Login | Register


All times are UTC - 7 hours


It is currently 2019-Aug-20 9:57 pm




Post new topic Reply to topic  [ 128 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6 ... 9  Next
Author Message
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 9:56 am 

Joined: 2012-Mar-31 11:52 am
Age: Elder Dragon
I believe they actually wanted to kill sheldor. Sheldon is quite safe.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 10:29 am 
User avatar

Joined: 2012-Nov-27 4:39 pm
Age: Elder Dragon
Location: Midgard
On the subject of joke answers in the survey, I can't help but wonder how much the data for Braids is tainted. The "FREE BRAIDS" I see people campaign on other social media sites feels only in part sincere--much more like a meme than anything else. Certainly there are people who legitimately want Braids unbanned, but I'm a little skeptical the number of responses for it aren't inflated by people clicking it 'just because.'

_________________
Current:
Decklists are posted here. They can all be found in the Decklist Forum.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 11:34 am 

Joined: 2019-Jun-06 6:25 am
Age: Drake
MRHblue wrote:
I did find it interesting OP skipped over how highly the RC rated on the poll, while previous to it they had said many people were telling him how bad a job they had been doing.

A 6.5 average is mostly middling and leaning towards the positive. Not exactly anything to write home about.
specter404 wrote:
I mentioned it briefly in my last post, but I'm going to put it at the top here. I am a maths teacher at a high school and I teach units on statistics and surveying. Now that doesn't make me an expert but I do speak with some moderate level of expertise when I say you really should not pick parts of responses to include, no matter how well intentioned you may be in that.

Unconscious bias is a significant factor in any survey and even if you don't think you are imposing your personal beliefs on the data, by making a choice about what gets in and what doesn't you run the risk of biasing them.

Given that you either set rules for removing a data set, like in political elections, and follow them rigidly, or you accept all of your data and adjust your analysis with an understanding of where it all came from.

The reason I have for this is how do you decide what is and isn't a troll response? Is "Ban Island" a troll? is "Kill Sheldon" a troll? is "survey are bullshit", did you throw out the whole response when the respondent used the free comment space to abuse you and threaten you personally or did you keep their numerical data? If you set the line on what is and isn't trolling then it's hard to accept the results as unbiased anymore.

The results are there for everyone to see, and I'm not gonna tamper with it. All I was saying is that the gag response in the free response section did not instantly invalidate the numerical data given, because even people who gave troll responses had data that seemed genuine (listing disfavor in the banlist + wanting a lot of bans, for example). All I was saying is that PERSONALLY I don't believe the island ban to be serious and I'm not going to focus on that for the sake of my interpretation of the data.

All and all, this is a big stink over at most 5 or 6 troll responses. I'm not cherrypicking data here or deleting anything.

_________________
ImageImage
ImageImage


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 12:39 pm 
User avatar

Joined: 2006-Dec-31 12:26 pm
Age: Elder Dragon
paragon_deku wrote:
MRHblue wrote:
I did find it interesting OP skipped over how highly the RC rated on the poll, while previous to it they had said many people were telling him how bad a job they had been doing.

A 6.5 average is mostly middling and leaning towards the positive. Not exactly anything to write home about.


How can you say the average is 6.5? Over 60% of the votes are 7 or higher. The most common response is an 8 with a full 1/4 of the votes. This is representative of when I looked at the data earlier today as well (so you can't say "oh, more people must have voted.")

And, unlike you randomly throwing numbers around, here's my proof:
Image

Sounds more and more like you have an agenda, but then can't be bothered to say you're wrong when your own methods of trying to back yourself up show that you are wrong.

paragon_deku wrote:
All I was saying is that the gag response in the free response section did not instantly invalidate the numerical data given, because even people who gave troll responses had data that seemed genuine (listing disfavor in the banlist + wanting a lot of bans, for example).

To repeat what was previously said and you've ignored: how can you possibly know the bolded part?

By doing what you are doing, you are introducing bias to your results. That you aren't publishing the raw data makes it seem like you have something to hide (proper scientific journals publish results and make raw data available for others to analyze -- as they are just as interested as others to know if any mistakes were made.)

Unless you justify what you are doing, you are willingly letting us dismiss the whole survey you did because you are deliberately tampering with the data. As much as you say you're going to leave it alone -- you're showing your bias by not dismissing the whole dataset for any obvious troll results.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:15 pm 

Joined: 2015-Jan-14 2:58 pm
Age: Elder Dragon
The mean score does appear to be 4117/627, or 6.566, based on that chart. The mode is 8, but since only 42.6% scored 8 or better, the median score is a 7.

I’d like to add that expunging only part of a response is indeed poor technique.

_________________
Deepglow Skate
Antis wrote:
I'm seriously suspicious of any card that makes Doubling Season look fair and reasonable.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:21 pm 

Joined: 2019-Jun-06 6:25 am
Age: Drake
Carthain wrote:
paragon_deku wrote:
MRHblue wrote:
I did find it interesting OP skipped over how highly the RC rated on the poll, while previous to it they had said many people were telling him how bad a job they had been doing.

