Life In 19x19
http://www.lifein19x19.com/

A Go Survey for Statistics Class
http://www.lifein19x19.com/viewtopic.php?f=10&t=12374
Page 1 of 2

Author:  scapetude [ Tue Oct 20, 2015 8:13 pm ]
Post subject:  A Go Survey for Statistics Class

Hello everyone!

I am doing a survey for my statistics class on Go. If you have a moment and don't mind helping, then please answer the following questions. If you're more comfortable messaging me the answers, that is fine too.

1) What is your gender? Male | Female (Choose one)

2) What is your age? __________ Years old

3) What is your current Go rank? __________ Kyu | Dan (Choose one)

4) How long have you been playing Go? __________ Years

5) Have you taken breaks from playing Go? Yes | No (Choose one)

If yes, how did it affect your Go games? Positively | Negatively (Choose one)

6) Do you do Go problems every week? Yes | No (Choose one)

7) On average, how many hours do you spend playing Go in a week? __________ Hours per week

8) Do you prefer fast or slow games? Fast | Slow (Choose one)

Thank you for your assistance with my class project!

Author:  Abyssinica [ Tue Oct 20, 2015 8:45 pm ]
Post subject:  Re: A Go Survey for Statistics Class

Female
18
4 kyu
2 years
yes
positively
yes
(I don't play much, but I do many things outside of playing so bleh) 2-3
Fast

Author:  Shaddy [ Tue Oct 20, 2015 9:27 pm ]
Post subject:  Re: A Go Survey for Statistics Class

Male
22 years old
4 dan KGS
8 years
Yes
Negatively
Yes
>10
Fast

Author:  jeromie [ Tue Oct 20, 2015 9:33 pm ]
Post subject:  Re: A Go Survey for Statistics Class

1) What is your gender? Male

2) What is your age? 34 Years old

3) What is your current Go rank? 6 Kyu (KGS)

4) How long have you been playing Go? 4.5 Years

5) Have you taken breaks from playing Go? Yes

If yes, how did it affect your Go games? Neutral (It slowed my growth, but I didn't regress while taking breaks.)

6) Do you do Go problems every week? Yes

7) On average, how many hours do you spend playing Go in a week? 3 Hours per week

8) Do you prefer fast or slow games? Slow (Though medium is probably my real preference. I like games that are about an hour long, and don't generally enjoy playing people with fast time settings as much. Playing bots quickly is ok, though.)

Author:  EdLee [ Tue Oct 20, 2015 11:39 pm ]
Post subject: 

Hi scaptude, may I ask which level of stats class ?
( High school, undergrad, grad school, etc. ... )

Author:  DrStraw [ Wed Oct 21, 2015 2:29 am ]
Post subject:  Re: A Go Survey for Statistics Class

I am a little skeptical when someone creates an account just to ask for personal data. I think some information about yourself and posting history is more appropriate.

Author:  scapetude [ Wed Oct 21, 2015 6:48 pm ]
Post subject:  Re: A Go Survey for Statistics Class

Thank you for participating!

EdLee wrote:
Hi scaptude, may I ask which level of stats class ?
( High school, undergrad, grad school, etc. ... )

Hi EdLee,
It is for an undergrad Intro to Stat class.

DrStraw wrote:
I am a little skeptical when someone creates an account just to ask for personal data. I think some information about yourself and posting history is more appropriate.

Hello DrStraw,
What you've stated is understandable. Unfortunately, I won't be able to provide a posting history since I do not have one. However, to know me a bit more, these are my answers to the survey questions:

1) What is your gender? Female
2) What is your age? 27 Years old
3) What is your current Go rank? 2 Kyu AGA
4) How long have you been playing Go? 5 Years
5) Have you taken breaks from playing Go? Yes
If yes, how did it affect your Go games? Positively
6) Do you do Go problems every week? Yes
7) On average, how many hours do you spend playing Go in a week? 1 Hour per week
8) Do you prefer fast or slow games? Slow

Author:  DrStraw [ Wed Oct 21, 2015 7:05 pm ]
Post subject:  Re: A Go Survey for Statistics Class

scapetude wrote:
Thank you for participating!

EdLee wrote:
Hi scaptude, may I ask which level of stats class ?
( High school, undergrad, grad school, etc. ... )

Hi EdLee,
It is for an undergrad Intro to Stat class.

DrStraw wrote:
I am a little skeptical when someone creates an account just to ask for personal data. I think some information about yourself and posting history is more appropriate.

Hello DrStraw,
What you've stated is understandable. Unfortunately, I won't be able to provide a posting history since I do not have one. However, to know me a bit more, these are my answers to the survey questions:

1) What is your gender? Female
2) What is your age? 27 Years old
3) What is your current Go rank? 2 Kyu AGA
4) How long have you been playing Go? 5 Years
5) Have you taken breaks from playing Go? Yes
If yes, how did it affect your Go games? Positively
6) Do you do Go problems every week? Yes
7) On average, how many hours do you spend playing Go in a week? 1 Hour per week
8) Do you prefer fast or slow games? Slow


Okay, as a college math professor who has regularly taught Introductory Statistics I have some questions for you.

