Good gods...
Aug. 20th, 2002 09:11 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
camwynsomeone who had the same experience in undergraduate science labs as me.
This poor bastich is probably going to go insane fairly soon, because he was chosen as User Friendly's Geek Link of the Day. I can sum up his article on his germanium experiments for you right here: 'blow me if the data looks anything like every single textbook ever has insisted it should'. I know I ought to either laugh or feel superior, but frankly, in my experience he's right....
See, when I was in high school, they told us that the mass of an electron was thus-and-such, and that it could be verified through Milikan's oil-drop experiment. They told us that we could and would conduct an experiment to determine the molecular mass of some compound or other, I forget what. They gave us procedures to follow. They gave us a high school chem lab full of equipment and assigned us partners - granted, this was private school equipment, so it was in good condition, but still. They told us that we would get THIS number, or THIS number, and that we had better document how we got there.
So I did. Every step of the way. Down to five decimal places, in the hopes that the outer edges of the scale's precision would help me be more accurate. And I got a number or two, all right... but not their number. Oh, it was close. But it wasn't their number. I tried again using only the first two digits after the decimal point, in case the last few weren't as accurate, but that didn't get me any closer. I tried the experiment again, and I got different results, but those weren't 'right' either. They were on the other side of right from where I'd been.
I got a B on that experiment, and a B+ or an A-, I forget which, on Milikan's oil drop BS. When we had a final exam that involved 'here is a substance that is one of four possible things, now put it through tests to determine which', I picked up my vial and wrote down its contents based on how the vial felt in my hand*. And I was right, but I had to run through the experiments... then I had to fool with the damn numbers again, because they were too fuzzy to be the 'source' of the 'accurate answer'!
They had told us at the start of chemistry that semester that the atomic masses were averages, based on the amount of the most common isotopes found in a mole of the substance. They had told us that Milikan had run his experiment lots and lots of times and finally come up with his result. Then they got angry with me because based on ONE, or based on TWO experiments, I couldn't replicate a year's worth of experimental averages. A number that was yielded by an entire bell curve's worth of results was treated as Holy Writ, the curve forgotten and collapsed down to a single point, as if that one data point were the only one that mattered and the only one anybody ever ought to see. Get it wrong due to the slightest slip-up in your experiment, and you were a Bad Science Student. Get it wrong due to natural variations - you know, the ones that the original scientists experienced, with their dedicated, powerful well-maintained grant-purchased equipment - and you were STILL a Bad Science Student. Lie about your numbers to say 'oo, yes, I got the exact same thing', and you were a Good Scientist, as long as you could produce documentary evidence of how you got that single number.
In the end, it comes down to this: scientists lie. They don't mean to, but they do. They tell you 'ah, this is the sacred number, this is the accurate result', and in the process they forget to tell you how they got it. They will say 'we know X, and X is to be found by doin g thus-and-so'. They say they have the equations that explain everything. The numbers are shiny and perfect and precise, and they are beloved, but because they are so shiny and perfect it is forgotten how those numbers were obtained in the first place... by experiment. By sweat. By hypothesis and testing and writing down a hundred or a thousand or ten thousand results, then drawing a conclusion from all of those results. NOT from one - from an average. When they say results have to be replicable to prove that something is true, everyone acts as if the entire future of a theory hangs on a single experiment or study. No. They mean 'someone who does the same thing we did, the same number of times as we did, in the same way that we did, will get the same results we did at the end of all of that'. To assume that one experiment or one series of experiments will yield the same results presented as Holy Physics (or chemistry, or biology) Gospel is as wrong as assuming that because one woman burst into flames during a Caesarian section, all women must necessarily give birth in the manner of the phoenix. Scientists forget that the public does not have the same background, the same assumptions, the same worldview as them. That people just learning science are not familiar with the procedure and policy and postulates that enable them to get the results they do. And so they lie. . . they don't mean to, it's not intentional, but the end result is that they lie.
So, Mr. Lucas Kovar, I salute you. And I hope your bandwidth spike doesn't blow your account up completely.
*Every bottle of silver nitrate I'd ever touched before that had been extremely cool on the outside. Even when it had been poured out into test tubes. Turned out the instructor kept the stuff in the lab fridge, and the other things we were testing were kept in a cabinet on the other side of the room, at room temperature. As far as I'm concerned this is valid empirical evidence. The instructor should've let the silver nitrate warm up if she didn't want me getting an answer within thirty seconds.
This poor bastich is probably going to go insane fairly soon, because he was chosen as User Friendly's Geek Link of the Day. I can sum up his article on his germanium experiments for you right here: 'blow me if the data looks anything like every single textbook ever has insisted it should'. I know I ought to either laugh or feel superior, but frankly, in my experience he's right....
See, when I was in high school, they told us that the mass of an electron was thus-and-such, and that it could be verified through Milikan's oil-drop experiment. They told us that we could and would conduct an experiment to determine the molecular mass of some compound or other, I forget what. They gave us procedures to follow. They gave us a high school chem lab full of equipment and assigned us partners - granted, this was private school equipment, so it was in good condition, but still. They told us that we would get THIS number, or THIS number, and that we had better document how we got there.
So I did. Every step of the way. Down to five decimal places, in the hopes that the outer edges of the scale's precision would help me be more accurate. And I got a number or two, all right... but not their number. Oh, it was close. But it wasn't their number. I tried again using only the first two digits after the decimal point, in case the last few weren't as accurate, but that didn't get me any closer. I tried the experiment again, and I got different results, but those weren't 'right' either. They were on the other side of right from where I'd been.