A 6.5 average is mostly middling and leaning towards the positive. Not exactly anything to write home about.


How can you say the average is 6.5? Over 60% of the votes are 7 or higher. The most common response is an 8 with a full 1/4 of the votes. This is representative of when I looked at the data earlier today as well (so you can't say "oh, more people must have voted.")

And, unlike you randomly throwing numbers around, here's my proof:
Image

Sounds more and more like you have an agenda, but then can't be bothered to say you're wrong when your own methods of trying to back yourself up show that you are wrong.

paragon_deku wrote:
All I was saying is that the gag response in the free response section did not instantly invalidate the numerical data given, because even people who gave troll responses had data that seemed genuine (listing disfavor in the banlist + wanting a lot of bans, for example).

To repeat what was previously said and you've ignored: how can you possibly know the bolded part?

By doing what you are doing, you are introducing bias to your results. That you aren't publishing the raw data makes it seem like you have something to hide (proper scientific journals publish results and make raw data available for others to analyze -- as they are just as interested as others to know if any mistakes were made.)

Unless you justify what you are doing, you are willingly letting us dismiss the whole survey you did because you are deliberately tampering with the data. As much as you say you're going to leave it alone -- you're showing your bias by not dismissing the whole dataset for any obvious troll results.

I like how you accuse me of lying and hiding data when not only did you try to disprove me by posting the full data I willingly provided but you also demonstrated a complete unawareness of how the statistical mean works.

Also, I'm not expunging anything. I said for a personal analysis of the data I'm not going to consider gag responses to the free response section.

Also why are you accusing me of tampering? That came out of nowhere. I feel like you're just looking for a chance to jump at my throat because you disagree with me. Not my fault you don't know what a mean is.

_________________
ImageImage
ImageImage


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:26 pm 

Joined: 2015-Jan-14 2:58 pm
Age: Elder Dragon
Unconscious bias is a real problem that is difficult to combat. I don’t see any reason you would willingly tamper with your data, but deciding what parts of responses are legitimate means making a judgement call on what is reasonable. That can easily lead to unintentionally eliminating only those parts of a response you disagree with. Best to throw out entire responses so there isn’t even an appearance of bias.

_________________
Deepglow Skate
Antis wrote:
I'm seriously suspicious of any card that makes Doubling Season look fair and reasonable.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:29 pm 

Joined: 2019-Jun-06 6:25 am
Age: Drake
Spectrar Ghost wrote:
Unconscious bias is a real problem that is difficult to combat. I don’t see any reason you would willingly tamper with your data, but deciding what parts of responses are legitimate means making a judgement call on what is reasonable. That can easily lead to unintentionally eliminating only those parts of a response you disagree with. Best to throw out entire responses so there isn’t even an appearance of bias.

I'm not touching anything about the data, though. Everyone seems to be missing the bit where I talk about "personal interpretation." MY personal interpretation involves not taking some of the free responses super seriously but still looking at their other opinions instead of just ignoring them. And unlike what other people are saying for whatever reason, I'm not trying to hide anything since the data is all there for everyone to see.

_________________
ImageImage
ImageImage


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:34 pm 

Joined: 2015-Jan-14 2:58 pm
Age: Elder Dragon
paragon_deku wrote:
Spectrar Ghost wrote:
Unconscious bias is a real problem that is difficult to combat. I don’t see any reason you would willingly tamper with your data, but deciding what parts of responses are legitimate means making a judgement call on what is reasonable. That can easily lead to unintentionally eliminating only those parts of a response you disagree with. Best to throw out entire responses so there isn’t even an appearance of bias.

I'm not touching anything about the data, though. Everyone seems to be missing the bit where I talk about "personal interpretation." MY personal interpretation involves not taking some of the free responses super seriously but still looking at their other opinions instead of just ignoring them. And unlike what other people are saying for whatever reason, I'm not trying to hide anything since the data is all there for everyone to see.


But that’s precisely the point. By not discounting the parts that are not obviously trolly along with the parts that are, you introduce a route for personal, unconscious bias to creep in. In most cases what you are doing is likely reasonable, but you can’t be sure of that. If you’re going to clean those responses out, doing it for the whole response is safer from a bias perspective.

That said, while I haven’t looked at the data myself, it’s good to have available for those that want it, especially if the trolly responses remain in it so others can see.

_________________
Deepglow Skate
Antis wrote:
I'm seriously suspicious of any card that makes Doubling Season look fair and reasonable.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:37 pm 

Joined: 2019-Jun-06 6:25 am
Age: Drake
Spectrar Ghost wrote:
paragon_deku wrote:
Spectrar Ghost wrote:
Unconscious bias is a real problem that is difficult to combat. I don’t see any reason you would willingly tamper with your data, but deciding what parts of responses are legitimate means making a judgement call on what is reasonable. That can easily lead to unintentionally eliminating only those parts of a response you disagree with. Best to throw out entire responses so there isn’t even an appearance of bias.

