4
$\begingroup$

I just logged into stack exchange for the first time to answer a very basic question about logistic models in R. I found out that I cannot answer the question because I don't have enough reputation points. So, I am curious what are the fastest ways to gain the 50 reputation points needed to answer this basic question about how to get a pseudo $r^2$ for a logistic regression?

The question is: How to calculate pseudo-$R^2$ from R's logistic regression?

The answer is just to use the pscl package with the pR2 command but I am really confused about how to get more points. Any useful tips for increasing reputation are appreciated.

| |
$\endgroup$
  • 2
    $\begingroup$ I cannot reproduce this circumstance: according to the help, you need only the minimum of one point to answer a question. How exactly did you try to answer and what caused that to fail? $\endgroup$ – whuber Jun 20 '14 at 15:26
  • $\begingroup$ I clicked on add comment $\endgroup$ – user48729 Jun 20 '14 at 16:03
  • $\begingroup$ Which apparently I can do on my own questions but not on other people's posts until I have 50 points. $\endgroup$ – user48729 Jun 20 '14 at 16:04
  • $\begingroup$ Playing around some more with voting up led me to my answer - stats.stackexchange.com/help/whats-reputation $\endgroup$ – user48729 Jun 20 '14 at 16:12
  • 3
    $\begingroup$ It's true. You must have 50 rep to 'comment everywhere'. The fastest way, in my opinion, is to provide a well argued answer. I think there are more questions than answers generally. In this case you are referring to, you can provide this as the answer in the 'Your Answer' text box. If you add a little bit more explanation to it than 'use this software' someone will upvote it which is worth 10 points. If you are lucky the person asking the question will also check mark that your answer helped which is worth 25 points. Or you can suggest 25 edits: 2 points * 25 = 50 points. $\endgroup$ – Meadowlark Bradsher Jun 20 '14 at 16:14
  • $\begingroup$ I guess that I am just still confused about how to get off of 1 reputation point because a user needs 5 user points to "answer" a question. I guess that the only way is to ask good questions? $\endgroup$ – user48729 Jun 20 '14 at 16:18
  • 2
    $\begingroup$ According to this page you should be able to provide an answer, you just can't provide a comment until 50. At 5 you can participate in meta but you are here already because your question was migrated. I don't see anything or recall anything that required 5 points to answer a question. $\endgroup$ – Meadowlark Bradsher Jun 20 '14 at 16:28
13
$\begingroup$

I think you are confused between answering a question in a comment (50 rep) and answering as the site intends for everyone to be able to do (1 rep). This response was done in the text box below that you see as 'Your Answer' with several formatting options readily on display. This is where you should be able to provide answers even with only 1 rep.

It appears that what you are attempting to do is answer in a comment which is where I had previously answered you. This requires 50 reputation and is generally not supposed to be where an answer that is intended to be final or complete should go.

Providing answers in comments is for partial answers, asking for clarification, interrogatively arriving at answers, providing guesses and other interactions that are not explicitly intended to be fully authoritative.

Try providing your answer in the huge text box instead of a comment and it should work. Lastly, if the answer I have provided you here was helpful please check the check mark next to the answer.

Welcome to Cross Validated!

