatment and the question is whether there's something even better), as opposed to a study where the control group gets nothing. I don't get what is so annoying about this.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
Highway
Posts: 12964
Joined: 12 May 2011, 00:22
Location: the Electric Ocean

Re: Mo's moderately interesting sci/tech thread

Post by Highway » 02 Sep 2016, 19:39

Some people have developed a technique to use vibrations of objects in video to create virtual models they can interact with:

"Sharks do not go around challenging people to games of chance like dojo breakers."

User avatar
Hugh Akston
Posts: 17206
Joined: 05 May 2010, 15:51
Location: El Pueblo de Nuestra Señora Reina de los Angeles

Re: Mo's moderately interesting sci/tech thread

Post by Hugh Akston » 18 May 2017, 16:24

Two studies find that musicians and audiences both prefer modern violins to Stradivarii
The first listener test took place in Vincennes, a suburb of Paris. Researchers gathered three Strads and three top-quality modern violins. An elite violinist played the same musical excerpt—for example, five measures from Tchaikovsky's Violin Concerto Opus No. 35—on each of the nine possible pairings of violins. Then, a second violinist played a different excerpt on all the pairs, with the order scrambled. The violinists wore modified welding goggles, so they couldn’t tell whether they were playing old or new instruments.

As the violins played solo and with orchestral accompaniment, 55 listeners rated which instrument in each pair projected better by making a mark on a continuous scale with one violin, labeled simply A, on one end and the other violin, labeled B, on the other. The researcher then averaged all those evaluations, and found that subjects generally thought the new violins projected better than the old ones—although the researcher left it up to listeners to decide what that meant. The effect was unambiguous, Fritz says.

The team then performed a similar test in New York City without the orchestra and with a different set of Strads and new violins. Again, the 82 listeners in the test reported that the new violins projected better. This time, Fritz and colleagues asked subjects which of the two violins in a pairing they preferred. Listeners chose the new violins over the old, they reported yesterday in the Proceedings of the National Academy of Sciences. The New York City study also showed that listeners' preferences correlated with their assessment of projection, suggesting the loudness of an instrument may be a primary factor in the quality of its sound.
"Is a Lulztopia the best we can hope for?!?" ~Taktix®
"Inexplicably cockfighting monsters that live in your pants" ~Jadagul

User avatar
Highway
Posts: 12964
Joined: 12 May 2011, 00:22
Location: the Electric Ocean

Re: Mo's moderately interesting sci/tech thread

Post by Highway » 18 May 2017, 17:08

What, they needed to do another test to confirm again that people think that louder music sounds better?
"Sharks do not go around challenging people to games of chance like dojo breakers."

User avatar
D.A. Ridgely
Posts: 17922
Joined: 26 Apr 2010, 17:09
Location: The Other Side

Re: Mo's moderately interesting sci/tech thread

Post by D.A. Ridgely » 18 May 2017, 20:52

Highway wrote:What, they needed to do another test to confirm again that people think that louder music sounds better?
They could have just checked with any sleazy speaker salesman ever.

User avatar
Aresen
Posts: 14602
Joined: 26 Apr 2010, 20:18
Location: Great White Pacific Northwest

Re: Mo's moderately interesting sci/tech thread

Post by Aresen » 23 May 2017, 00:17

An interesting Infomercial from Northrup Grumman about the James Webb Space Telescope:



I am eagerly awaiting the science results from the JWST, but the number of critical steps between 'go for launch' and first data is intimidating. So many things that have to go right.

I wonder how the mission engineers will get any sleep for the six months between the launch and the time it is actually operational. (TBS, I am unclear why they didn't build two of them so they could have a backup. The cost is 99% engineering and design, actual labor and materials is nominal.)
If Trump supporters wanted a tough guy, why did they elect such a whiny bitch? - Mo

Those who know history are doomed to deja vu. - the innominate one

Never bring a knife to a joke fight" - dhex

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 23 May 2017, 00:21

Aresen wrote:I am eagerly awaiting the science results from the JWST, but the number of critical steps between 'go for launch' and first data is intimidating. So many things that have to go right.
Supposedly a lot of the critical steps have been done umpteen million times already for spy satellites. They're probably just under orders to make it look hard so nobody will realize how much it's been done already.

I don't generally buy the theory that classified technology is light-years ahead of civilian tech (I think that's highly sector-dependent) but "satellites that you point at distant things to take pictures" are one area where I assume that civilian astronomers really are a couple decades behind the military and intelligence services.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
Aresen
Posts: 14602
Joined: 26 Apr 2010, 20:18
Location: Great White Pacific Northwest

Re: Mo's moderately interesting sci/tech thread

Post by Aresen » 23 May 2017, 00:38

thoreau wrote:
Aresen wrote:I am eagerly awaiting the science results from the JWST, but the number of critical steps between 'go for launch' and first data is intimidating. So many things that have to go right.
Supposedly a lot of the critical steps have been done umpteen million times already for spy satellites. They're probably just under orders to make it look hard so nobody will realize how much it's been done already.

I don't generally buy the theory that classified technology is light-years ahead of civilian tech (I think that's highly sector-dependent) but "satellites that you point at distant things to take pictures" are one area where I assume that civilian astronomers really are a couple decades behind the military and intelligence services.
I can see the deployment of the mirrors as things that may have been done before (though we don't know the failure rate of military/NSA satellites, for obvious reasons), but the sunshield and operation at L2 are new AFAIK.
If Trump supporters wanted a tough guy, why did they elect such a whiny bitch? - Mo

Those who know history are doomed to deja vu. - the innominate one

Never bring a knife to a joke fight" - dhex

User avatar
Kolohe
Posts: 13484
Joined: 06 May 2010, 10:51

Re: Mo's moderately interesting sci/tech thread

Post by Kolohe » 23 May 2017, 11:09

There has literally never been a topic on grylliade with more expertise than you on this one, T, so I'll ask you - is the physics and tech of getting good resolution images of something a few square centimeters in area from a thousand kilometers away transferable to getting good resolution images of something thousands of square kilometers in area from hundreds or thousands or millions light years away?

My layman's feel isn't that its not, anymore than its easy to make a telescope from a microscope.
when you wake up as the queen of the n=1 kingdom and mount your steed non sequiturius, do you look out upon all you survey and think “damn, it feels good to be a green idea sleeping furiously?" - dhex

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 23 May 2017, 11:14

Kolohe wrote:There has literally never been a topic on grylliade with more expertise than you on this one, T, so I'll ask you - is the physics and tech of getting good resolution images of something a few square centimeters in area from a thousand kilometers away transferable to getting good resolution images of something thousands of square kilometers in area from hundreds or thousands or millions light years away?
Yes, because it's all about angular resolution. If you double the distance to the object and the object size then the angle between rays from different points doesn't change, and for distant objects that's what matters.

EDIT: But I'm told that for adaptive optics there are differences between looking up from the earth (the atmospheric distortion happens in the last part of the beam path) and looking down from space (the distortion happens in the first part of the beam path).

Space telescopes looking away from earth don't need adaptive optics, so that simplifies some of it.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
Kolohe
Posts: 13484
Joined: 06 May 2010, 10:51

Re: Mo's moderately interesting sci/tech thread

Post by Kolohe » 23 May 2017, 11:24

Ok, good to know.
when you wake up as the queen of the n=1 kingdom and mount your steed non sequiturius, do you look out upon all you survey and think “damn, it feels good to be a green idea sleeping furiously?" - dhex

User avatar
Aresen
Posts: 14602
Joined: 26 Apr 2010, 20:18
Location: Great White Pacific Northwest

Re: Mo's moderately interesting sci/tech thread

Post by Aresen » 23 May 2017, 11:28

I am going to venture slightly into thoreau's territory - and hope he is gentle in correcting any errors - and add that spy satellites looking in and space telescopes looking out both have to solve problems of platform stability and optimal wavelengths to look through dust or clouds or other matter in between the source and the camera.

The wavelengths are also important because the resolution possible decreases inversely as the wavelength increases.
If Trump supporters wanted a tough guy, why did they elect such a whiny bitch? - Mo

Those who know history are doomed to deja vu. - the innominate one

Never bring a knife to a joke fight" - dhex

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 23 May 2017, 13:27

The dust and gas cloud problems for astronomy are (usually) much less severe than for spy satellites, and if there was in fact a dust cloud as dense as a desert dust storm that itself would be damn interesting. Though a less dense but thicker cloud could admittedly pose the same problems, there are plenty of interesting things that can be seen without clouds in the way. On earth, OTOH, even regions of mild climate will tend to have clouds from time to time.

In astronomy almost any wavelength range will show you something interesting, though what type of interesting thing it will show you varies with wavelength. From gamma rays to radio waves it's all interesting, and for the things that can be seen with devices that look like telescopes as we usually think of them you can get interesting info from UV to far-IR. On earth, you really want either visible, near IR, or close to thermal IR.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
thoreau
Posts: 26382
Joined: 06 May 2010, 12:56
Location: Back to the lab again

Re: Mo's moderately interesting sci/tech thread

Post by thoreau » 23 May 2017, 13:32

I will add that a friend in astronomy* was once at a meeting with some of the contractors working on a space telescope. The contractors had great confidence that a certain problem would be totally manageable. It seemed like a hard problem, and no research astronomer had ever attempted this before with a space telescope. (I forget which space telescope it was; this was a while ago.) At a subsequent meeting, the contractors said that they had finally gotten clearance to say that this task has been successfully performed with satellites seven times previously, so that's why they were quite confident that it would work. They could say no more than that.

If you want to know more about top-secret space telescope technology, tweet to @RealDonaldTrump and maybe he'll tweet back some classified data.

*Yes, I have friends in astronomy. #NotAllAstronomers are SJWs.
"ike Wile E. Coyote salivating over a "4000 Ways To Prepare Roadrunner" cookbook without watching his surroundings, the Road Runner of Societal Inertia snuck up on them both and beepbeeped them off the mesa."
--Shem

User avatar
Kolohe
Posts: 13484
Joined: 06 May 2010, 10:51

Re: Mo's moderately interesting sci/tech thread

Post by Kolohe » 23 May 2017, 13:47

Are the number of working space telescopes in the single digits or the dozens or some order of magnitude more? I honestly have no idea but if I guess before reading the above I would have guessed single digits.

Is it just Hubble hogs all the glory and so that's why it seems there are so few?
when you wake up as the queen of the n=1 kingdom and mount your steed non sequiturius, do you look out upon all you survey and think “damn, it feels good to be a green idea sleeping furiously?" - dhex

User avatar
Kolohe
Posts: 13484
Joined: 06 May 2010, 10:51

Re: Mo's moderately interesting sci/tech thread

Post by Kolohe » 23 May 2017, 13:48

when you wake up as the queen of the n=1 kingdom and mount your steed non sequiturius, do you look out upon all you survey and think “damn, it feels good to be a green idea sleeping furiously?" - dhex

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests