Mod Browsing Overhaul Discussion

[showler]

39 minutes ago, Demorphic said:

The window is gone but your recent searches are saved and can be run again.

 

Also, you can always bookmark well-used, custom searches, such as

One thing is that if I open the quick search and type "witcher" then the results appear in the overlay window.  If I type "witcher" and hit enter, the results appear in a "search results" page.  And if I open the quick search and click on my previous search of "witcher" then the results also appear in the "search results" page.  There appears to be no way to stay in the quick search overlay while re-using a previous search without re-typing it and avoiding hitting enter.

[showler]

23 minutes ago, kat1004 said:

The users you're getting feedback from are either casual users who don't go on the site often and don't have adblockers, or people who thought it was worth the potential risk to their computers/privacy.

And I one again must ask:  What about either of those things is a disqualifying factor for randomly selecting people to get their opinion on the website change?

[Glitchfinder]

6 minutes ago, showler said:

And I one again must ask:  What about either of those things is a disqualifying factor for randomly selecting people to get their opinion on the website change?

It's no longer random, for one. The entire point of a random sample is to eliminate sampling bias, so adding a sample bias back in after randomly selecting users will affect the results. The actual impact on the survey can only be guessed at, because there is no control group we can compare against to see what eliminating the more tech savvy users from survey responses would change in the results.

[kat1004]

5 minutes ago, showler said:

And I one again must ask:  What about either of those things is a disqualifying factor for randomly selecting people to get their opinion on the website change?

Harming the users who use the site most frequently because they weren't randomly picked from a hat to risk their computers' safety is... very much not a good look in general. 

I NEVER said anything about those being "disqualifying factors" though, so don't put words in my keyboard. They should run an actual poll though, one that gets feedback from as many users as possible. See how popular the changes actually are.

[Demorphic]

4 minutes ago, Glitchfinder said:

It's no longer random, for one. The entire point of a random sample is to eliminate sampling bias, so adding a sample bias back in after randomly selecting users will affect the results. The actual impact on the survey can only be guessed at, because there is no control group we can compare against to see what eliminating the more tech savvy users from survey responses would change in the results.

As a premium user / mod author you don't see ads on the site, but I understand many users will have popups blocked by default.

[kat1004]

Just now, Demorphic said:

As a premium user / mod author you don't see ads on the site, but I understand many users will have popups blocked by default.

The popup search window is unwieldy and unpleasant to use. Please at least consider an option to switch back to the way Quick Search used to function.

[Demorphic]

24 minutes ago, showler said:

One thing is that if I open the quick search and type "witcher" then the results appear in the overlay window.  If I type "witcher" and hit enter, the results appear in a "search results" page.  And if I open the quick search and click on my previous search of "witcher" then the results also appear in the "search results" page.  There appears to be no way to stay in the quick search overlay while re-using a previous search without re-typing it and avoiding hitting enter.

Yep, this is true, hitting a recent search takes you out of the quick search to the search results page.

I can further refine it by hitting the "view all" on the right hand side. It is a different approach though, you are right.

[showler]

54 minutes ago, Glitchfinder said:

It's no longer random, for one. The entire point of a random sample is to eliminate sampling bias, so adding a sample bias back in after randomly selecting users will affect the results. The actual impact on the survey can only be guessed at, because there is no control group we can compare against to see what eliminating the more tech savvy users from survey responses would change in the results.

All surveys have biases like that.  A phone survey only selects people who have a phone and are willing to answer it.  An in-person survey at a mall only selects people who are at the mall and willing to answer questions.  And a survey that pops up for every person who visits Nexus Mods for a week will only select people who visit during that week and are willing to fill out the survey.

The question is whether the selecting factor invalidates the results.  And there is no evidence that whether or not someone blocks hotjar has any bearing on whether or not they like the new site.  People seem to be objecting to the results of that survey simply because it doesn't agree with their expectations.  And those expectations are largely the result of the complaints they see in this thread, on Discord or on Reddit...despite the clear fact that those results are much more biased in a way that definitely skews the results.

[Glitchfinder]

17 minutes ago, showler said:

The question is whether the selecting factor invalidates the results.

Yes, probably. We can't actually know without better controlled data, but it would be reasonable to assume that power users (e.g. repeat customers) are more likely to be higher frequency internet users and thus more likely to be excluded from the sample by being unwilling to answer a survey on a site that requires them to disable their adblocker.

Edit: Just to be clear about my reasoning here: This is a modding website. Gaming is already a hobby that trends more towards the technologically-inclined, and modding your games trends even harder towards the more technically inclined. There will be a much stronger correlation between "People who frequently mod their games" and "people who use adblockers while browsing the internet" than there will be between "people who regularly browse the internet" and the latter, or "people who frequently use computers" and the latter.

By taking survey data that penalizes that correlation, Nexus is likely skewing the data towards casual visitors who visit the site infrequently, and away from users who regularly use the site and are thus more likely to directly contribute via membership subscriptions. If I were to take an uncharitable perspective, this skew could be interpreted as deliberate due to the nature of advertising and how it helps support the site, but I doubt that is something they would have actively chosen to do.

[kat1004]

2 minutes ago, showler said:

All surveys have biases like that.  A phone survey only selects people who have a phone and are willing to answer it.  An in-person survey at a mall only selects people who are at the mall and willing to answer questions.  And a survey that pops up for every person who visits Nexus Mods for a week will only select people who visit during that week and are willing to fill out the survey.

The question is whether the selecting factor invalidates the results.  And there is no evidence that whether or not someone blocks hotjar has any bearing on whether or not they like the new site.  People seem to be objecting to the results of that survey simply because it doesn't agree with their expectations.  And those expectations are largely the result of the complaints they see in this thread, on Discord or on Reddit...despite the clear fact that those results are much more biased in a way that definitely skews the results.

So then have Nexus run a real poll, one that isn't based on an untrustworthy website that adblockers, not users, have deemed unsafe and blocked. One that everyone, not just a fraction of a percent of users, can answer. Have it locked so only logged in users who have existed for more than... say, a couple days to a couple weeks can fill it out, so there's a guarantee that only real accounts can answer the poll in order to minimize poll fraud (users filing out the poll multiple times.

And display the results publicly so that people don't get upset at the utter lack of real transparency - which is a lot of our issues with the HotJar poll outside of it being an untrustworthy site. There seems to be no transparency.

If a real poll shows similar results, I'll still be upset at the drastic drop in usability, but I'll shut up and take it (and use Stylish... which really shouldn't be necessary to make a site usable for the majority but here we are) and I'm sure a lot of other people would too.