My prior here is that we’re just seeing less visitors overall
Trike should be able to check this. My guess is that the site has a lot higher view rate than 2-3 years ago, from both casual and registered users, but maybe they do not vote as often as before (or are more negative than before?), keeping the average karma/post steady.
Also, does the median karma per post match the mean, or is the latter skewed by high-karma outliers?
I have access to LW google analytics. Traffic on LW has trended down since its peak in 2012, but not as steeply downwards as Discussion posts… perhaps a 15% drop.
My pet theory is the same as the one I’ve always offered: the LW user moderation is too heavy-handed, writing LW posts isn’t that much fun, and there’s a culture of “how dare you write that post” (e.g. “was this really appropriate for (Main|Discussion)? it really should have gone in (Discussion|an Open Thread)” is a common refrain). And there’s become a kind of deflationary phenomenon where what was once appropriate for Main becomes appropriate for Discussion becomes appropriate for an open thread (e.g. this was a featured post in the early days of LW; nowadays a link with explanatory text is frequently an open thread post). I think we should try (a) telling people in threads like these they should write up interesting post ideas if they have them (to save LW!) and (b) go friendly/easy on those who do write posts.
Note that something like this has been discussed as a problem since 2011.
The nice thing about user moderation in the form of voting is that it’s easy to throw a lot of content at the forum and see what sticks… it will get filtered automatically. So why not do that?
After I look at the old Main or Discussion, I mostly remember the best posts and am hesitant to post lower quality stuff. Not sure if this is a common sentiment.
Consider that your internal estimate of the quality of the post you will make is noisy. For example, I did not expect this to become my most popular post, and I did not anticipate all the negative feedback I received on this post (also potentially interesting to note: although that second post ended at +5, I think it was lower than 0 at times and overall I felt pretty punished for writing it).
Consider that your internal estimate of the quality of the post you will make is noisy.
Oh, I agree that my estimate of how my post would be rated is piss-poor, judging by how my comments are rated. I was talking about how I personally feel about their quality vs how I personally feel about the quality of the best posts. Maybe I’ll dig through my drafts and post something after a slight polishing, just to see what happens.
Growth in survey answers, presumably? We’re talking about hits here, which includes the lurking masses. And should be a much more solid number, if we have it.
If you mean growth in the status associated with taking the survey, which is what actually matters for gaining responses, then I’m not sure about that. I haven’t run the regression I’d need to generalize, but my own ritual “I took the survey” responses have gained me less karma each year.
There are other population dynamics that could explain this, but they all look a little far-fetched to me.
It would be great to see some numbers, and which parts of the site they’re viewing. For this particular question, we’d want to look at the visits to the Discussion page, rather than visits to general parts of the site such as the Sequences or Wiki.
Trike should be able to check this. My guess is that the site has a lot higher view rate than 2-3 years ago, from both casual and registered users, but maybe they do not vote as often as before (or are more negative than before?), keeping the average karma/post steady.
Also, does the median karma per post match the mean, or is the latter skewed by high-karma outliers?
I have access to LW google analytics. Traffic on LW has trended down since its peak in 2012, but not as steeply downwards as Discussion posts… perhaps a 15% drop.
My pet theory is the same as the one I’ve always offered: the LW user moderation is too heavy-handed, writing LW posts isn’t that much fun, and there’s a culture of “how dare you write that post” (e.g. “was this really appropriate for (Main|Discussion)? it really should have gone in (Discussion|an Open Thread)” is a common refrain). And there’s become a kind of deflationary phenomenon where what was once appropriate for Main becomes appropriate for Discussion becomes appropriate for an open thread (e.g. this was a featured post in the early days of LW; nowadays a link with explanatory text is frequently an open thread post). I think we should try (a) telling people in threads like these they should write up interesting post ideas if they have them (to save LW!) and (b) go friendly/easy on those who do write posts.
Note that something like this has been discussed as a problem since 2011.
The nice thing about user moderation in the form of voting is that it’s easy to throw a lot of content at the forum and see what sticks… it will get filtered automatically. So why not do that?
After I look at the old Main or Discussion, I mostly remember the best posts and am hesitant to post lower quality stuff. Not sure if this is a common sentiment.
Consider that your internal estimate of the quality of the post you will make is noisy. For example, I did not expect this to become my most popular post, and I did not anticipate all the negative feedback I received on this post (also potentially interesting to note: although that second post ended at +5, I think it was lower than 0 at times and overall I felt pretty punished for writing it).
Oh, I agree that my estimate of how my post would be rated is piss-poor, judging by how my comments are rated. I was talking about how I personally feel about their quality vs how I personally feel about the quality of the best posts. Maybe I’ll dig through my drafts and post something after a slight polishing, just to see what happens.
Sounds excellent!
What do you base your guess of higher readership upon? e.g. there’s been no new HPMOR this year, so no n00bs from that.
Well, the surveys have consistently shown growth, although IIRC the last survey showed less growth.
Growth in survey answers, presumably? We’re talking about hits here, which includes the lurking masses. And should be a much more solid number, if we have it.
And growth in status of the survey.
If you mean growth in the status associated with taking the survey, which is what actually matters for gaining responses, then I’m not sure about that. I haven’t run the regression I’d need to generalize, but my own ritual “I took the survey” responses have gained me less karma each year.
There are other population dynamics that could explain this, but they all look a little far-fetched to me.
It would be great to see some numbers, and which parts of the site they’re viewing. For this particular question, we’d want to look at the visits to the Discussion page, rather than visits to general parts of the site such as the Sequences or Wiki.