You are asking here for a probably very small volunteer sample. Has your instructor told you about how unreliable volunteer samples are? And has he explained about how sample size severely affects the reliability of results. To produce anything of significance you need to have a sample size of around 35. Are you likely to get that from a posting here? Even if you do, it is a volunteer sample and so not reliable.

I am not criticizing you for trying to gather data. I am just pointing out that you need to be aware of the reliability of what you may gain from posting here. I hope you have other sources of data.

Author:  EdLee [ Wed Oct 21, 2015 7:11 pm ]
Post subject: 

scaptude wrote:
7) On average, how many hours do you spend playing Go in a week? 1 Hour per week
8) Do you prefer fast or slow games? Slow
Hi scaptude,

Thanks for the info. I also prefer slow games,
and one hour (shared between two people) is very fast (way too fast) for me. :)
I like the US Open time setting of 90 initial minutes per person.
But it's also understandable time is limited for a busy undergrad. :)

Good luck. :mrgreen:

Author:  scapetude [ Wed Oct 21, 2015 7:22 pm ]
Post subject:  Re: A Go Survey for Statistics Class

DrStraw wrote:
Okay, as a college math professor who has regularly taught Introductory Statistics I have some questions for you.

You are asking here for a probably very small volunteer sample. Has your instructor told you about how unreliable volunteer samples are? And has he explained about how sample size severely affects the reliability of results. To produce anything of significance you need to have a sample size of around 35. Are you likely to get that from a posting here? Even if you do, it is a volunteer sample and so not reliable.

I am not criticizing you for trying to gather data. I am just pointing out that you need to be aware of the reliability of what you may gain from posting here. I hope you have other sources of data.

Thank you for your inputs. I am aware of the unreliability of convenience samples. I am not expecting a large sample size from a posting here, and will also get my data elsewhere. My instructor is also aware of how convenience samples will affect the project. Since he has approved it, I have decided to go forward with it. But, I am open to suggestions.

Author:  DrStraw [ Wed Oct 21, 2015 7:30 pm ]
Post subject:  Re: A Go Survey for Statistics Class

scapetude wrote:
Thank you for your inputs. I am aware of the unreliability of convenience samples. I am not expecting a large sample size from a posting here, and will also get my data elsewhere. My instructor is also aware of how convenience samples will affect the project. Since he has approved it, I have decided to go forward with it. But, I am open to suggestions.


If you are open to suggestions then I would suggest that you look at all the people who have posted here during that last week and directly PM them with the questions. This may seem like a small difference, but if you specifically target regular contributors you are much more likely to get a non-volunteer sample and also get more responses. You should, of course, explain the reasoning behind your request. It would also be a good idea to clear such a mass PM with the moderators here to make sure that it is within the TOS.

Just simply asking in an open forum is a notoriously bad way of gathering data. Those motivated to respond are more likely to be bias in some form or other.

Author:  scapetude [ Wed Oct 21, 2015 7:33 pm ]
Post subject:  Re: A Go Survey for Statistics Class

DrStraw wrote:
If you are open to suggestions then I would suggest that you look at all the people who have posted here during that last week and directly PM them with the questions. This may seem like a small difference, but if you specifically target regular contributors you are much more likely to get a non-volunteer sample and also get more responses. You should, of course, explain the reasoning behind your request. It would also be a good idea to clear such a mass PM with the moderators here to make sure that it is within the TOS.

Just simply asking in an open forum is a notoriously bad way of gathering data. Those motivated to respond are more likely to be bias in some form or other.

I haven't thought of it like that. Thank you for the suggestion.

Author:  Abyssinica [ Wed Oct 21, 2015 10:55 pm ]
Post subject:  Re: A Go Survey for Statistics Class

I was told you needed 30+ not 35, Drstraw.

Author:  Uberdude [ Thu Oct 22, 2015 12:15 am ]
Post subject:  Re: A Go Survey for Statistics Class

There is no magic number 30 or 35 that makes a good sample size. It depends on the population size, with what confidence you want the result, the type of question...

In what context were you taught/did you teach this magic number Abyssinica/DrStraw?

Author:  wineandgolover [ Thu Oct 22, 2015 12:58 am ]
Post subject:  Re: A Go Survey for Statistics Class

DrStraw wrote:
I am a little skeptical when someone creates an account just to ask for personal data. I think some information about yourself and posting history is more appropriate.