I'm not touching anything about the data, though. Everyone seems to be missing the bit where I talk about "personal interpretation." MY personal interpretation involves not taking some of the free responses super seriously but still looking at their other opinions instead of just ignoring them. And unlike what other people are saying for whatever reason, I'm not trying to hide anything since the data is all there for everyone to see.


But that’s precisely the point. By not discounting the parts that are not obviously trolly along with the parts that are, you introduce a route for personal, unconscious bias to creep in. In most cases what you are doing is likely reasonable, but you can’t be sure of that. If you’re going to clean those responses out, doing it for the whole response is safer from a bias perspective.

Yes, I agree. But I won't be because I want people to see the results in their entirety.

_________________
ImageImage
ImageImage


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:45 pm 
User avatar

Joined: 2006-Dec-31 12:26 pm
Age: Elder Dragon
paragon_deku wrote:
Also, I'm not expunging anything. I said for a personal analysis of the data I'm not going to consider gag responses to the free response section.

Also why are you accusing me of tampering? That came out of nowhere.

Because you're telling us that you're tampering with the data by still accepting some data from datasets that are shown to be bogus. And by calling it 'tampering' you finally responded to it. You've been asked at least twice now how you can possibly know that the rest of those datasets are legit, and you've just ignored it.

So all we have is that you are selectively choosing which data to use and which not to use -- that's tantamount to tampering with the data.

You probably have some good data in there -- but we can't trust it because you haven't released the raw data to us and you are willingly telling us that you are picking and choosing which of the data you're including or not. And that you can somehow tell that even though the answer in the free-form had nothing to do with the question, that the other results are worthwhile. That you keep skipping that question throws everything you do with that data into question.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:47 pm 

Joined: 2019-Jun-06 6:25 am
Age: Drake
Carthain wrote:
paragon_deku wrote:
Also, I'm not expunging anything. I said for a personal analysis of the data I'm not going to consider gag responses to the free response section.

Also why are you accusing me of tampering? That came out of nowhere.

Because you're telling us that you're tampering with the data by still accepting some data from datasets that are shown to be bogus. And by calling it 'tampering' you finally responded to it. You've been asked at least twice now how you can possibly know that the rest of those datasets are legit, and you've just ignored it.

So all we have is that you are selectively choosing which data to use and which not to use -- that's tantamount to tampering with the data.

You probably have some good data in there -- but we can't trust it because you haven't released the raw data to us and you are willingly telling us that you are picking and choosing which of the data you're including or not. And that you can somehow tell that even though the answer in the free-form had nothing to do with the question, that the other results are worthwhile. That you keep skipping that question throws everything you do with that data into question.

I've... not chosen in favor of any data? I haven't touched it at all? I'm leaving all the data in its rawest form for you to look at and interpret. I think you're just paranoid.

EDIT: Okay, yeah, you guys are just deliberately ignoring what I'm saying. The data is all there, warts and all. I haven't touched it and you can see it all for yourself.

EDIT EDIT: In case you want it in a more malleable form instead of the charts google forms provides, here you can see the spreadsheet:
https://docs.google.com/spreadsheets/d/ ... sp=sharing

_________________
ImageImage
ImageImage


Last edited by paragon_deku on 2019-Jun-18 1:50 pm, edited 1 time in total.

Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:49 pm 

Joined: 2015-Jan-14 2:58 pm
Age: Elder Dragon
Is there a link we can follow to a read-only dataset? It’s possible you’ve included it already but I don’t really feel like digging back through the thread.

_________________
Deepglow Skate
Antis wrote:
I'm seriously suspicious of any card that makes Doubling Season look fair and reasonable.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:51 pm 
User avatar

Joined: 2006-Dec-31 12:26 pm
Age: Elder Dragon
Or did you ever consider that, as there are multiple people telling you the same thing -- that there might be something to it?

Are you even trying to understand what we're saying? Or are you just sitting back saying "you can see all the data!"

Also -- we can't see all the data. We see all the results for the multiple-choice questions, but not the free-form ones. We also have no way to associate the free-form ones with the multiple-choice ones to see if there's any patterns there that could be significant.


Top
 Offline Profile  
 
 Post subject: Re: A General EDH Survey
AgePosted: 2019-Jun-18 1:54 pm 

Joined: 2019-Jun-06 6:25 am
Age: Drake
Carthain wrote:
Or did you ever consider that, as there are multiple people telling you the same thing -- that there might be something to it?

Are you even trying to understand what we're saying? Or are you just sitting back saying "you can see all the data!"

Also -- we can't see all the data. We see all the results for the multiple-choice questions, but not the free-form ones. We also have no way to associate the free-form ones with the multiple-choice ones to see if there's any patterns there that could be significant.

You could have just asked me to post the spreadsheet at any time, dude. I had no way of knowing you couldn't see the associations since they show up on my end. I'll admit that's my mistake, but the fact you jumped to tampering and, again, didn't know what a mean was is pretty silly

_________________
ImageImage
ImageImage


Top
 Offline Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 128 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6 ... 9  Next

All times are UTC - 7 hours


Who is online

Users browsing this forum: No registered users and 37 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to: