5 ms·

"researchers played music to a group of chimps (three adult males and four females) and found that all of the chimps responded to the music by moving in dance-l

"researchers played music to a group of chimps (three adult males and four females) and found that all of the chimps responded to the music by moving in dance-like ways—though the degree to which they danced varied greatly between them. [...] the chimps had different moves." So, we have 7 chimps that responded to an unusual sound (we really don't want to call that danceable music) from a source unknown to them in 7 different ways, and we have n researchers that were set out to interpret that as dance? And these researchers can make a living out of that and get published? p.s. I skimmed through the original article. It looks like science and this might even be a standard approach in this field (I don't know) but ... I cannot describe what I think of it.

nikkwong 7y ago
No kidding. This is a quintessential example of what is wrong in research. If you were to just read the paper you’d definitely come to the conclusion that the researchers want to make—that monkeys dance much like we do. Then you go and watch the video and find their conclusion...biased at the least. The instances in which this type of tomfoolery becomes truly malignant are the ones in which it’s less obvious. I.e researchers have hidden or hashed the original data and then published something that we can’t unscramble. I’m literally expecting to hear “everybody dance now” and end up hearing that charade instead. Lol.
- ChuckMcM 7y ago
  Which is the really super awesome part of science. You find chimps that don't care when they hear music. You find chimps that make "dance like moves" when you play atonal non-syncopated note structures, you find chimps that make "dance like moves" when you play dialog. Then you go through and plot your results and you show that this particular study isn't really indicative of what is going on. My mother once lamented "How can they call that science? That isn't a valid way to test that statement." And I reminded her, publishing a paper isn't "science", forming a hypothesis and a set of experiments to prove or disprove that hypothesis, and then having others test your hypothesis in different ways and getting the same result. THAT is the scientific process. And after that happens you know something new and you can demonstrate it on demand, each time, every time, because you know what is going on.
ordu 7y ago
> It looks like science and this might even be a standard approach in this field One of them. I don't see anything wrong with it. Science works that way: you start with informal observation, if you see something curious enough, you move to formalizing observation. You search for operational definition of "dancing". After that you could watch a little more and to test your definition and to find other variables you need to control in an experiment. And only then you probably can stage an experiment, controlling everything that needs to be controlled, using "blind" observers having instructions on what to observe and how to measure, but not knowing goals of observation. It is very important to know how to watch Nature, how to see something new and do not miss it. Your proposed idea "to avoid known words and to describe in terms of observed behavior" wouldn't work for this task. You would drown in describing words and you drown your readers. And probably no one would see few important observations in a heap of unimportant. The funny thing -- scientists do not what dance is. Even in people we do not know what dance is. How we could call something "the dance" if we do not know what dance is? Should we believe a person when she says she was dancing? Or maybe we should believe our perception which has its ways to distinguish dance from not dance? I think, that it is the case when an informal perception based approach is the best. Just because we have no satisfying formal definition of a dance.
- stewbrew 7y ago
  Okay, let's discuss just two details: - They discuss an effect size related to gender based on the observation of n=7 chimpanzees. Unfortunately, the software didn't stop them from calculate t-tests etc. but this is simply ridiculous. At this point, we simply ignore the fact that the chimpanzees "participate in experiments on a voluntary basis". Self-selection bias is unheard of. - Other statements are based on the observation of n=1 chimpanzee. Again, unfortunately the software didn't stop them from calculating an anova. At best, this is a case study. Don't get me wrong, case studies can be very useful and informative. I'm a big fan of good case studies. But you shouldn't disguise it as something else. If you don't have data suitable for statistical analysis, forget about it. Do something else. Don't pretend you're a physicist. BTW I don't question that animals respond to music. Why not. But just repeating piano samples doesn't make sound music -- not even techno. Don't ask animals to dance to sound no human would dance to.

nikkwong 7y ago

No kidding. This is a quintessential example of what is wrong in research. If you were to just read the paper you’d definitely come to the conclusion that the researchers want to make—that monkeys dance much like we do. Then you go and watch the video and find their conclusion...biased at the least. The instances in which this type of tomfoolery becomes truly malignant are the ones in which it’s less obvious. I.e researchers have hidden or hashed the original data and then published something that we can’t unscramble. I’m literally expecting to hear “everybody dance now” and end up hearing that charade instead. Lol.

ChuckMcM 7y ago
Which is the really super awesome part of science. You find chimps that don't care when they hear music. You find chimps that make "dance like moves" when you play atonal non-syncopated note structures, you find chimps that make "dance like moves" when you play dialog. Then you go through and plot your results and you show that this particular study isn't really indicative of what is going on. My mother once lamented "How can they call that science? That isn't a valid way to test that statement." And I reminded her, publishing a paper isn't "science", forming a hypothesis and a set of experiments to prove or disprove that hypothesis, and then having others test your hypothesis in different ways and getting the same result. THAT is the scientific process. And after that happens you know something new and you can demonstrate it on demand, each time, every time, because you know what is going on.

ordu 7y ago

> It looks like science and this might even be a standard approach in this field One of them. I don't see anything wrong with it. Science works that way: you start with informal observation, if you see something curious enough, you move to formalizing observation. You search for operational definition of "dancing". After that you could watch a little more and to test your definition and to find other variables you need to control in an experiment. And only then you probably can stage an experiment, controlling everything that needs to be controlled, using "blind" observers having instructions on what to observe and how to measure, but not knowing goals of observation. It is very important to know how to watch Nature, how to see something new and do not miss it. Your proposed idea "to avoid known words and to describe in terms of observed behavior" wouldn't work for this task. You would drown in describing words and you drown your readers. And probably no one would see few important observations in a heap of unimportant. The funny thing -- scientists do not what dance is. Even in people we do not know what dance is. How we could call something "the dance" if we do not know what dance is? Should we believe a person when she says she was dancing? Or maybe we should believe our perception which has its ways to distinguish dance from not dance? I think, that it is the case when an informal perception based approach is the best. Just because we have no satisfying formal definition of a dance.

stewbrew 7y ago
Okay, let's discuss just two details: - They discuss an effect size related to gender based on the observation of n=7 chimpanzees. Unfortunately, the software didn't stop them from calculate t-tests etc. but this is simply ridiculous. At this point, we simply ignore the fact that the chimpanzees "participate in experiments on a voluntary basis". Self-selection bias is unheard of. - Other statements are based on the observation of n=1 chimpanzee. Again, unfortunately the software didn't stop them from calculating an anova. At best, this is a case study. Don't get me wrong, case studies can be very useful and informative. I'm a big fan of good case studies. But you shouldn't disguise it as something else. If you don't have data suitable for statistical analysis, forget about it. Do something else. Don't pretend you're a physicist. BTW I don't question that animals respond to music. Why not. But just repeating piano samples doesn't make sound music -- not even techno. Don't ask animals to dance to sound no human would dance to.