I'm sorry, this makes even less sense.At best they are "Golden Ear Tests" without statistical significance.
I'm sorry, this makes even less sense.At best they are "Golden Ear Tests" without statistical significance.
I'm sorry, this makes even less sense.
Lol, what? You can absolutely do double blind at home.So I have gone from nonsense to not stating the objective.
Now you can slink away.
Thanks DT
No you can not do Double Blind at home.
Lol, what? You can absolutely do double blind at home.
I’m not sure what you’re talking about or why people would laugh at me. Are you taking about a statistically significant double blind test? I’m not saying that’s easily doable at home. (Although I do have a lot of nerdy friends.)Okay, sure do Double Blind Testing at home. Your sample size will way to small for any statistical significance.
Your results are likely due to chance or you peaked.
If it is fun go ahead, just don't be surprised if someone laughs at your results.
LOL
Hello,I’m not sure what you’re talking about or why people would laugh at me. Are you taking about a statistically significant double blind test? I’m not saying that’s easily doable at home. (Although I do have a lot of nerdy friends.)
Thanks for your expert opinion.Hello,
Now that I look, I see that you have 27 posts. I was a little harsh.
You and your nerdly friends go ahead on, do some audibility testing for fun.
Thanks DT
We are still waiting for you to state the objective ie specific question to be answered by the experiment from you as SIY already pointed out.Okay, sure do Double Blind Testing at home. Your sample size will way to small for any statistical significance.
Your results are likely due to chance or you peaked.
If it is fun go ahead, just don't be surprised if someone laughs at your results.
LOL
We are still waiting for you to state the objective ie specific question to be answered by the experiment from you as SIY already pointed out.
Until then see the answer from Amir. Sure you can do DT with one person to determine if that one person can statistically reliably hear something or is just flipping a coin.
Harman studies answer a different question (objective) and are therefore designed differently. Yes here you need larger populations.
That’s your problem right there. You assume the wrong research question for the home DBT. It does not seek to say anything statistically about a population. Its only purpose is to see if the test taker can hear difference between de DUTs.If we choose to study some characteristic of a defined population
Again you are confusing two objectives or are unable to clearly state them and that alone casts doubt on your understanding.First off I assume that your purpose is to give me grief and not to ask me a knowledge question so I will be brief.
If we choose to study some characteristic of a defined population we need to determine what level of confidence is required, what level of error is acceptable and ideal sample size? There is an app for that.
Here is an example:
Population size is 100
Required level of confidence is 95%
Allowed Margin of error 5%
The idea sample size is 80. It only takes a couple of outliers to skew the results.
If @amirm asked a question it was about a population size of 1. So the only sample size is 1. This post was largely meant as a distraction. @mirm does not want to discourage ASR fun.
I made the general to specific statement that you cannot do statistically significant double blind studies at home. I should include the exception that the population being studied lives there. SIY attempted to obfuscate my statement by trying to redirect the conversation from specific to general.
If meant for fun banter in ASR posts the rigors of defining population, level of confidence and margin of error does not matter so much. So have at it.
Thanks DT
That's the question that started this discussion, reference:A) Can one person hear statistically reliably differences?
All the rest is smoke and mirrors.On some days my audio lacks incisive clarity and I discovered if I add one of my modded extension cords using these replacement plugs things seem to brighten up.
"If we choose to study some characteristic of a defined population we need to determine what level of confidence is required, what level of error is acceptable and ideal sample size? There is an app for that."A) one person can hear statistically reliably (takes several samples) differences. For this it only takes a population size of one and can easily be done at home. The result is only valid for that population (person).
B) a majority of (many) people either prefer or can hear a difference. For this it takes obviously many people (large population) each doing several tests (samples). This is usually beyond an at home test for obvious reason. The tests are valid for that population and within a certain confidence can be extrapolated to be valid for other people as well (best example: election polling).
This is what @SIY asked you to begin with. Is it a or b or something else. The objective determines the test design.
The discussion was about A) not B) as nobody said anything, but you, about studying a large population. This is just you trying to deflect (smoke and mirrors)."If we choose to study some characteristic of a defined population we need to determine what level of confidence is required, what level of error is acceptable and ideal sample size? There is an app for that."
This statement covers both A and B above.
How do you accomplish Double Blind Testing on yourself at home.
How do you blind the experimenter?
How do you blind the test subject?
Sounds like Golden Ear circle logic to me.
Using a recording of what you want to compare and software like Foobar ABX comperator.How do you accomplish Double Blind Testing on yourself at home.
As you describe, I have done that testing many times.Many explained it to you already how it’s done and many did it and I am not wasting more time as you clearly want to ignore it. Why you want to ignore it, I don’t know, but maybe it shatters the world of your beliefs that when you would do such as test as described in A) you would indeed not able to pass it and claim the differences you allegedly hear.
Now you are going off on another tangent. “Cheating” would certainly invalidate any test and your wife (experimenter) not being blinded and suggesting there is a difference even invalidates the prerequisites for single blind (test subject). That doesn’t make it impossible to conduct it at home (single blind) and it is certainly possible to instruct the experimenter not to give away any telltales (remarks, sighs, facial expression) or even better leave the room before you enter it. It is not that complicated if you try a bit being honest with yourself.As you describe, I have done that testing many times.
I have all the high tech equipment on my bench. AP, GRAS, ....
I have done all the A X B comparison testing, I know my preferences.
Thing is, it is still not Double Blind. Same as SIY, I grab my wife as she walks past the the door, "can you hear the difference"? "Yes dear".
The thing that I can reliably identify is noise.
I believe that there is a lot of peaking going on under the blindfold.
DT
I have never grabbed your wife.Thing is, it is still not Double Blind. Same as SIY, I grab my wife as she walks past the the door, "can you hear the difference"? "Yes dear".
Want to involve your wife? Let her listen and you do the switching. She might not be an 'experienced listener' (pun intended), but she should be able to notice 'night and day differences' no? Well, from people who did so I know the outcome does not confirm the ussual 'she heard it from the kitchen' story.your wife (experimenter) not being blinded and suggesting there is a difference
Made my dayI have never grabbed your wife.