How To Design A Monster Vs Coat Hanger Experiment?

One of our readers is an enterprising psych major and he would really like to recreate the Monster Cable vs Coat Hanger test with laboratory-grade methodology, controls, and statistical measures. However, Adam needs your help. What is the minimum equipment he should buy, both audio equipment and coat-hanger-wise?

Comments

Edit Your Comment

  1. I’m no scientist, but I did own an A/V store for the last four years. If you use a cheap amp and speakers, you are going to run the risk of people saying that you couldn’t hear the difference because your equipment sucked.

  2. Traveshamockery says:

    Speaker cables don’t need shielding – only line-level cables do. For a speaker wire test, the “standard cable” could be lamp cord, as long as it’s the same gauge as the fancy cable being tested.

  3. Traveshamockery says:

    @suburbancowboy: This is true. You’re going to need to remove the objection that “the cables were held back by poor speakers or amplifiers”.

    In other words, you’re going to be spending at least 5 grand on amplifiers and a pair of speakers to convince any skeptics.

  4. apowers says:

    I’m the aforementioned psych major, and I’d like to add that it doesn’t necessarily have to be a Monster vs. coat hanger experiment — ANY comparison between products and services that Consumerist readers might find useful is fair game, especially if it’s outside the scope of something like Consumer Reports (so I get the added warm glow of making a unique contribution to the blogosphere).

    But, really, any suggestions are welcome.

  5. geekpdx says:

    No need to buy equipment – just rent two 2 hours in a recording studio. Contrary to popular belief, they don’t all run for thousands of dollars an hour. In fact, you might find a small label right in town that will lend you the use of a studio for very little money, though they may want a little free publicity. This way, you don’t have to buy a bunch of high end equipment, and you also have the luxury of performing the listening test in a sound room.

  6. disavow says:

    Borrow an oscilloscope from electronics lab. Connect its input jack to an amplifier, first with the Monster cable and then with the coat hanger. Use the same audio for each. Decent oscilloscopes will let you record and superimpose input, so the two signals could be superimposed to check for variations. If there’s no real difference, there ya go.

  7. e-gadgetjunkie says:

    From a psychology point, the best thing to do is a double blind study. Get some people, (preferable who know nothing about psychology or electronics) and have them turn on and off stereos for the test subjects. That way only the person in charge of the experiment knows which set is connected with what.
    Also, I agree with get good quality equipment. Two identical sets of everything. And maybe throw in generic cables too instead of just top of the line and coat hanger.

  8. Michael Belisle says:

    First think I’d do before you buy anything is design the experiment: [www.itl.nist.gov] . If you survive that, then you’ll know what to buy.

    As for equipment, I’d find an oscilloscope and a function generator. If you’re doing line-level signals, then that’s all you need. Just compare a known input signal to the output signal over the two cables.

    But once you add in the variety of Monster Cables available and start picking out an amplifier for real-life tests, I think there are too many variables and its too subjective for anyone to be converted.

    @InfiniTrent: I agree with the lamp cord suggestion since the coat hanger is so impractical that it’s stupid. You might wrap aluminum foil around the lamp cord to shield it for line-level tests.

  9. Michael Belisle says:

    @apowers: Do you have any equipment at your disposal already? Access to a lab, maybe? It’s going to be hard if you don’t.

  10. Michael Belisle says:

    @e-gadgetjunkie: Why people who know nothing about electronics? Why not get oldsters who are hard of hearing? People with cochlear implants, maybe? I’m sure they’ll all agree that the coat hanger sounds just as good, which will conclusively prove the null hypothesis for discerning Consumerists.

    I suggest musicians and audio engineers for the sound test.

  11. AlCzervic says:

    Are you trying to measure the human perception of the output signal or are you trying to measure the actual distortion of the signal over the line?

    If you are trying to measure the distortion over the line, ask around the electrical engineering department of your university and try to get access to a spectrum analyzer and source generator. Better yet, ask about the possibility of using the department’s network analyzer. If you can’t get access to any of those tools, an oscilloscope would probably work just fine. I wouldn’t expect them to just let you have at any of this equipment as it is all very expensive, but if you could enlist the help of an electrical engineering student, you might be able to get some time with the tools you need.

  12. Traveshamockery says:

    Actually, you could improve on the current experiment a lot by doing an ACTUAL double blind test. The person asking the questions of the subjects should have no idea which cable is playing, and the subjects shouldn’t know either.

  13. satoru says:

    Here’s how I would do the test protocols. I think the original article gave a decent description already. I don’t think that having a bad speaker or amplifier is going to make a big difference. I mean you don’t want to have static on the line constantly, but it’s not like the amp or speaker is going to ‘adapt’ and make things worse/better depending on the cable.

    1) Get an amplifier with an A/B switch for speakers. This will allow you to easily switch between the 2 speakers during the test. However, to eliminate a variance that perhaps the A switch is better than B, repeat each experiment with the speakers swapped between A/B. So one time put the coathangers on A, the other time on B.

    2) Perform each test multiple times. This gives you more data points and thus will tend to average out any variances. Like the original article 6 sets should give you a decent sampling.

    3) Standardize the music chosen. I would suggest a short 30 second to 1 minute clip. Since you’ll need to repeat this lots of times, it keeps your subjects more focused. The music should be varied as in the original test.

    4) Consistency is key. If you have multiple test subjects, either do each person separately, or have them all together but sitting in the exact same spot for every test.

  14. satoru says:

    @InfiniTrent: This is a good point. I think the A/B speaker switch on the amp would go a long way for that.

  15. satoru says:

    I guess one other point would be that subjects should not know what the other subjects answers are. Doing the tests individually would ensure that there’s no selection bias because I saw that John rated cable A higher.

  16. apowers says:

    @Michael Bellisle: I don’t have access to a sound lab; I originally intended on purchasing or renting whatever equipment I needed, but it’s looking like it’ll be pretty expensive. :P

    As I mentioned before, I’m not interested solely in redoing the Monster experiment; I was hoping to get some feedback on other experiments that might also be of interest. I would think that most people are pretty sure that Monster cables are a sham.

  17. Elvisisdead says:

    @apowers: I’d suggest running this past the human subjects board and the professor that’s sponsoring your activities before you get any further. Unless you have somebody backing this idea, it’s a long walk off a short pier.

  18. blacketj913 says:

    Get all the technical details figured out, enough people have already gone into that.

    But to make this a more interesting Psychology experiment I would suggest three different tests. A double blind test as described by earlier posters, a second test where the listeners are told exactly what they are listening through, and a third test where the listeners are told the opposite of what they are really listening through.

    Comparing the results of all three tests against each other would really show how consumer expectations can change what the listener thinks he/she hears in quality.

  19. anonymouscoworker says:

    @Michael Belisle:
    I’m not sure strictly experimental designs for engineering are the way to go here. In this case I think social design coupled with engineering design might be helpful. For the audio equipment piece (or for anytime you’re comparing something technical) you want to know as much about keeping the technology constant as you want to know about your participants. So keep a Solomon 4 Group Design in mind as well.

    @Elvisisdead: I agree, you might get some great help/feedback from your school’s Internal Review Board (IRB).

  20. Michael Belisle says:

    @anonymouscoworker: I see now that my post can pretty much be summed up as “Stick to engineering and objectively quantifiable tests. That psychology part is where the smoke and mirrors comes in. You might as well quit now.”

    Yup, seems I’m still an engineer.

  21. Michael Belisle says:

    @satoru: The quality of the amp is one of the arguments Monster Cable pushes. If the amp doesn’t accurately reproduce sound in the first place, why do you care if your cables do?

    There’s no argument that spending $1000 more on a better equipment will do more than spending $1000 on better cables. The question comes in when you have a $15,000 system: Are you really going to wire it up with coat hangers?

  22. TechnoDestructo says:

    @apowers:

    Yeah, but the cheaper and seemingly crappier the opponent to Monster, the more it rubs Monster’s face in it.

  23. framitz says:

    Step 1.
    Take a science class and learn the scientific method.
    Step 2. Design your experiment.
    . . .

  24. @InfiniTrent:

    Rent it from Costco :-) (j/k)

  25. Michael Belisle says:

    @framitz:
    Step 3. ???
    Step 4. Profit!

    The “scientific method” is a load of bullshit.

  26. muddgirl says:

    The problem with comparing the input and output of an oscilloscope is that, if we’re comparing a coat hanger to Monster cables, there probably will be a technical difference However, the question is, will that difference be audible to listeners?

  27. zarex42 says:

    I am a professional scientist, and would be qualified to do an experiment like this myself. So I’ll chime in.

    satoru is correct in pretty much everything. The A/B switch needs to be absolutely inaudible, so the levels must match perfectly, and there should be no switch noise (not quite trivial).

    Use a high quality amp (no, an audiophile grade amp is NOT necessary, but it will disarm the morons who try to use this as an argument against the experiment).

    Double-blind is absolutely essential, and you need as much data as possible. Lots and lots of datapoints, since people are “noisy”. Use plenty of different music.

    You may also want to use TWO sets of speakers. Tell the people listening that they are judging SPEAKER QUALITY, not cable quality. This will remove some bias. Obviously one set of speakers is just a dummy, but make sure the listeners don’t know this and can’t figure it out.