I got a B on that experiment, and a B+ or an A-, I forget which, on Milikan's oil drop BS. When we had a final exam that involved 'here is a substance that is one of four possible things, now put it through tests to determine which', I picked up my vial and wrote down its contents based on how the vial felt in my hand*. And I was right, but I had to run through the experiments... then I had to fool with the damn numbers again, because they were too fuzzy to be the 'source' of the 'accurate answer'!
They had told us at the start of chemistry that semester that the atomic masses were averages, based on the amount of the most common isotopes found in a mole of the substance. They had told us that Milikan had run his experiment lots and lots of times and finally come up with his result. Then they got angry with me because based on ONE, or based on TWO experiments, I couldn't replicate a year's worth of experimental averages. A number that was yielded by an entire bell curve's worth of results was treated as Holy Writ, the curve forgotten and collapsed down to a single point, as if that one data point were the only one that mattered and the only one anybody ever ought to see. Get it wrong due to the slightest slip-up in your experiment, and you were a Bad Science Student. Get it wrong due to natural variations - you know, the ones that the original scientists experienced, with their dedicated, powerful well-maintained grant-purchased equipment - and you were STILL a Bad Science Student. Lie about your numbers to say 'oo, yes, I got the exact same thing', and you were a Good Scientist, as long as you could produce documentary evidence of how you got that single number.
In the end, it comes down to this: scientists lie. They don't mean to, but they do. They tell you 'ah, this is the sacred number, this is the accurate result', and in the process they forget to tell you how they got it. They will say 'we know X, and X is to be found by doin g thus-and-so'. They say they have the equations that explain everything. The numbers are shiny and perfect and precise, and they are beloved, but because they are so shiny and perfect it is forgotten how those numbers were obtained in the first place... by experiment. By sweat. By hypothesis and testing and writing down a hundred or a thousand or ten thousand results, then drawing a conclusion from all of those results. NOT from one - from an average. When they say results have to be replicable to prove that something is true, everyone acts as if the entire future of a theory hangs on a single experiment or study. No. They mean 'someone who does the same thing we did, the same number of times as we did, in the same way that we did, will get the same results we did at the end of all of that'. To assume that one experiment or one series of experiments will yield the same results presented as Holy Physics (or chemistry, or biology) Gospel is as wrong as assuming that because one woman burst into flames during a Caesarian section, all women must necessarily give birth in the manner of the phoenix. Scientists forget that the public does not have the same background, the same assumptions, the same worldview as them. That people just learning science are not familiar with the procedure and policy and postulates that enable them to get the results they do. And so they lie. . . they don't mean to, it's not intentional, but the end result is that they lie.
So, Mr. Lucas Kovar, I salute you. And I hope your bandwidth spike doesn't blow your account up completely.
*Every bottle of silver nitrate I'd ever touched before that had been extremely cool on the outside. Even when it had been poured out into test tubes. Turned out the instructor kept the stuff in the lab fridge, and the other things we were testing were kept in a cabinet on the other side of the room, at room temperature. As far as I'm concerned this is valid empirical evidence. The instructor should've let the silver nitrate warm up if she didn't want me getting an answer within thirty seconds.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
um, *grin*
Date: 2002-08-20 10:19 am (UTC)it made me laugh my head off.
In defense of scientists...being one and not being able to resist :P :
I would say, that what you're describing isn't scientists lying, but science teachers, and those are two different things. Any scientist who thinks their numbers and results are shiny perfect and exact needs their head examined, and most scientists I know laugh at the idea of being able to get really accurate results out of a high-school science lab.
More importantly, most of the scientists I know, especially my colleagues in physics, are hugely bothered by the way high school science is taught for just this reason. No science teacher I've ever met teaches correct handling of experimental errors, either in terms of statistics or systematics. Instrumental precision is only one part---and a small part at that---of determining the precision of your final result.
A result that's too good out of a high school science lab, frankly, generally means someone was cooking the books---consciously or unconsciously. (Yup, unconsciously. There are all kinds of ways to bias your data to be what you expect without realizing you're doing it). And it sure sounds to me like what you were getting, most of the time, were perfectly reasonable results if other experimental errors had been correctly taken into account.
IMHO, all of this is what really should get taught at the high school level, the labs and their results are worse than useless otherwise. But most science teachers aren't qualified to do it, so we get the cookbook "_This_ is the answer" approach instead. :(
Sorry, you hit one of my top ten pet peeves about science education ....
That being said, yes, scientists do lie :P Not because we mean to, or because we want to, but because it is sometimes difficult to translate from a very specialized language, and intuition based on years of experience, and still accurately explain what we're talking about. So we give up and cheat.
no subject
Date: 2002-08-20 03:45 pm (UTC)Fortunately my teachers were more interested in getting our actual results and our margin of error and seeing that we'd calculated it all correctly and that our results weren't impossible.
Also fortunately, my father is a research scientist so if they tried to give me any of that I could politely and annoyingly explain bell curves and averages to them and that perhaps they should review their own texts. It's obnoxious, but it generally shuts them up.
Ok, so my teachers also never graded experiements on how close to the ideal we were, they graded on whether we'd recorded everything properly and gone through the steps and if we could offer reasonable exlpanations for why our data wasn't the average.
no subject
Date: 2002-08-21 07:15 am (UTC)It's been years. I don't have to do that any more. I just have to remember this, for if I ever have kids. Even the most enlightened science teacher I've ever met could still ding you for failing to come up with what was supposed to be the right number. I'll make sure my kid knows why.