Wikipedia:Reference desk/Archives/Science/2012 September 13

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

September 13

Storm chasing from the sky

Here in the Oklahoma City television market the local news stations will in times of tornadic thunderstorms send up their news helicopters to chase the storms and any tornadoes they produce. Strangely, despite flying in close proximity to large supercell thunderstorms, there seems to be very little turbulence (or at least the cameras on the helicopters holds a very steady picture). Why is this? Also, what distance might the helicopters have to stay away to avoid hazards from the thunderstorm, such as the tornado, hail, etc? Ks0stm ^{(T•C•G•E)} 03:10, 13 September 2012 (UTC)

I know you asked specifically about supercells and tornados, but you may find 53d Weather Reconnaissance Squadron and Hurricane Hunters interesting, as it has some information on arial reconnaissance of major storms. --Jayron32 03:25, 13 September 2012 (UTC)

Cells that produce tornadoes can be highly localized, so the weather may be fine where the chopper is filming. This is in contrast, say, with a hurricane. StuRat (talk) 04:14, 13 September 2012 (UTC)

Also, most the cameras in most television station helicopters are mounted with gyroscopic stabilizers.    → Michael J Ⓣ Ⓒ Ⓜ 08:59, 13 September 2012 (UTC)

Gyro stabilizers prevent the camera from being rotated by the motion of the aircraft - but they can't prevent positional motion from producing problems. It may be that they also record a larger image than is displayed and use motion tracking (eg Digital image correlation) to crop each frame to eliminate the motion. That is the basis of the Image stabilization provided by many modern digital cameras. In this case, since the camera and the subject are many hundreds of feet apart, a foot or more of helicopter motion would only produce a few pixels of image motion between frames - and that can easily be compensated for in this way. SteveBaker (talk) 13:31, 13 September 2012 (UTC)

I guess they would keep enough distance to be able to get away if it came towards them... Thunderstorms are steered by the winds in the area, maybe they preferably film it from the side, outside of it's predicted path. The FAA advices to stay at least 20 miles away from bad weather, but that's general advice for all aviation. Looking for info I came across a law firm page that from the looks of it specialised in "news chopper related injuries to employees" cases. Not sure if that's an indication of the risks they take, or if the firm has similar pages for every potentially lucrative law suit they could think of.

Keep in mind that those crews have quality equipment with powerful zoom lenses, so it may look closer than it really is. Ssscienccce (talk) 19:21, 13 September 2012 (UTC)

How accurate are bathroom scales?

Seems like a simple question. How accurate are digital bathroom scales? My weight has been being shown as either 222.2lbs or 220.0lbs all through this week...never any other number. I suspect that it's internally measuring to the nearest kilogram - but displaying in pounds with a readout that's showing tenths of a pound. Just to check, this morning I was able to weigh myself then pick up two soda cans (12oz each...so 1.5lbs) and see no change whatever in the readout - so clearly there is an internal error that's at least ten to twenty times the displayed precision...and I also kinda suspect that the machine is remembering the last weight it displayed and showing the exact same number again if you weigh yourself again and the results are similar enough.

But Wikipedia has no specific article about these machines - and the general Weighing scale article doesn't mention them beyond a mention in the intro that basically says that they exist! Surfing around the web for a while, reviewers who claimed to have done careful testing gave only vague results ("Very accurate", "Accurate", "Not very accurate") with no actual numbers. Manufacturers and retailers said things like "Display is accurate to 0.2lbs"...which anyone with an ounce of critical thinking skills can parse to mean "We put a ridiculously accurate LCD display on a crappy weighing machine and we're trying to pursuade you that this is an OK thing to do"...which is evidently what happens here.

I'm trying to diet - but any reasonably sustainable diet only drops my weight by a pound or two per week...this horrifying inaccuracy in weighing myself means that I can't track my progress at any kind of reasonable rate...even if I'm careful to weigh myself right after peeing & pooping in an effort to eliminate that source of noise in the measurements - and I average my readings over a week to track progress.

Obviously I don't care much about my absolute weight so much as the rate of change - so I don't care if the machine has my weight off by a few percent, so long as it's consistent from one day to the next and able to show variations of better than 1%.

