A poll for interest's sake on some controversial topics.

[March Audio]

52 participants so far.  @MLXXX how long will you run this?

[LHC]

For perspective SNA currently has 110,746 active members capable of voting in this poll. 

[March Audio]

14 minutes ago, LHC said:

For perspective SNA currently has 110,746 active members capable of voting in this poll. 

 

And how many do you need to say satisfy 95% confidence and say 10% accuracy?

 

Multiple sources suggest about 100.

Edited August 16, 2023 by March Audio

[LHC]

16 minutes ago, March Audio said:

 

And how many do you need to say satisfy 95% confidence and say 10% accuracy?

 

Well according to this calculator one needs a minimum of 96 responses. The tricky variable is the "Population Proportion" which is unknown across the membership, so I just used 50%. Also not explicit is the underlying assumptions used by the calculator and how applicable they are to the format of this poll. Worth bearing in mind.

 

[March Audio]

1 minute ago, LHC said:

 

Well according to this calculator one needs a minimum of 96 responses. The tricky variable is the "Population Proportion" which is unknown across the membership, so I just used 50%. Also not explicit is the underlying assumptions used by the calculator and how applicable they are to the format of this poll. Worth bearing in mind.

 

 

There are of course many variables, but the takeaway is that is always surprising (to me at least) how small the sample size needs to be.

[LHC]

2 minutes ago, March Audio said:

 

There are of course many variables, but the takeaway is that is always surprising (to me at least) how small the sample size needs to be.

 

Yes, the sample size can be 'relatively' small, but again that depends on the assumptions made. If there are hidden bias in the sampling then one needs a much larger sample size to counter-balance their influence. 

[March Audio]

1 minute ago, LHC said:

 

Yes, the sample size can be 'relatively' small, but again that depends on the assumptions made. If there are hidden bias in the sampling then one needs a much larger sample size to counter-balance their influence. 

 

I guess we will never know

[MLXXX]

19 hours ago, March Audio said:

52 participants so far.  @MLXXX how long will you run this?

The tally of voters was 47 three days ago.  You say it was 52  a bit earlier this evening. That may have prompted  some to vote as I see the tally has now reached 61!

 

My thinking is to wait until the end of this coming Saturday night. That would be one week and one  day after the poll started.

 

At that  stage the poll would still be running but we could disclose and summarise the results to that point, and some people might wish to make some general comments!

Edited August 17, 2023 by MLXXX

[LHC]

I would encourage all eligible SNA members to vote and see if we could reach that theoretical number of 96. Even if one is in principle or philosophically oppose to the construct of this poll, just do it anyway. Statistics needs large (enough) numbers to have a chance of being reliable. 

[MLXXX]

1 hour ago, March Audio said:

 

There are of course many variables, but the takeaway is that is always surprising (to me at least) how small the sample size needs to be.

That has surprised me too, ever since I first learnt about statistical sampling!

 

A potentially confounding issue is that the sampling is only of the member population who read this part of the forum (and who choose to vote).   That may not be a good proxy for SNA membership at large, as quite a high percentage of members may rarely if ever look here (for whatever reasons).

 

I hope though that this little poll may help us get a bit of an insight into which side of the fence forum members are more likely to gravitate to, even if that insight might only more clearly apply in respect of the types of members who choose to come to this part of the forum.

Edited August 16, 2023 by MLXXX

[aussievintage]

6 hours ago, MLXXX said:

A potentially confounding issue is that the sampling is only of the member population who read this part of the forum (and who choose to vote).   That may not be a good proxy for SNA membership at large, as quite a high percentage of members may rarely if ever look here (for whatever reasons).

 

That's ok, from the perspective of seeing the results as indicative of how the the crowd in "Great Audio Debate" see the world.    The other way people will see it and vote, is in the "Selected since last visit" pages.  They don't have to be specific readers and come in here to get news of what's happening  

[POV]

10 hours ago, LHC said:

I would encourage all eligible SNA members to vote and see if we could reach that theoretical number of 96. Even if one is in principle or philosophically oppose to the construct of this poll, just do it anyway. Statistics needs large (enough) numbers to have a chance of being reliable. 

