11
$\begingroup$

While searching for a "duplicate" question, the only post I found worth mentioning was this one: Decline in activity which shows a "peak" in 2017 followed by a decline until 2019 (which is when the post was written). What I have noticed is something slightly different, since I was looking at questions/day as archived sporadically in the months of August (or the closest snapshot to August) each year on the WayBackMachine. I actually see a rise in activity from 2018 to 2019 and even 2020, but then a rapid decline from 2020 to now.

Site Questions/Day Year and hyperlink for proof
Chemistry 11 2022
Chemistry 14 2021
Chemistry 20 2020
Chemistry 23 2019
Chemistry 17 2018
Chemistry 20 2017
Chemistry 22 2016
Chemistry 18 2015
Chemistry 9.3 2014
Chemistry 3.7 2013
Chemistry Didn't exist 2012

Using the same archived pages as above, it might be useful to compare to SO and other science SE sites, so I will do that below.

Site Questions/Day in 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
StackOverflow 5.2k 6.6k 7.3k 7.7k 7.5k 7.5k 7.1k 6.6k 7.1k 5.7k 5.4k
Mathematics 175 386 361 446 399 438 375 465 455 339 277
Statistics 27 49 68 82 90 96 93 85 86 61 66
Physics 19 47 51 73 74 70 72 77 81 75 62
MathOverflow TBD TBD 32 36 34 35 34 34 40 31 30
Chemistry Didn't exist 3.7 9.3 18 22 20 17 23 20 14 11
Computer Science 6.7 11 8.8 12 11 16 15 22 12 12 7.9
Biology 5.9 5.6 10 12 12 18 11 10 8.1 5.7 5.1

When looking at earlier months of 2020, it is apparent that there was an increase in SE activity overall during the first few months of strict COVID-19 related lockdowns (e.g. in March to July 2020), if we remove that factor which might have inflated the 2020 number, what we're seeing is a decline in activity from a "peak" of 23 questions/day in August 2019 to a trough of 11 questions/day in August 2022 (somewhat ironically since the only related post I found, which is the one to which I provided a link at the beginning of this post, observed a "peak" in 2017 and a "trough" in 2019). I'm aware that SE got in a legal dispute with a diamond moderator later in 2019 which caused many diamond moderators to resign, but I'm asking if anyone might have alternative theories or supplementary or complementary theories regarding the decline in activity here from 2019 to now.

The second table suggests that other sites have experienced this too. Maybe people left scientific and technical jobs during the pandemic? That doesn't explain why fewer questions from students would be asked though.

$\endgroup$
3
  • 4
    $\begingroup$ My guess is the culture. It went from a place to ask and answer questions freely to a place to be scrutinized at every angle. Members of my Facebook group no longer like to participate here due to the negativity on this SE site. I can't name anyone in particular, but it seems like we're missing people who were here years ago that made the site worth using. $\endgroup$ Aug 10 at 20:40
  • $\begingroup$ @MelanieShebel I think that might be the best answer so far, especially if more details and/or examples were added. I can't click accept on a comment though! I'm not saying I've experienced the same thing, but if what you're saying is true then it's a better explanation that just "all good questions have been asked already" or "the core users left" or "Monica", since none of those theories, even collectively, explain the data observed in my post. $\endgroup$ Aug 10 at 20:44
  • 4
    $\begingroup$ @MelanieShebel, You are right. It could be the negative energies here. There is one person who habitually downvotes every other question in SE. Chemistry and rarely writes an answer. What a useful contributor. Of course, new comers will run away. $\endgroup$
    – AChem
    Aug 10 at 22:24

5 Answers 5

12
$\begingroup$

It's interesting that the effect is network-wide, which actually comes as a bit of a relief to me. Anyway, let me share some musings -- please don't read too much into it, it's just something I've thought about for a while.

Basically, I wonder whether this is somehow kind of natural. Most Internet communities, even social media platforms, seem to have a finite lifespan. In the case of SE it seems to be like this -- at the start you have a core of dedicated users who invested time into developing the platform, and the userbase slowly grows as people think "oh this is quite cool". But as the original core users leave the activity slowly dwindles and people move on to the next thing. The Monica situation may well have been a catalyst too.

Maybe it's just confirmation bias on my part, but I think a lot of the original 'power users' of Chemistry.SE have left, or are just inactive. I hesitate to include myself in this category (because I joined quite late), but the graph of my involvement definitely peaked a few years ago.

By the way, this not only includes main site, but also meta. I think that meta has been stale for a while now and I'm personally not sure if that's really a healthy sign of a community (hence my submission for the election questionnaire).*

Take this with a giant pinch of salt, though. It's possible that I'm just projecting my own experiences onto others......


* I hope that me saying this won't influence any nominees' answers. My opinion is not really worth much.

$\endgroup$
6
  • $\begingroup$ I wouldn't say that the effect is network-wide, because Chem.SE dropped by more than 50% from 2019 to 2022 and none of the other sites in my table dropped by nearly as much (Biology comes closes, but that's just 1 site out of the 8 examples, and the activity during this time period dropped by less than 50%). Apart from the "Monica situation" which I did mention in the question already, it seems that your answer is that many of the early power users have left. If that's the only other reason, why does it seem not to have happened as much on other sites? And also they answer much more than ask! $\endgroup$ Aug 7 at 23:53
  • 1
    $\begingroup$ @user1271772 That's a fair point. You know, perhaps I'm just nostalgic for 'good old times', especially because I'm quitting; but basically I think the community building aspect on Chem has not been great. Anyway, I really don't mean to be definitive; it's just a feeling I get, which can't be substantiated with actual evidence, and I don't really spend enough time on other sites to be comfortable making direct comparisons especially on matters like 'site culture' which require one to be fairly involved -- I can only say what we did or didn't do here. I wish people didn't upvote it so much. $\endgroup$ Aug 8 at 0:03
  • $\begingroup$ @user1271772 I do want to ask, though, whether you have some theory of your own? I get the sense that you have some kind of idea, maybe not a definitive one either, but I'd still be interested to hear it. $\endgroup$ Aug 8 at 0:04
  • $\begingroup$ I mentioned the "Monica situation" in my question here, which is one "theory", but it wasn't satisfactory to me on its own because it doesn't explain why Chem declined in activity so much more (at least as a percentage of 2019 usage) compared to Physics, Math, Statistics, etc., and it also doesn't hold all water when we look for example at 2020 (which was after the "Monica situation" and yet activity increased on many of the sites in my table, likely because of the pandemic). Why do you "get a sense" that I have theory of my own, other than the one theory mentioned in the question? $\endgroup$ Aug 8 at 0:12
  • $\begingroup$ @user1271772 Because your comments show you've definitely thought a bit about it! Perhaps even more than me, to be honest. I don't mean that in a snarky way at all, btw. $\endgroup$ Aug 8 at 0:13
  • 1
    $\begingroup$ Haha, yea I've thought about it, but I haven't managed to come up with a satisfactory theory. I did also have a second theory which had to do with the launch of MMSE, but with only 2-4 questions/day over the last 2 years, and many of those questions being stolen not from Chem.SE but from Physics.SE, QuantumComputing.SE, ComputationalScience.SE, ArtificialIntelligence.SE, Mathematics.SE, StackOverflow, etc., it can only account for a small fraction of Chem.SE's drop from 23 questions/day in August 2019 to 11 questions/day in August 2022. $\endgroup$ Aug 8 at 0:37
10
$\begingroup$

Here is the graph from the referenced 2019 post, updated to the currently available data:

enter image description here

From my own teaching experience, I learned there are now commercial sites dedicated to cheating. It is possible we are getting fewer attempts at getting answers to homework or exams on StackExchange Chemistry. This would not be a bad thing.

I do think there is a saturation of questions you would ask for general chemistry on our site. This is one goal of the site - answer all questions that folks might have. In fields like computer science, there are new languages and even new disciplines (data science, artificial intelligence). In introductory chemistry, there is less change, so there is less need to ask new questions instead of just looking up old questions.

Here is a list of the current four most-viewed questions:

  1. What are the maximum number of electrons in each shell? (1,166,616 views)
  2. How do I figure out the hybridization of a particular atom in a molecule? (788,635 views)
  3. Positive or Negative Anode/Cathode in Electrolytic/Galvanic Cell (572,597 views)
  4. Why does ice water get colder when salt is added? (491,198 views)

(They are mostly physical chemistry because this lends itself to general questions rather than asking about a specific substance or reaction in - say - organic chemistry.)

More than 2.5 million people figured out how to get an answer to their question without posting another version of these questions.

When I google "maximum number of electrons in each shell", I get an answer from the BBC first, and then answers from eight different Q&A sites (including a link to the StackExchange post). So there are a lot of choices that potential question posers get exposed to.

From the graph, you can also see that the numbers have large short-term variation. If you just take August numbers (or worse, near-August), you are getting a lot of noise. However, the overall conclusion that the number of questions has declined since a peak sometime around 2017 seems correct.

$\endgroup$
3
  • 3
    $\begingroup$ Yes, the number of questions is not necessarily an accurate indicator of usage. Also, your answer points to a good place to put emphasis in order to improve the site: curation, (editing posts) or posting new answers to old questions. Mod Melanie Shebel (unfortunately is not very active right now) has in the past emphasized this point. $\endgroup$
    – Buck Thorn Mod
    Aug 7 at 9:12
  • $\begingroup$ It seems that you propose two theories for why the usage has dropped: (1) cheating websites have risen in popularity, which makes asking homework questions on Chem.SE less necessary for cheaters, (2) enough questions have already been answered here, so askers don't have to post new questions as much. As for (1), I don't see such a big drop (in terms of percentage of 2019 questions/day) at Physics.SE or Mathematics.SE or Stats.SE (for example). As for (2), again I don't see why this is more true on Chem than other sites, but also $\endgroup$ Aug 7 at 23:16
  • $\begingroup$ [cot'd] I don't see why activity was so steady in 2015, 2016, 2017, 2018, 2019, 2020, then dropped off so much in 2021 and 2022. About using August numbers, it seems from your graph that peaks were always around March-April and troughs have always been in August, so it seems fair to focus on August. About CS.SE, I agree that there's new fields like data science and AI (those are the two that you mentioned), but they both have their own SE sites now, which might even take away from the CS.SE activity. OTOH, those fields lend themselves to an increase in chemistry research-level questions. $\endgroup$ Aug 7 at 23:21
8
$\begingroup$

We have some very knowledgeable regular ("core") participants, but their focus is mostly on answering (and it's great to see the great answers they post btw),as well as editing, commenting etc. There are gems among these questions and answers, but questions from "core" users are less frequent.

The core users referred to in orthocresol's answer were not responsible for the bulk of questions posted during the peak, unless that base or individual participation expanded significantly up to 2016. Much of the waning participation since the peak of 2016 reflects changes in the number of questions, probably due to changes in usage across the internet as a whole and specifically the rise of other platforms competing for attention. Why users might choose other sites, and whether that is sometimes or often a good thing or not remains to answer. Given the goals around which this site is designed, that might not be a bad thing. We might want to focus on quality rather than quantity.

Reasons that might reduce or limit participation include:

  • other sites have a lower barrier to entry (or retention, or to obtaining answers)
  • other sites cater more specifically to users lost since 2016
  • Chemistry SE has a reputation that attracts a particular type of user
  • SE sites attract people already on the network
  • people stick to platforms they become familiar with
  • the culture of the site has evolved
  • new users are discouraged from further participation
  • the bulk of the low-lying fruit (the "easy" questions) has been answered, and more people visiting the site successfully find answers among existing posts
  • difficult questions are difficult to answer, and are more likely to go unanswered, discouraging posting

This is not a fully digested answer, I also recommend adding a little salt. On the plus side, some of the statements can be tested more rigorously using database queries, and it would be cool to see more such information reported.

$\endgroup$
6
  • 4
    $\begingroup$ I would add one more: formatting. Some of us find formatting, and all the details involved with that, to be a chore. For new users, it might be a deal breaker. $\endgroup$
    – Ed V
    Aug 6 at 16:51
  • 1
    $\begingroup$ @EdV I don't know whether requesting that users observe certain rules of usage is very important for participation. I can agree that all else being equal it will reduce participation somewhat (unless the loss of people discouraged by this requirement is compensated by fans of latex, markdown and the like - highly unlikely). One concept that appeals to me though is that of "proof of work" - that in exchange for use of the site you should help maintain it. The requirement helps discourage lazy questions. $\endgroup$
    – Buck Thorn Mod
    Aug 7 at 8:35
  • $\begingroup$ I agree, especially with the “proof of work” concept! $\endgroup$
    – Ed V
    Aug 7 at 13:09
  • 1
    $\begingroup$ I think this is the best answer so far. orthocresol blames it on "core" users leaving, but I agree 100% with you that those users were answering more than asking, so questions/day wasn't affected so much directly by them leaving (although there could be the indirect effect that when questions don't get answered, people are less motivated to recommend the site to others, and they'll also be less likely to ask questions again, but if this were the reason I think one should say "fewer questions are getting answered" rather than "core users left"). But your central theory seems to be that... $\endgroup$ Aug 7 at 23:49
  • $\begingroup$ [cot'd] usage has declined here because of "changes in usage across the internet as a whole and specifically the rise of other platforms competing for attention", and that theory doesn't satisfy me because Chem.SE usage dropped by more than 50% of the 2019 value in my table, and none of the other sites in the table fell in usage by nearly as much in that time period (biology coming the closest, but still less than a 50% decline). $\endgroup$ Aug 7 at 23:52
  • $\begingroup$ @EdV For frequent writers and editors, installing a browser extension might make it a bit less tedious, see chemistry.meta.stackexchange.com/questions/5218/… and vimeo.com/741609155 $\endgroup$
    – Karsten Mod
    Aug 21 at 15:06
3
$\begingroup$

I can't speak for everyone, but I will conjecture it is, in part, the culture here. It went from a place to ask and answer questions freely to a place to be scrutinized at every angle. Members of my Facebook group no longer like to participate here due to the negativity on this SE site.

I can't name anyone in particular because I'm awful at names, but it seems like we're missing people who were here years ago that made the site worth using.

We used to mark questions as homework questions and shut them down when they showed no effort at all. When some effort was shown, we would walk them through it kindly or close it for being a duplicate. It now appears that we close them for being vague even when it's clear what the user is asking.

"Why did everyone downvote my post" is a type of question that regularly appears in meta, but that's not really a new occurrence.

Not accounting for a drop in use over the years, but more recent development is that Chem.SE usage has tanked. We used to receive a rough average of 30k views a day but now receive less than 10k. (I'm just eyeballing this.) Chem.SE traffic

This is a significant drop. I can only conjecture that this is due to Google search viewership as other sites have tanked due to a change in the Google search algorithm that roughly corresponds with the date our stats dropped. My goal as a moderator has always been to improve searchability by making improvements to questions as misspelled or poorly written questions and answers can actually harm a site in search results. I believe there needs to be an increased focus on making the content our site offers as nice as possible.

TL;DR: Polish the turds. Don't just write beautiful answers to questions. Make the questions look like they were worth your time answering.

$\endgroup$
8
  • 2
    $\begingroup$ That drop in traffic is precisely because of the change on Google Analytics, so including that part seems to be distracting. But I totally agree with other things you wrote :) $\endgroup$ Aug 10 at 21:27
  • 4
    $\begingroup$ I completely agree with the negativity. Even your recent question or reactive vs. corrosive has downvotes. This is completely reasonable useful query! And these downvoters have nothing useful to contribute except their negativity. $\endgroup$
    – AChem
    Aug 10 at 22:29
  • $\begingroup$ @user1271772 It’s wholly relevant because site usage has decreased across multiple algorithm updates and IS a cause of decreased traffic. Must you be contrary with everything? $\endgroup$ Aug 11 at 5:24
  • $\begingroup$ @MelanieShebel LOL. Traffic did NOT decrease on Chem.SE like your screenshot suggests. On that specific day, Google Analytics changed to stop counting any views from people who have not accepted cookies. I would highly recommend removing every part related to that screenshot, however I agree with other aspects of this answer :) $\endgroup$ Aug 11 at 9:56
  • 1
    $\begingroup$ To clarify @user1271772's point: meta.stackexchange.com/questions/379947/… $\endgroup$ Aug 11 at 11:37
  • 1
    $\begingroup$ @orthocresol thanks. Here's another one: meta.stackexchange.com/q/379047/391772 $\endgroup$ Aug 11 at 11:50
  • 1
    $\begingroup$ @AChem and everyone: Could we discuss down-voting in chat? I chose a chat rather than a meta-post because rather than having answers, it seems we would benefit from a discussion. chat.stackexchange.com/rooms/138453/downvoting $\endgroup$
    – Karsten Mod
    Aug 11 at 15:49
  • $\begingroup$ @KarstenTheis, Good idea. $\endgroup$
    – AChem
    Aug 11 at 16:04
-2
$\begingroup$

The policy by moderators and the allowed behaviour killed the site.

I have been active on pretty much all the high stem SEs: Mathematics, Physics, and Chemistry. I'd say roughly speaking, mathematics is the most lenient, physics is the middle and chemistry is the least lenient. I mean, of course, the type of questions considerably acceptable in MathSE is entirely different from that of PhysicsSE or Chem SE but besides the point.