SteveBaker (talk) 13:23, 13 September 2012 (UTC)

Interesting I had noticed what seemed to be a memory effect on scales. However please clarify the obvious: that the scales are on a hard surface in the same place, with your feet in the same place and your centre of mass similar in all cases? The old scales with counterweights seem able to spot a few grams but modern digital ones I am less convinced about. Good luck with the diet. --BozMo talk 13:33, 13 September 2012 (UTC)

Yeah - I'm very careful to place them in the same spot on my hard bathroom floor - and to stand reasonably consistently...although shifting my position by an inch or two or leaning my weight onto one foot or the other doesn't change the readout. The "memory effect" could well be a simple software hack to make the scales SEEM more reliable by simply repeating the last readout if the measured weight is within the internal error tolerance...but I've tried weighing myself, then putting one foot on the scale and pushing down until it reads some highish number...then weighing myself again...and the results don't change. So either they are remembering the last N weighings in order to defeat my test - or the weighing mechanism is indeed heavily quantized. The problem here is that not knowing how it works makes experiments to determine it's accuracy quite challenging! SteveBaker (talk) 14:10, 13 September 2012 (UTC)

Is your weight consistant to within 2 pounds over the course of any arbitrary length of time? It may be that the scale is accurate enough, but your weight isn't consistent, given the amount you have eaten and drunk, whether you just took your morning dump or not, how much sweating you've done, etc. Your weight is basically fluctuating +/- 1 percent and I don't know that that is outside of the normal range for what a person will do on a given day or week. The way to test this would be to get an object whose weight you know isn't changing, and then weigh it at various times. Your body is far too dynamic and unpredictable a thing to test the reliability of a bathroom scale. --Jayron32 13:47, 13 September 2012 (UTC)

Sure, there is indeed no way that my weight could be that consistent...if I drink a can of (diet!) soda - that's 12 ounces...add a light lunch and I could easily gain a pound or more - and lose it again when I next pee/pooh. So of course it's necessary to weigh oneself after a morning trip to the toilet - and to average over a week or so...all of which I'm careful to do. But I'd hope that this would get the error bars down to within a pound or two...which is what I need to know in order to figure out whether my diet is still keeping on track. Should I walk further next week? Can I let up on the semi-starvation a bit? I know that dieting too fast is a bad idea because it can put your body into crisis/starvation mode and drop your metabolic rate to the point where your diet hits a plateau. The best advice seems to be to try to stick between one and two pounds per week of loss. That's 1% for chrissakes...I really ought to be able to measure that as an average over a week and see if I actually did lose somewhere between one and two pounds over that time.

Bottom line is that if this machine I have is really quantizing in ~1kg increments internally - and maybe isn't even accurate to 1kg - I could easily be getting a 2% error in my measurements - and that's equivalent to an entire month of dieting - and I can't track whether my dieting behavior is excessive or inadequate. If that's the case then I need to invest in a better machine...if such a thing exists for the bathroom at a reasonable price. It's frustrating that I can't find out how accurate these things actually are in order to make an intelligent choice over whether to upgrade my bathroom scales!

SteveBaker (talk) 14:04, 13 September 2012 (UTC)

As a seasoned dieter, I can only repeat the advice every diet programme I've ever seen over the years gives: Only weigh yourself once a week. Resist the temptation to weigh yourself every day. Obviously this advice has the raison d'etre that (a) your weight will fluctuate naturally from day to day, and (b) the scales may not be as well calibrated as you'd like. So same time, same place every week to weigh yourself. Regardless of whether your scales are that accurate, you will at least get an idea of the weekly result and be able to compare it over time. --TammyMoet (talk) 14:55, 13 September 2012 (UTC)

The problem with the weigh-once-a-week and the weigh-daily-and-compare-with-yesterday approaches is that both give too much 'weight' to a single measurement. One's measured weight can swing through two or three pounds based purely on minor changes in the amount of water in one's body and the fullness of one's bowels and GI tract. Daily measurements prompt both the "Sweet - I lost two pounds yesterday!" delusion as well as the "Oh my God - I was fasting and still gained half a pound!" panic. Individual weekly measurements can give similarly misleading impressions about the overall success of a diet. (If I lost two pounds in week one, and no pounds in week two, was there really a thousand-calorie-per day difference in my eating between the two weeks, or am I just seeing noisy data?)