| |
$\endgroup$
  • $\begingroup$ Thank you meadowlark but you need to have 5 reputation points to answer in the "Your Answer" text box. Until you have gained 5 reputation points, you see "You must have at least 5 reputation on Cross Validated to answer a question." Which according to the StatExchange email that I received a few minutes ago, indicates that you first need to have an upvoted post in order to be able to answer a question. So, would you be willing to upvote this post just to let me answer posts? I am a stats prof and I am extremely amused (and somewhat distracted) by this minor inconvenience. $\endgroup$ – user48729 Jun 20 '14 at 17:37
  • 2
    $\begingroup$ We are sympathetic and really appreciate your bearing with us. We (the community members) have little control over these basic aspects of how the SE system works, but this thread helps me appreciate the wisdom of most of the restrictions. You ran into trouble because you were trying to create material that would have taken considerable intervention on the community's part to clean up. I hope you won't mind spending a few minutes reading over our help center: it will help you decide if this is a good place for you to be and it will help you get up to speed quickly if you decide to participate. $\endgroup$ – whuber Jun 20 '14 at 17:54
  • 1
    $\begingroup$ whuber, I believe (not sure) you'd need to ping @user48729, for calling his attention on someone else's post. $\endgroup$ – Andre Silva Jun 20 '14 at 18:23
  • $\begingroup$ @user48729 I had already upvoted this question but in 'meta' upvotes do not increase reputation. $\endgroup$ – Meadowlark Bradsher Jun 20 '14 at 18:37
  • $\begingroup$ @whuber I am curious to know what exactly happened. The documentation says he should be able to answer but he was not? I can't see the question he is referring to. Is there a link? $\endgroup$ – Meadowlark Bradsher Jun 20 '14 at 18:38
  • $\begingroup$ @MeadowlarkBradsher The original post was here: stats.stackexchange.com/questions/82105/… But, I have tried to answer other questions today and have found this note at the bottom of my page "You must have at least 5 reputation on Cross Validated to answer a question." Thanks for upvoting my post and then explaining why it isn't raising reputation points. $\endgroup$ – user48729 Jun 20 '14 at 18:45
  • $\begingroup$ @MeadowlarkBradsher would it be appropriate to post this to the stats.stackexchange.com site? I guess that I am now confused by the difference between the meta site and the regular site. Thank you for being a good community citizen and pointing out many of these basic things. $\endgroup$ – user48729 Jun 20 '14 at 18:52
  • 7
    $\begingroup$ User48729: (1) according to our help, you need only 1 point to answer a question on the main site, but 5 to answer here on meta. The role of the meta site is explained at stats.stackexchange.com/help/whats-meta. @Andre: One does not need to ping either the thread originator nor the answerer; in fact, an attempt to do so at the beginning of a comment will automatically be deleted! Both those people will always be notified of any comments made. Pinging serves to notify others who might have been part of a comment conversation. $\endgroup$ – whuber Jun 20 '14 at 19:07
  • $\begingroup$ This answer is a good example of how to make more growth out of smaller value questions. When the answer you provide is compelling then it grows in value over time. When it evokes discussion there is also point growth there. The natural product of these is a rule that for a sufficiently good question (not hard to find here), a great answer can elicit response and value that increases over time. Best of luck. $\endgroup$ – EngrStudent Jul 3 '14 at 15:40
8
$\begingroup$

To clarify (though all the parts of this information is here in various places):

  • On the main site, you can normally answer a question with 1 reputation.

  • Here on meta you need 5 reputation to answer

  • You need 50 reputation to comment in either location.

$\,$

To gain reputation quickly:

  • answer questions on the main site. Good answers will generally get upvoted; this is the fast route to reputation. An average answer should typically get you about 30 points or so (but even an experienced answerer with a higher average will get 0 points sometimes). [However, it might take a while to figure out the kinds of answers that tend to be worth more - my average was below the site average for a long while. Other people may be faster on the uptake than I was.]

    There are many unanswered questions, so there's plenty of opportunity for points. An even slightly determined person can get 50 points in a session.

  • ask good questions on the main site. This is often a little slower, but a very good question gets a lot of attention, and one that attracts good answers gets a lot more. You don't need to be able to answer questions to garner a lot of reputation.

  • if you have good reputation on other SE sites and the site knows they belong to the same person (i.e. you register the accounts from the same email), you should get a bonus (this one doesn't look like it applies to you but may apply to other people who read this question).

There's a list of sources of reputation in the help (I know you already found this though - again, I am putting it here in an answer for later readers).

| |
$\endgroup$
  • 1
    $\begingroup$ I guess it also helps to answer the pure statistics questions. At least in my experience answers to econometrics questions attract much less attention and upvotes. $\endgroup$ – Andy Jun 21 '14 at 12:29
  • $\begingroup$ R related question were good hunting ground, but nowadays the easiest ones are moved to the stackoverflow. Econometrics and time series really do attract much less attention. $\endgroup$ – mpiktas Jul 2 '14 at 8:21
  • 1
    $\begingroup$ Among my top 100 tags, I find simulation (0.33 upvotes per answer, 24 answers) does much worse for me than time-series (2.4 upvotes per answer, 48 answers), in spite of my feeling that simulation is a subject I am a bit more knowledgeable on. However, they're both down the low end (in my top 100 tags my average is roughly twice that for time series.) $\endgroup$ – Glen_b Jul 2 '14 at 9:30
6
$\begingroup$

You can also earn reputation by editing the questions and answers. Your edit will be placed in a queue, but on stats.SE the queues disappear quickly. You can read more about how editing works here: How do suggested edits work?

| |
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .