Why are so few questions being answered?

[SlideRule]

I am a software developer, currently in graduate school for general mathematics, who intends a career in statistical programming. I have begun taking on small statistical programming projects and I have questions that could benefit from the advice of the many of you who have good experience.

While lurking in Talk Stats I am often delighted to find clever questions posed, and I look forward eagerly to see what answers are given, but unfortunately many questions asked here go unanswered. Why is that?

Are the questions too hard? Are they stated poorly? Most of what I see seems to be well stated and not too advanced, yet few responses are forthcoming. Are the answers too obvious?

Some technical questions like, "How do I do such-and-such in SAS?" do get answered, but questions about mathematical statistics or general applied methods seem ignored.

Those of you who are "regulars," please give your suggestions on what we newbie members can contribute to Talk Stats to invigorate discussion. Thank you.

 

[bugman]

I am less familar with the mathmatical side of things. Generally, I will reply on questions realting to statistical test chice or study design and others related to biological applied stats. I tend to leave PDF questions and equation based questions alone.

 

[SmilingSara]

Learn to Google first

My issue is that there are posts where the author has clearly made no attempt to find the solution first (even though we have that sticky note at the top of the threads). I'm a Google hound so I know that there's a lot out there that would clearly answer many posters questions, yet they do not indicate that they've done any work on their end before they make the post.

 

[TheAnalysisFactor]

Hi Sara,

My experience and interest is in helping researchers analyze their data. So I guess I do have criteria about kinds of questions to answer.

- I don't answer homework questions, both because I often don't remember the details of the mathematics, and because I figure students have TAs, but researchers often have nobody.

- There are certainly a number of questions I just don't know the answer to.

- Many questions are too hard to answer in this format, not necessarily because the questions are stated poorly, but because there isn't enough detail. I'd have to ask so many questions about the design, the purpose of the study, etc., that it would take forever to write back and forth. There are a lot of situations in statistics where the devil is in the details. There are a lot of times I want to say, "you need to TALK to a statistician," but that seems too rude. So I just keep quiet.


Karen

 

[SlideRule]

Reasons questions might not receive an answer

In the year-plus that I have been a visitor to this website I've found that there are many reasons why a question here might go unanswered, or poorly answered. Here are a few:

1. The question makes clear sense to the poster who asked it, but is incomprehensible to anyone not intimately familiar with the problem under study.

2. In order to answer the question accurately would require asking 20 more questions of the poster to gain necessary details that were omitted.

3. The simple-seeming question has a simple answer that would take four-to-eight years of college studies to understand.

4. Even with years of college studies the answer would require very long explanation.

5. The question appears to be a homework question that the poster could easily answer for themselves through studying their school assignment or through the aid of their school's resources.

6. And finally, the problem may simply be one that nobody here knows the answer to.

I ask you frequent forum members if you have suggestions of how we can help new visitors to express their questions in ways that may better bring forth useful answers.

I ask that especially for those "non-statistician" graduate students who seem out of their depth when posing their questions here. I often find their questions to be among the most interesting, and I am usually curious myself to know more about the problems they are struggling with and what directions may best lead to the answers.

And to you posters, let me say thank-you to those who post data files. I enjoy studying your data, trying to comprehend your problems, discovering possible answers for myself, and then I wait patiently for someone more knowledgeable than me to post a correct approach. As a graduate student myself, it is a pleasure to sometimes puzzle over problems that I don't actually have to solve, and it is always fun to have real-world data to explore, instead of textbook stuff.

 

[gianmarco]

I totally agree with the above replies. As for me, as archaeologist with a non deep knowledge of math, I limit myself to reply to posts regarding issues that match my general inclinations and interest: mainly descriptive stats, hipothesis testing (parametric and non-parametric), multivariate analysis and expecially correspondence analisys.

Regards
gm

 

[TheEcologist]

I ask you frequent forum members if you have suggestions of how we can help new visitors to express their questions in ways that may better bring forth useful answers.

I think you have started a very useful topic (esp. for new visitors). I will make this tread a sticky, as I believe this would aid new visitors. Feel free to add more ideas and suggestions.

 

[Mike White]

I will answer questions relating to writing code for R and S-Plus or formulas and macros in Excel. I can also help with some multivariate analysis problems like clustering and PCA, which I use a lot in my work. However, I am a scientist not a statistician so I tend to avoid questions about statistics or the interpretation of the output from statistical functions.

I agree than many members could solve their problems themselves if they made more of an effort to search for a solution, however that does not prevent me from replying because hopefully the posted solutions will be helpful to many other people.

It is also helpful if members posting queries include an example of the data they are using.

 

[Dason]

I'm pretty new here but I was wondering if there was some sort of forum rules thread. I couldn't find one but I do think that if there was something that laid of a few guidelines for posts it could help out.

I know that I've completely ignored posts that are clearly homework problems (they're too perfectly set up to NOT be homework) and the poster hasn't put forth any sort of effort to answer the question or shown any previous work.

I guess I was wondering if there was something like a rules/guidelines thread that would lay out the guidelines. A few things like "Clearly label if the question is for homework and show any previous work you've attempted" "Clearly lay out the problem and give any and all relevant information that could help in answering the question"

A lot of these things seem like common sense but without listing what is expected it's harder to expect everybody to follow the unwritten rules. Then if somebody doesn't follow the rules we can kindly send them over to the rules thread and ask them to elaborate or show their work instead of completely ignoring them.

Edit: I saw the sticky in the first two sections of the forum for homework help but should there be a forum wide announcement laying out the general rules?

 

[TheEcologist]

Edit: I saw the sticky in the first two sections of the forum for homework help but should there be a forum wide announcement laying out the general rules?

I would second, support and sticky a post from anyone who can come up with a general set of forum 'posting guidelines' specific to our various forums. Of course on condition that other 'regulars' agree on these.

If anyone wants: feel free to start with a set of 'guidelines' (I woundn't call them rules).

We can then start a 'peer review' process among the regulars, to come up with a set of general and fair guidelines for posting homework questions.

What do you guys/girls think? Would it add to the forum or are quark's basic rules here sufficient?

 

[bugman]

I think that Quark has covered most of the important stuff, but may i suggest (follwing from Mike Whites suggestion) that when posting in the R forum, that reference to the appropriate package is given and that the code is "clean" for those trying to offer assistance, this makes things much easier and there is more chance that the poster will get a reply.

Phil

 

[CowboyBear]

I agree with Phil that Quark's posts pretty much cover the important stuff - though perhaps we could add one or two things, like a sticky with guidelines for the R forum perhaps?

In general my suggestion to people asking questions (esp. consulting questions) would be make your question interesting. I don't mean interesting necessarily in a statistical sense, but tell us what you're actually trying to find out and why it matters!

I really think simply keeping regulars' interest is a big issue here - for each of us there's a narrow window between boring questions we've seen a thousand times before, and questions that are too hard for us to answer expertly on (or not without a fair bit of work!) If posters want regulars to venture outside this window, they really need to convince the reader that their research question is substantively important enough to justify spending time on.

(maybe this could be something to add to the consulting forum in specific?)

 

[bugman]

Nice point Cowboybear!

 

[Link]

I'm pretty new here, but don't mind helping at all and am more than glad to share my knowledge and experience. That being said...

Two points mentioned already that I would like to emphasize on are:

1) Put in a significant amount of effort into solving your question first.
2) Be clear and detailed in what your question is.

I tend to ignore posts where i) people either just want a question answered that they could at least have easily googled or solved after reading a textbook, or ii) they don't even know what they're trying to ask or do. Try to understand what it is you really want solved before you ask.

On a side note, I (and most likely the other knowledgeable members here) initially came here to share and expand my knowledge. If it comes to the point where I become mainly a tutor answering homework questions, it's likely that I'll lose my motivation for coming here. So please, even if you're less knowledgeable, please try to contribute too.

 

[hlsmith]

Look at Dason giving out thanks like nobody's business.


This thread contains 2.5% of his thanks. His "Thank" program must have had a glitch that day, or that capability is now in reserve mode.

 

[spunky]

Look at Dason giving out thanks like nobody's business.


This thread contains 2.5% of his thanks. His "Thank" program must have had a glitch that day, or that capability is now in reserve mode.

why did you summon this back from TEH DEADZ!?

it's been 5 yrs already! LET IT DIE! LET IT DIE!

 

[bryangoodrich]

Dason had a fresh niceness chip inserted five years ago. Unfortunately, it hasn't been replaced and its power output has degraded considerably.

 

[Dason]

This thread contains 2.5% of his thanks.

I thought you were joking but the numbers check out.

but since you pointed it out...
NO THANKS FOR YOU

 

[hlsmith]

There must be a virus embedded in this thread. BG has given 0.9% of his overall thanks within in this thread.




Dason, just getting you to acknowledge the Thanks button exists or is Thanks enough for this 12 month human calendar cycle. Though I wouldn't mind seeing your use of thanks over time graphed. Did it plummet when you were jaded by Hamid?

 

[bryangoodrich]

I don't respond to too many threads, so if I'm engaged in one, maybe constituting 1% is admissible behavior