    I would recommend also these materials, if there is time:

    - Cheap lamp cord from home depot (16ga or bigger)
    – Crumpled up aluminum foil in long strips (just make sure the connection to wire is good and strong)
    – An old car bumper

    The connection between the wire hangar (or whatever) and wire should be as good as possible as well – weld if possible, and make sure that any coatings are removed.

    Good luck! Come ask me/us for help if you need it!

  28. zarex42 says:

    @Michael Belisle:
    Any decent amp will reproduce the signals without any audible byproducts at all. There is no need to go crazy, except to eliminate a dumb argument from the audiophiles.

  29. feralparakeet says:

    I don’t know about the equipment necessary, but being a social science type and a methods buff (public administration, yay!), here’s my two cents.

    I would use at least 3 different types of wiring and double-blind it (obviously, this requires you to have a co-PI, but I’ll mention that later). I’d also have subjects listen to multiple types of music (classical, rock, and something else) that will get the widest spectrum of frequencies/sounds possible, without being irritating.

    You’ll want to collect a good pre-test, and ask them about their listening preferences, whether they consider themselves audiophiles, etc., and demographics of course. I’m sure you can find some good background research in cognitive psych that may note some differences.

    As far as strictly comparing groups, t-tests and ANOVA will be fine. But since you’ll end up getting enough data for some real research, I’d strongly suggest using some logit or probit techniques to predict possibilities of selecting a given type of wiring as superior. I suppose you could also use OLS regressions, but that depends on how you decide to measure their preferences. I’d say do a 1-10 scale rating, and let them rate their preferences, and analyze it both ways. Since it seems that you’re an undergrad, I doubt you’ll be doing any logit or probit anytime in the next few years, but a friendly prof in your department would surely be glad to help with analysis in exchange for co-authoring some papers off the data. You’ll need IRB Human Subjects approval to do this anyway, which requires faculty sponsorship when you’re a student. If you’re wanting to do grad school, it’ll be worthwhile.

    Quite grateful for the distraction from my own research and analysis to babble about someone else’s. Good luck!

  30. pfeng says:

    @geekpdx: Ooooh, the recording studio is a very good thought!

  31. theron says:

    Oscillioscopes and spectrum analysis are helpful, but only get you so far in subjective testing like audio quality. There will be differences in the signals between any two cables, but the question is can the human ear perceive them? There is a standardized methodology for these types of tests, called ABX double blind testing. The listener has three choices: source A, source B, and X – which is either A or B (chosen at random before the test). The tester can switch between the inputs all they want. The test is not can they hear a difference between A and B, but can they tell which of A or B does X match. A single tester can guess correctly 50% of the time, but with enough testers, if you don’t see a stastically significant improvement above random, there is no subjective difference.

    There are purpose-built ABX comparators for such tests. You have to provide a suitable listening environment, and hide the cabling, of course. It is in theory possible to do the same thing in software (freely available programs do exist) but that opens up a lot of variables and possible excuses when the golden ear types can’t distinguish between Monster and coat hanger.

    This ONLY applies to analog cables. A scope (and a knowledge of what to look for) will tell you if a digital cable is working or not. Unless your run is very long or you are putting the cable where it cannot be easily replaced (like inside a wall), there is no point spending more than $10 for any digital cable. A 2 meter HDMI cable selling for $120 is the real rip-off.

  32. moore850 says:

    You need several other setups as control’s, i.e. a typical boombox, a component system that doesn’t use either monster or coat hanger, and an extreme component system with cabling way more expensive than monster cable. You randomly switch between A and B where A is one of these , and B is monster cable or coat hanger. Then you plot the results against a chart and see how close A and B came in the lineup to the other systems. Expensive but comprehensive. Maybe a good component store would lend you the equipment for the test.

  33. cerbie says:

    Buy cables you think are good enough, and move on. This is a tired subject, done over and over, and the conclusion is always the same: people disagree. As long as you can show Monster and the real ripoff folks are just that, let people do as they wish. You will not succeed is proving that it sounds the same, better, or worse. Even if you think you do, you haven’t. You will only serve to frustrate yourself.

    Think of it like having proof God does not exist, and trying to tell it to someone who believes fossil records are a hoax. It does not matter how good your experiment is.

  34. WylieCalidus says:

    since you are a doing this for a psych project, don’t think equipment at all. Think perception. Use what ever equipment you have on hand, even an mp3 player. Just set up the experiment to look like your testing cables. Have the test subject listen to music then tell them you are switching to another cable and have them listen again. One group you tell them your switching to particular brands of cables the other you just say your switching. You will get your data from questions that you have prepared for your subjects to answer. Use answers like Y/N or on a scale from 1 to 5. You will also need to include, age, gender, social background and other factors that you may want to include. Try to narrow things down after you do the experimental portion because if you use alot of variables the data can become overwhelming. If you haven’t had a good statistics class yet, find someone who has to give you a hand.