5
$\begingroup$

For some reason I never liked the number $180$, so

enter image description here

and together with the 90-days and 120-days and 150-days numbers

\begin{array}{| r | r |} \hline \hline \text {Metric} & \text{90 B-days} & \text{120 B-days} & \text{150 B-days} & \text{181 B-days}\\ \hline \text {Q per day} & 5.4 & 4.5 & 7.4 & 5.1\\ \text{% answered} & \text {80%} & \text {79%} & \text {80%} & \text {82%}\\ \text{avid users | %} & 54 | \text {5.5%} & 58 | \text {4.8%} & 63 | \text {4.3%} & 86 | \text {5.1%}\\ \text {total users} & 981 & 1210 & 1455 & 1689\\ \text {Engaged users %(*)} & 34\text{%} & 37\text{%} & 37\text{%} & 38\text{%}\\ \text {answer ratio} & 1.8 & 1.7 & 1.7 & 1.6\\ \text {Visits/day (median)} & 192 & 225 & 308 & 496\\ \hline \end{array}

(*)"Engaged Users" are those that do not have reputation $1$ or $101$, i.e. those that had at count day performed some reputation-affecting activity on the site (I count out also the $101$'s because the $100$ points are SE-registration bonus if they are active on other sites). I decided to show "engaged" rather than "unengaged" users, since all other metrics in the table have the association "larger value = better".

(P.S. The 30-days numbers can be found here, and the 60-days numbers here

COMMENTARY

Median Visits/day were just short of $5$ to be characterized as "Okay", and very much increased compared to one month before.

Participation increased (both the % Avid Users and the % Engaged Users), given also that the rate of new registrations per day continues to be unbelievably constant during all these 6-month period (7-9 new accounts per day, month per month, what kind of spooky stationary stochastic process is this?)

Economics.SE starts to acquire a life of its own, I think.

I will continue to monitor monthly, but maybe I should leave the site to show us its seasonality (by reporting again in 3-months from now, in mid-August).

$\endgroup$
2
  • 3
    $\begingroup$ I only worry about questions. The number dropping coincides with the decrease in standard Keynesian questions from first/second year undergraduates. In general, I feel that we're getting less "expert questions". $\endgroup$
    – FooBar
    May 20, 2015 at 14:19
  • 1
    $\begingroup$ And now we can celebrate not having any red status on Meta, even if it partly may come from our single high traffic question :) $\endgroup$
    – FooBar
    Jun 8, 2015 at 15:28

2 Answers 2

Reset to default
4
$\begingroup$

My suggestion is that we should try to ask and answer econometric questions, both basic and advanced. If we do this then this could cover that niche and increase the numbers by a large amount. In particular how about the most cited work in econometrics robust standard errors and the most cited recent paper Arrellano Bond Dynamic GMM models

$\endgroup$
10
  • 2
    $\begingroup$ That's a specific and targeted idea, with which I personally agree. $\endgroup$
    – Alecos Papadopoulos
    May 21, 2015 at 22:32
  • $\begingroup$ That niche is already well covered by another site of this family: stats.stackexchange.com $\endgroup$
    – user4239
    May 22, 2015 at 5:04
  • $\begingroup$ @AndréPeseur Are you seriously suggesting that we should not consider econometrics questions in economics.SE? $\endgroup$
    – Alecos Papadopoulos
    Jun 8, 2015 at 15:31
  • $\begingroup$ I think that econometrics is sufficiently differentiated from statistics, think monopolistic competition $\endgroup$
    – user157623
    Jun 8, 2015 at 17:30
  • $\begingroup$ I AM SERIOUS. YES. The site I mention is really excellent for question about statistics. They have an econometrics tag with currently 671 questions stats.stackexchange.com/questions/tagged/econometrics More than 3600 questions with a time-series tag, almost 7500 tagged with regression etc. $\endgroup$
    – user4239
    Jun 8, 2015 at 18:00
  • $\begingroup$ "econometrics is sufficiently differentiated from statistics" Ah? Econometrics is the application of some statistical method (read: regression) to "economic" problems. And I think that the line should be drawn here. Questions about methods should go on the statistics site and questions about the economic content should go here. That's an efficient divide. $\endgroup$
    – user4239
    Jun 8, 2015 at 18:03
  • $\begingroup$ @AndréPeseur Why there should be a divide? In what sense it will be "efficient"? $\endgroup$
    – Alecos Papadopoulos
    Jun 8, 2015 at 18:50
  • $\begingroup$ Faster and better answers. Have a look at the other site. $\endgroup$
    – user4239
    Jun 8, 2015 at 19:25
  • $\begingroup$ I respectfully disagree, I think it is just not only about having faster and better answers if we do not have sufficiently specialized topics, think GMM-SYS Arellano Bond Blundell or BLP, more important IV, RD and diff in diff all the Angrist like literature, an econometrician is more likely to know more about that than an average statistician. In any case, I think that in the spirit of the market, people should be let to choose freely where to post their questions and where to answer them. Forced specialization at this stage I think is too hasty. $\endgroup$
    – user157623
    Jun 8, 2015 at 19:32
  • 1
    $\begingroup$ I already see a ton of Arellano-Bond / Blundell-Bond questions on cross validated, and virtually none here. 'fraid that ship may have sailed... Also few of those questions involve what I see as the real economics of those estimators (when are they appropriate, what do their results mean) and instead ask questions that are basically answered by pointing to the J-stat. To the extent there's a meaningful distinction, that sounds more like statistics to me. $\endgroup$
    – jayk
    Jun 15, 2015 at 0:22
0
$\begingroup$

Just a few thoughts:

  • The number of visits per day is steadily increasing. That's the real positive thing here.

  • The number of users is increasing too, but at a decreasing rate. Not so good.

  • As compared to 150 days, there has been a surge in the number of avid users: +37%, as compared to 150 days. Having a look at the past evolution and the current list of users, this is probably more an outlier than anything else. Unfortunately. This means that the site needs to attract and retain more new users. The reticence to upvote questions and answers is probably part of this problem.

  • The number of questions per day and the percentage of answered questions are pretty constant. One question out of five doesn't receive an answer. That doesn't give a good signal to newcomers.

$\endgroup$

You must log in to answer this question.

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