Far better to accept that one's measured weight varies naturally from day to day and hour to hour, and apply some sort of smoothing or averaging to one's measurement history. The Hacker's Diet provides instructions and spreadsheets for doing this; it plots a trendline based on weighted averages of preceding days measurements, to give a plausible impression of one's 'real' weight without the superimposed day-to-day noise. Even a straight three-, five-, or seven-day moving average is going to be more resistant to irrelevant fluctuations than pure daily or weekly measuring. TenOfAllTrades(talk) 15:32, 13 September 2012 (UTC)

Yeah - I think the advice to not weigh yourself every day is for the benefit of stupid people. Intelligent, math-aware, scientific thinkers (which I aspire to be!) can only gain from having more information - providing they are smart about using it. So I gather data by averaging over a week - and look at the trend of averages to give myself some idea of whether I'm doing OK or not. However, not having a decent measurement device to collect that data - one that induces error comparable to several weeks of expected weight loss isn't making life any easier! So the question remains...how accurate are bathroom scales? SteveBaker (talk) 16:58, 13 September 2012 (UTC)

As someone once said, I resemble that remark... I guess there's no accounting for those who would obsess over their weight. If you're not losing weight, don't blame the scales! --TammyMoet (talk) 17:39, 13 September 2012 (UTC)

OK, it's time we stopped guessing. If you have access to Consumer Reports magazine, they have done the testing. You can't read the results of the test online unless you have a subscription, but here is their methodology for studying the accuracy of bathroom scales. For any sort of consumer product testing, that's usually where I go first; they have a reputation for being thorough and impartial. If you don't have or don't want to pay for a subscription online just to view this one article, the magazine is widely availible in many libraries, so perhaps you can get it that way. Good luck! --Jayron32 17:09, 13 September 2012 (UTC)

Well, I don't have a subscription, so I'm going to guess some more. Steve, I think you should dump the digital scales. I have done some of the same style of testing as you, and I hypothesize that they try to make themselves look more consistent than they are, by having logic that attempts to figure out if the person on them is the same one as a few minutes ago, and if so, give the same reading.

Analog scales have limitations — the spring moves more when you step on it than the pressure sensor in a digital scale, so you might think it would soften faster. I don't know if that's true or not. But at least they're too stupid to be trying to fool you.

Best of all should be a real doctors'-office-style balance with the moving counterweights. A little more money but might be worth it if you're looking at this level of detail. --Trovatore (talk) 17:30, 13 September 2012 (UTC)

This would likely depend significantly on both scales. My last 2 analog scales would change about 5-10 kg depending on which way you were leaning and precisely where you were standing on the scale, even when on a flat hard surface. Nil Einne (talk) 09:44, 14 September 2012 (UTC)

That's a .... pretty bad scale. But — at least you could tell how bad they were. I assume neither was of the doctors'-office sliding-counterweight variety? --Trovatore (talk) 09:59, 14 September 2012 (UTC)

My small digital scales (500 gr and 10 gr max) weigh continuously, i.e. the readout changes when the applied weight changes. My bathroom scale gives only one reading, for a second one it has to turn off and on again. The difference is most likely because people move, so it has to average the weight over a longer timeframe. If your's is like that, picking up two soda cans will have no effect. The minimum difference of 2.2 lbs you get may be because they use a non-linear strain gauge to provide the same relative accuracy over the whole range. So a weight of 20lbs would be within 0.2lbs, but 300lbs would only be accurate within 3lbs. It would allow them to use a cheaper (lower resolution) A/D converter. But that's pure speculation on my part. Ssscienccce (talk) 21:30, 13 September 2012 (UTC)

My girlfriend has been weighing herself every morning for about 14 weeks, while losing 0.95 lbs / week. The data actually plot pretty smoothly on a line. At 1-sigma, the typical daily difference from her measurements and the line is +/- 0.7 lbs. So, I would infer that our scale is at least that reproducible, and probably more so since her daily values should also fluctuate based on variations in the daily routine from day to day. Of course your mileage might vary, but the digital scale we have seems to do a pretty good job of tracking relative changes. Dragons flight (talk) 21:37, 13 September 2012 (UTC)

We had a discussion of this very topic maybe a month ago here. Yes, cheap digital scales do have logic in them to provide false accuracy by not changing the reading if it's similar to a recent one. How recent and how much change it will mask may vary by model. To overcome this:

A) Use a known weight which is over the range your scale masks. 10 pounds should do it. So, get on the scale holding the ten pound weight, and you should get your actual weight plus ten pounds. Just subtract. The next day, weigh yourself without the weight, and you should get your true weight. Just alternate days with and without the weight.