For understanding, have a look at my MathSE question page vs my PhysicsSE one, you can see that PhysicsSE the ones which are most upvoted is not particularly due to my presentation of the idea but rather the idea that they got voted up. Meanwhile, on MathSE, it can be seen that generally seen that the presentation of the idea also matters quite a bit.

Of course, it may be that the different types of questions the amount people caring about quality of post vs quality of actual content is different, but I asked most of my questions when I was in HS, so I think it would be a good representative of the experience of a person at that level.

Now, have a look at the questions which I got upvoted on ChemSE, it seems to me the questions that got upvoted are of type "help with technical calculation on specific step". I believe these questions were upvoted mostly by people studying for JEE exam (an Engineering exam in India) and any negative reception that they got was from everywhere else. For instance, this question got closed after initially posting and it was reopened after much effort from me.

And it's usually so that the most upvoted questions here are of this type. The issue is that a broad audience doesn't really care about these types of problems much. For instance, compare the views of the type of question I ask on PhysSE vs ChemSE:

enter image description here

enter image description here

Of course, there are about 900k Indians writing this exam per year, so it looks like a large population of users can be derived from India but at the same time, only 10% of Indians speak English/ 65%-ish have internet, so tapping into the full potential of this pool within the SE framework is not possible.

Here is an example of me trying to ask a type of question that may get popular on PhySE here. Initially it was downvoted to death and I think closed few times before finally opened.


Tl;dr: For gaining users, need to allow questions that actually attract the attention of people to thrive. For keeping users, it boils down to what type of behavior is allowed on the site.

The selfishness of a few users who wanted to rule a personal SE site for themself made this site inhabitable for anyone else.

:/


How are the other SEs faring?

MSE and PhysicsSE are suffering a similar decline but for very different reasons. Their decline is because all the new questions which are being asked are found to be dupes of old ones. Since PhysicsSE policy was so strict before, their first line of defense was to make the policy less strict and be more inclusive of the type of questions that can appear.

MSE I think so alleviated the problem by explicitly allowing people to post their own solutions for problems and then ask the community for feedback. And, I don't think this really does any harm because most of the important and general questions have been answered.

$\endgroup$
8
  • 6
    $\begingroup$ I'm very happy to take criticism, but your unhappiness is misdirected. Policy is not dictated from above by moderators, but rather the community. The moderators are responsible for enforcing site policy. If you think Chemistry is too strict with downvoting, or close voting, that's one thing, and I don't have any issue with you saying that. However, I completely reject the claim that this stems from 'the selfishness of a few users who wanted to rule a personal SE site for [themselves]' -- especially when it is heavily implied that these 'selfish users' are the mods. $\endgroup$ Aug 16 at 9:35
  • 1
    $\begingroup$ Just so you know, had over 1.5k rep here , in total gained about 9k rep in physics and over 12k in MathSE. Ultimately this is just my reading, take it as you like. $\endgroup$ Aug 16 at 9:55
  • 3
    $\begingroup$ 880 posts on MathSE, 500 posts on Phy and 50 posts on Chem.. that's not the best way to get info on activity.. some SEs are bigger than others $\endgroup$ Aug 16 at 12:50
  • $\begingroup$ What is meant by "...it may be that the different types of questions the amount people caring about quality of post vs quality of actual content is different..." I was trying to clean up your post because I had a hard time reading it but I couldn't figure out what was meant here. $\endgroup$ Aug 17 at 21:56
  • $\begingroup$ @MelanieShebel a post can be poor quality but with high quality content. If the question is very valuable, but the presentation, grammar, formatting, etc. is bad, then it's a poor quality post. $\endgroup$ Aug 18 at 2:48
  • 1
    $\begingroup$ If a post is presented beautifully ( well researched , formatting etc), then I would say the quality as a post itself is high but if there is a nice idea in the post then even if its a weak presentation that'd also be a quality post but in a different sense $\endgroup$ Aug 18 at 8:01
  • $\begingroup$ @user1271772 I wasn’t asking you. I was asking OP intent. $\endgroup$ Aug 19 at 5:51
  • 1
    $\begingroup$ Hey, that could have sounded a bit rough. I think @user1271772 had only said what they said because they felt strongly they understood what I said ( which I believe they did!) $\endgroup$ Aug 19 at 5:55

You must log in to answer this question.

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