 

hmm...why would someone have a principled or phisosophical opposition to a poll about hifi things on a hifi forum?

[Hi-Fi Whipped]

Great Poll, could render the Great Audio Debate Sub Forum extinct by the looks of the results! 

Edited August 17, 2023 by Hi-Fi Whipped

[aussievintage]

9 minutes ago, Hi-Fi Whipped said:

Great Poll, could render the Great Audio Debate Sub Forum useless by the looks of the results! 

 

Glass half empty  - I say it could render it useful  

Edited August 17, 2023 by aussievintage

[rantan]

41 minutes ago, POV said:

 

hmm...why would someone have a principled or phisosophical opposition to a poll about hifi things on a hifi forum?

 

Precisely.

It appears to me that some people are overthinking this. It is not life and death or an examination of one's core belief system.

 

 

 

[POV]

1 hour ago, Hi-Fi Whipped said:

Great Poll, could render the Great Audio Debate Sub Forum extinct by the looks of the results! 

Does that mean we can have robust discussions in the rest of the forum without people getting overly upset!?

[Hi-Fi Whipped]

1 minute ago, POV said:

Does that mean we can have robust discussions in the rest of the forum without people getting overly upset!?

One can only dream!

[rantan]

1 hour ago, POV said:

Does that mean we can have robust discussions in the rest of the forum without people getting overly upset!?

 

  I wouldn't bet a lot of money on that.

Robust discussions here tend to be people from two irreconcilable camps trying to nuke one another whenever possible, with circular arguments based on established prejudices which have no prospect of being changed.

[aussievintage]

10 minutes ago, rantan said:

established prejudices which have no prospect of being changed.

 

the truth cannot be changed      

[rantan]

1 minute ago, aussievintage said:

 

the truth cannot be changed      

 

What is one person's truth is another's illusion.

[aussievintage]

Just now, rantan said:

 

What is one person's truth is another's illusion.

 

yeah, but it still remains true, no matter what perceptions anyone might have.

[LHC]

6 hours ago, Hi-Fi Whipped said:

Great Poll, could render the Great Audio Debate Sub Forum extinct by the looks of the results! 

 

No it doesn't ... not yet anyway. There are already several posts cautioning about (mis-)interpreting the poll's results in its present state. 

[LHC]

22 hours ago, LHC said:

 

Well according to this calculator one needs a minimum of 96 responses. The tricky variable is the "Population Proportion" which is unknown across the membership, so I just used 50%. Also not explicit is the underlying assumptions used by the calculator and how applicable they are to the format of this poll. Worth bearing in mind.

 

 

Without going into the underlying assumptions of this calculator, one could improve on the 10% margin of error if more people were sampled:

• 5% margin of error requires a sample size of 383
• 4% margin of error requires a sample size of 598
• 2% margin of error requires a sample size of 2351

It is intuitive that more poll data would lead to greater precision. The diagram below is a good depiction (but using a different population size) 

 

[Hi-Fi Whipped]

1 hour ago, LHC said:

 

No it doesn't ... not yet anyway. There are already several posts cautioning about (mis-)interpreting the poll's results in its present state. 

I’m a visual guy, so in the case of the charts I’m seeing above, is it a case of regardless of the amount of shafts in the room it’s a proportional thing and the biggest shaft wins?

[LHC]

17 minutes ago, Hi-Fi Whipped said:

I’m a visual guy, so in the case of the charts I’m seeing above, is it a case of regardless of the amount of shafts in the room it’s a proportional thing and the biggest shaft wins?

 

I am sorry, I don't quite understand your question.

 

If you are asking about the graphs relating margin of error to sampling size, this link article provide a good explanation on how to interpret them. I think the article will be useful once we start examining the our poll results and turning them into percentage points.

 

https://journalistsresource.org/media/margin-error-journalists-surveys-polls/