B) Pull the battery out between uses. I bet the previous reading is in volatile memory, and will go bye-bye.

As I mentioned in the previous discussion on this topic, the most annoying thing is that they don't tell you prior to purchase that this is what it's doing, you only figure it out later. We need some consumer protection laws here. StuRat (talk) 04:06, 14 September 2012 (UTC)

Ideally, digital bathroom scales should behave like Ssscienccce's 500g scale — the reading stays on all the time you stand on it, and changes to reflect the weight it senses at that moment. Then you could stand there statue-like and wait for it to stop changing, and then you'd know all the transients had damped out. If they did that, I would withdraw my objection in favor of analog scales.

Analog balances, with the movable weights, would still be better, of course. --Trovatore (talk) 04:16, 14 September 2012 (UTC)

(Although, those actually measure your mass...not your weight...but that's probably what you ought to care about anyway! Gravity can vary by as much as half a percent depending on where you are on the earth's surface. Fortunately, I'm only planning on weighing myself (er "massing myself") in my bathroom...so it's OK.) SteveBaker (talk) 12:59, 14 September 2012 (UTC)

Yes, as you say, mass is the important thing; weight-in-the-sense-of-gravitational-force is just a proxy for it. I'll take the opportunity to mention that it's perfectly respectable, historically, to use weight to mean "quantity of matter" rather than "gravitational force" — this is the sense in which bulk goods are sold by weight, the Bureau of Weights and Measures, and so on. The retrospective linguistic prescription imposed in high-school physics seems to be more for the convenience of physics teachers than anything else. Similar remarks could be made for velocity versus speed — if you're teaching a physics class, you need some way to explain how a body can be accelerating even though it never goes any faster or slower, but etymologically they mean the same thing, and you can find velocity as a scalar quantity attested all through the technical literature. --Trovatore (talk) 19:14, 14 September 2012 (UTC)

Maybe digital scale manufacturers figure (or focus groups have shown) that a dancing digital number (like many a digital multimeter measuring a jittery voltage source) that never stops due to the user's shifting his/her weight, which frequently tops 0.1lb on the sensor, frustrates the user to the point where they return the scale and the manufacturer doesn't make money. 20.137.18.53 (talk) 14:45, 14 September 2012 (UTC)

They could eliminate that without reported your previous weight instead of your current weight. It could ignore any readings the first second you are on the scale, then average out the readings over the next second, then give you a reading of your current weight, after 2 seconds. Some type of display to tell you it was "working" would also be in order.

The reason they do it the way they do is to make inaccurate scales appear to be accurate, since weighing yourself immediately again always gives the same result. This is deceptive marketing, and should be illegal (at least unless the customer is notified that it works this way). StuRat (talk) 18:18, 14 September 2012 (UTC)

They could do what you say in your first paragraph, but why? I would prefer to see the numbers change, until they don't. Then I know directly when the transients have damped out and don't have to rely on the scale's firmware. I'm not going to get into the political question of whether it should be illegal, but like you I definitely don't like this practice. --Trovatore (talk) 19:51, 14 September 2012 (UTC)

I believe the average consumer just wants to jump on the scale and have it tell them their exact weight. They don't want to have to watch for it to settle down. Also, they seem to put absurdly small watch batteries in them (mine is powered by a single CR2032, when there's room for a 9 volt battery, or maybe 4 of them). Therefore, you want to display as little as possible, so the battery might last a little longer. StuRat (talk) 21:26, 14 September 2012 (UTC)

WP:OR of course, but mine is the same observation as StuRat's three posts above. When weighing myself within a short time frame (minutes), the first time it takes a while to settle, but then when I weigh myself again, I immediately get exactly the same readout. No matter if I shift the position of my feet, take a pee, drink a glass of water or whatever. So the scales clearly have some sort of memory. The next morning, I get a different readout. So upon reading this thread, I weighed myself several times, each time getting the readout 80.0 kg. I then weighed myself carrying a load of approximately 20 kg. Readout: 99.8 kg. And then, again without the load: 79.9 kg. Repeated measurement: 79.9 kg. Irritating... --NorwegianBlue ^talk 01:10, 15 September 2012 (UTC)

Did you trying pulling the battery out ? StuRat (talk) 01:31, 15 September 2012 (UTC)

I didn't, but decided to try it now after reading your question. First: five consecutive measurements, 78.0 kg (the 2 kg difference from last night accounted for by not having eaten breakfast and wearing less clothes). Then, turned the scales upside down to remove batteries, found I needed my glasses, spent about two minutes looking for them, turned the scales the right way up and repeated the measurement before proceeding. Five consecutive measurements of 77.8 kg. Then another two minute interruption (phone rang). Five consecutive measurements of 78.0 kg. So, first conclusion: this particular device remembers your weight for a minute or two, no more. I then got curious about what difference in weight that was necessary to get a new, real measurement. I found two 1/2 liter bottles of water. Holding one: first readout: 78.0 kg. Second readout: about to settle on 78.0 kg, then changes its mind, settles on 78.7 kg. Five consecutive readouts: 78.7 kg. Two bottles of water: First and second readout: 78.7 kg. Third readout: about to settle on 78.7 kg, then changes its mind, 79.1 kg. Five consecutive measurements: 79.1 kg. When I finally got around to removing the batteries, I found that that would involve bending some parts that looked rather brittle. So I didn't proceed. My curiosity about this particular device was satisfied. --NorwegianBlue ^talk 09:59, 15 September 2012 (UTC)

Unfortunately, you may soon need to replace the battery after all that testing. :-) Your model is better than mine, which seems to remember my weight from yesterday. StuRat (talk) 20:48, 15 September 2012 (UTC)

That doesn't sound right, Stu. It'd be off the scale with those sorts of stratospheric numbers. It's probably just recalculating your IQ. :) -- ♬ Jack of Oz ♬ ^{[your turn]} 09:06, 16 September 2012 (UTC)

Merck Manual/s

Is/are Merck Manual/s worth it/are the buy and etc? Or if not what is/are worth and etc?

Basically, I was looking at it earlier and it seems its worth the buy and etc. In the end didn't get it because I'm still totally unsure about it and etc. — Preceding unsigned comment added by Mybodymyself (talk • contribs) 22:32, 13 September 2012 (UTC)

I couldn't say, but keep in mind that a number of Merck's products have had to be yanked from the market, so use your best judgment. ←Baseball Bugs ^{What's up, Doc?} carrots→ 22:58, 13 September 2012 (UTC)

I don't see any direct connection between a Merck drug being pulled, and the quality of The Merck Manuals. They have a long and well reputed history. And they're worth it if you like that sort of thing. It's not a question, OP, we can answer without knowing more about what it is that you're looking for in a medical handbook. --Tagishsimon (talk) 23:05, 13 September 2012 (UTC)

I second that. Ssscienccce (talk) 23:12, 13 September 2012 (UTC)

The manual is aimed at health care professionals and medical students. Still makes a good read imo, if you like that sort of thing. Don't expect layman terms, it mostly stuff like: "Immature WBCs and RBC precursors are found in the peripheral blood, and marked anisocytosis and poikilocytosis , with microcytes, elliptocytes and teardrop-shaped cells develop." (first random page I picked) If you can find the older centennial version, you get a copy of the 1899 first edition as well. Ssscienccce (talk) 23:12, 13 September 2012 (UTC)

I own a "Merck manual of medical information, Home edition" and it seems to be a reliable source for medical information. The paperback is over 1600 pages, and is not in "medacalese" terminology requiring medical training to interpret. I recommend against buying an obsolete edition, if any reliance is to be placed on the contents for the well-being of self or family, since medical practice and recommendations are constantly changing. Any good local public library is likely to have a recent edition. Your physician should be your primary source of medical information with respect to your health, but he is unlikely to sit and give you as much information as the relevant chapter in the Merck book. Edison (talk) 18:18, 14 September 2012 (UTC)

Also, not sure which Merck Manual the OP is interested in, but all Chemists sleep with two books on their nightstand: the Merck Index and the CRC Handbook of Chemistry and Physics, both of which are fantastic and useful reference books. From a chemists perspective, the Merck Index is worth it. Again, though, not sure which Merck publication you are interested it. --Jayron32 00:59, 14 September 2012 (UTC)

You can buy old editions rather cheap, I have bought or seen them at library sales, and you can get them at online used book sellers like Amazon and, in the US, abebooks.com μηδείς (talk) 03:14, 14 September 2012 (UTC)

Thank you for all of your wonderful responses to my section here. All of them were insightful for sure.--Jessica A Bruno (talk) 20:53, 14 September 2012 (UTC)