Mgoblog Survey Results

Submitted by wingedsig on April 10th, 2011 at 6:07 PM

A few weeks ago I posted a survey for all Mgoblog users to take. First of all, THANK YOU to the 500 participants! You have immensely helped me out for a class, and hopefully you have helped to make Mgoblog a better online community. As many of you requested in the survey thread, below are the statistical results from the survey and a few insights from the essay responses.

Untitled Image Untitled Image 1 Untitled Image 2 Untitled Image 3 Untitled Image 4 Untitled Image 5 Untitled Image 6 Untitled Image 7 Untitled Image 8

Insights (based on feedback received from survey questions and essay responses):

The point system does not affect motivation to post on a large scale, but it does largely affect the content being posted. While many Mgoblog members appreciate the value of a point system, they believe that the members and moderators are incorrectly using it because of opinion-based voting (rather than content-based), inflated reputations due to over-posting (quantity, not quality) and the exploitation of such reputations. Based on this research, I would make the following recommendations to improve the Mgoblog online community:

  • Stress quality (not opinion) based voting to the members. Make it clear to all members that “Upvoting” and “Downvoting” is for quality-based feedback only. All members should have the right to their opinions, and their value and reputation as a poster should not be jeopardized because of varying opinions.
  • Continue to support the voting function, but eliminate points for posting. As a result, quality of posts will be stressed, rather than the quantity of posts.
  • Set an automatic ban number for members that obtain a certain amount of negative points. This will guarantee that members of poor value and reputation will be excluded as soon as possible.
  • If/when these changes (or similar) are enacted, return all posters’ points to 0. This should be done so true reputations can be built within the better system of moderation.

The last "fun" question was meant to judge the Mgoblog community's confidence of win percentage for the upcoming season.

Untitled Image 9

The Mgoblog community expects Michigan to lose to Iowa (close one), Nebraska, and Ohio State, finishing with a record of 9-3. Our most likely loss, according to the amount of votes, is Nebraska.

Thanks again to all participants, and GO BLUE!

Comments

justingoblue

April 10th, 2011 at 6:41 PM ^

Very interesting read. The demographics are about what I expected (although 4% females is really, really weak), it'd be interesting to see what most people studied, as this place seems really skewed towards the math side. I always wondered what that representation was.

I like the auto-banhammer for certain point totals idea. What did Brian say about the figures?

justingoblue

April 12th, 2011 at 7:03 PM ^

I don't know, I only took it once. There's also michichick right? And apparently Shirtless every once in a while. I'd be interested in who all the girls are on here, and if they're posters or lurkers. I'd imagine more lurkers than posters, just because the way conversations are held sometimes seems completely male in perspective (which, duh).

I don't know how well I would do posting on a site of almost all women, I assume at least some women feel that way in reverse around here.

Moe Greene

April 10th, 2011 at 6:55 PM ^

20 women respondents?

For the coolest blog in the universe, this is a total sausage hang.

 

(Not indistinct from far too many parties I was at at the alma mater, he said sadly)

Cope

April 10th, 2011 at 8:04 PM ^

Brian's market is mostly male 20 something's. It would be nice to hear more from the older population of M fans as far as opinions and perspectives for the sake of discussion.

SysMark

April 10th, 2011 at 8:11 PM ^

Impossible to say.  I don't even know how many total members there are.  One thing I would guess is that the older ones may be on far less often...a statistic probably just as important as member count.  Besides being an alum I have a daughter who is a current UM student.  If she were not there I would probably pay less attention.

4godkingandwol…

April 10th, 2011 at 10:00 PM ^

... so I believe they provide an estimate for him that is not survey based (it's cookie based).  Would be interesting to see those stats and see how they compare to the sample to see if there is some sort of selection bias.

GoBlueInNYC

April 11th, 2011 at 2:08 AM ^

I think it's funny that the board-in-aggregate thinks Minnesota has just as much of a chance of beating Michigan as EMU.

Odd question for the EMU alum, any idea if anyone still makes apparel with the old EMU Hurons logo?

Interesting side note, apparently back when EMU was called "Michigan State Normal College" their nickname was (for a time) the "Normalites," which strikes me as very awesome.

DanRareEgg

April 11th, 2011 at 6:14 PM ^

Unfortunately I've never seen any Hurons gear in my trips to the bookstores.  The alumni association used to have "Once a Huron, Always a Huron" bumper stickers for the alums who were students in those days, but I don't even see those anymore.

Also, I would definitely sport a "Normalites" hoodie.

Sgt. Wolverine

April 10th, 2011 at 7:57 PM ^

and I'm on board with removing the point for posting.  It does serve to devalue points.

But...wait, you want to take all my points away?!  NOOOOO THEY BE STEALIN' MAH BUCKET OF POINTS!

MGoSoftball

April 10th, 2011 at 7:57 PM ^

some of the data shows normally distributed curves as well.  Very nice.

The biggest surprise is the curve is skewed toward the younger population.  I would like to know of that 21-30 crowd, are most of you students/graduates?????

MGoSoftball

April 10th, 2011 at 8:15 PM ^

Eastern is a good school too though.  I applaud you good sir for not getting discouraged and walking away.  You took the next best thing.  Something very similar happened to me and I was broken.  But I got up and marched forward.  Everything happens for a reason.

DanRareEgg

April 10th, 2011 at 8:25 PM ^

I did enjoy EMU, especially the political science profs (my major was Public Law and Government).  In some ways I'm glad I didn't get accepted to UM.  I likely would not have met my wife or be at my current job, and frankly I didn't deserve to get in.  I had the test scores, but not the grades.  I'm quite sure a more studious applicant than myself got in...unless it was that one girl whose father was on the UM faculty.  She rated a "64" on the Crayola brightness scale.

BlueDragon

April 10th, 2011 at 9:05 PM ^

1/3 of posters don't post.  Interesting.  The 4% of female posters make up for it by being very vocal.  I can think of 4 different female posters who have been fairly active in the last few months.  The key, though, will be what happens in the fall.

BlueDragon

April 10th, 2011 at 9:25 PM ^

No, I haven't installed the Ultra MGoBoard.  I thought it was for mobile devices only.  Did he release a browser version?

Also, tell the GF that we are all tall, statutesque blond he-men with six-packs who love giving gentle back rubs and watching romantic comedies.  That'll get her posting!  Seriously though, what's she got against DRob?  Is she jealous of his beautiful smile?

Edit:  I checked the link and yeah, he's got a Chrome version, I'll install it the next time I'm on my laptop.

justingoblue

April 11th, 2011 at 11:56 AM ^

It's pretty awesome. You can "highlight" some posters, so yours (and some others) show up in bright yellow at the top. You can "hide" some so they show up as a tiny red box that you have to click to see (I should not ever be subjected to "Cowhoke" again, I personally instacaved that guy), and there are a couple more features that I haven't used, like blocking people below certain point thresholds.

There's also a cool "hide user" button below the persons "online/offline" status, for caving someone on sight.

The_Doctor

April 10th, 2011 at 9:30 PM ^

i am a forty-something alum who is on nearly daily, and occasionally posts...yet missed the request for participation completely!

and hard to argue with the overall votes, and individual votes on the game by game results, but for some reason I am skeptical about 9-3.  can't point to where those predicitons will go wrong, and would love to see 9-3 in the Hokester's first year, but just not sure at this point

 

GO BLUE

YankeeGold

April 10th, 2011 at 10:07 PM ^

A minor comment:  I would think that the people that visit the site less frequently are also less likely to respond to your survey.  (i.e.  They are not as familiar with the board or read the email instead.  The population that saw the survey request would be the only population that could respond.)

Thanks for doing the survey!

wingedsig

April 11th, 2011 at 12:13 AM ^

This is very true! While this might pose a problem, I'm not sure what I could've done since it was on the board for about a week before I calculated the results (and got pushed off the front page pretty fast). Having 500 participants alone helps legitimize the results too.

Thanks for all of the great feedback everyone!

beenplumb

April 11th, 2011 at 11:11 AM ^

Cool study, but in order to draw meaningful conclusions, this survey should've been done from a random sample to avoid selection bias. The above insights should be taken with a grain of salt. You can't rely on sample size alone to legitimitize your findings. The central limit theorem doesn't apply to sample bias when you're trying to draw conclusions on such a varied population. At least, not in survey sampling.

wingedsig

April 11th, 2011 at 11:48 AM ^

I'm no statistics major (never took 350 either), so I'm not sure how I could've drawn a random sample from Mgoblog. Could you explain that further for any future surveys I might conduct? Keep in mind I had a time limit for this survey if that comes into play (would've loved to do the survey during football season when there's more traffic from those who do not normally read/post).

beenplumb

April 11th, 2011 at 2:20 PM ^

I think the best way to do it would be dive in head first and have the site moderators approve a system that randomly selects visitors to the site and asks them to complete the survey. This could be done by picking a sample size that you feel confident with using, and then selecting every nth visitor to the site (where n is a randomly selected seed number which can be generated through a number of methods), surveying that many people, and then analyzing the data that way. This is referred to as systematic sampling. There are other ways to do it, but for simplicity's sake, I think it would be the best to use in this instance. It's also the least invasive, because you don't need a list of GoBlog users to go off of, you can just collect information from anyone who comes across the site (and in doing so are collecting information from the population at-large). Of course, this would have to be something approved by Brian and the mods, but I think it'd be hella cool to find out more about. I would definitely be interested in being a part of it.

There is no perfect way to do internet surveys; even offering some sort of incentive can invalidate the survey, and doesn't garauntee you'll get responses. Certainly, though, asking people to participate is the least reliable way to draw information, for the simple reason that you're only getting data from people who want to go out of their way to answer the questions, and ignores the people who are more moderate in their opinions (see: ratemyprofessors.com).

Again, it was a cool idea, and there is some good information to be gleaned from this, but in terms of actually drawing useful conclusions, the only way to be sure you're arriving at statistically significant interpretations is to make sure your sample accurately represents the population, and the best way to do that is through random sampling.