My reservation is slightly different. It is highly unusual to ask people to post private information in a public forum, and I wouldn't comply with the request even if I were sympathetic, which I am. I'd strongly recommend using an online survey tool like surveymonkey. This should:
1) keep your respondents' replies confidential,
2) increase your response rate, and
3) provide you the results in a spreadsheet-compatible format.

Please do not spam direct message frequent posters. So long as your teacher is aware of and approves sample shortcomings, you'll be cool. (Consider addressing them in your report.)

You could also consider offering to post the final results or report.

Good luck!

Author:  EdLee [ Thu Oct 22, 2015 2:37 am ]
Post subject:  Re: A Go Survey for Statistics Class

scapetude wrote:
But, I am open to suggestions.
Hi scapetude,

Your class is likely covering something of this nature: sample size determination.

If you happen to be in the vicinity of Los Angeles this weekend, you can poll a few Go people at the annual Cotsen tourney. :)

Author:  Pio2001 [ Thu Oct 22, 2015 3:42 am ]
Post subject:  Re: A Go Survey for Statistics Class

Hi scapetude
1) What is your gender? Male
2) What is your age? 40 Years old
3) What is your current Go rank? 8 Kyu (kgs)
4) How long have you been playing Go? 3 Years
5) Have you taken breaks from playing Go? Yes
If yes, how did it affect your Go games? Negatively (in fact it had no effect since I was beginner)
6) Do you do Go problems every week? Yes
7) On average, how many hours do you spend playing Go in a week? 9 Hours per week
8) Do you prefer fast or slow games? Slow

wineandgolover wrote:
It is highly unusual to ask people to post private information in a public forum, and I wouldn't comply with the request even if I were sympathetic, which I am. I'd strongly recommend using an online survey tool like surveymonkey. This should:
1) keep your respondents' replies confidential,
2) increase your response rate, and
3) provide you the results in a spreadsheet-compatible format.


For me, it is rather the opposite : sharing information with other members is the only interest I see in answering. I wouldn't have bothered posting anonymous information for a project that I'll never hear about again.

But now, if the final stats are published here at the end, it is different. Then I'd be interested in posting, be it anonymously or publicly, to help build the statistics and see how mine compares to the average.

Author:  sybob [ Thu Oct 22, 2015 7:21 am ]
Post subject:  Re: A Go Survey for Statistics Class

DrStraw wrote:
I am a little skeptical when someone creates an account just to ask for personal data. I think some information about yourself and posting history is more appropriate.

I agree.
And there is an other reason why I have not responded (yet).
I find the questions a bit awkward.

What is the purpose of this enquiry? What will you be trying to derive/conclude?
Do you have a theory or hypothesis, like: taking a break positively effects go strength? Or: a break positively effects go strength for older people, but not for younger ones (or man vs. woman?
Are you interested in mean answers, or distribution?
Do you want to distinguish kyu players from dan players?
And, who will use (for what purpose) the results? Are you a tour operator offering goplayers a break or what? Or perhaps a go tournament organizer who wants to decide if it is to be a slow or fast game tournament?

I also think not all questions are not adequately worded. What do you consider a break? A week, a month? A year or more? And how should I answer if I like both fast and slow games, either once in a while?

I do not want to sound sceptical or critical, but I think such a survey is not very meaningful.
Good luck.

Author:  Abyssinica [ Thu Oct 22, 2015 7:48 am ]
Post subject:  Re: A Go Survey for Statistics Class

Uberdude wrote:
There is no magic number 30 or 35 that makes a good sample size. It depends on the population size, with what confidence you want the result, the type of question...

In what context were you taught/did you teach this magic number Abyssinica/DrStraw?



The centrral limit theorem for approxamate normality

Author:  sparky314 [ Thu Oct 22, 2015 8:22 am ]
Post subject:  Re: A Go Survey for Statistics Class

Also, a big factor: where you're retrieving the data. The population that frequents a Go forum is likely to be a little more dedicated towards Go than the average Go player. That'd make the statistics (if there was a large enough sample size) strongly influenced by this particular forum.

I'd prefer an anonymous survey...but, since I find this interesting...

1) What is your gender? Male

2) What is your age? 31

3) What is your current Go rank? 6 kyu KGS (barely...)

4) How long have you been playing Go? 10 months (non-consecutive)

5) Have you taken breaks from playing Go? Yes

If yes, how did it affect your Go games?
No effect. I learned how to play last year, but only played a few games over 2 months before stopping. Then I started again in March.

6) Do you do Go problems every week? Yes, daily

7) On average, how many hours do you spend playing Go in a week? Play games: ~3 hrs, Go problems/books: 4-5 hours

8) Do you prefer fast or slow games? Concrete definitions here would be preferred. I only play online, but on KGS I prefer 25 min + 30s byo-yomi, as it gives me more time to read and play. I'll play 10 min + 30s byo-yomi in order to train myself to read quicker.

Page 1 of 2 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/