Comments on: Analyzing the final and intermediate results of the iversity MOOC Fellowship online voting

By: simonjames

simonjames — Tue, 25 Jun 2013 09:31:25 +0000

Nice post. One hour subsequent the world wide web science mooc had 428 votes and was recorded on the 7th result sheet. There are some methods going on and they are hard to understand if one does not have get get access to to to the full log documents.

By: Jürgen Handke

Jürgen Handke — Thu, 30 May 2013 02:29:39 +0000

We made similar observations during the MOOC voting period. We are the Linguistic Engineering Team of Marburg University, are currently running an extremely successful MOOC on our platform http://www.linguistics-online.com (420 participants from 58 nations) and are taking part in the MOOC-competition with “Teaching English as a Foreign Language – Part I: Pronunciation”.
On 17 May , we realized that voting with fake e-mail registrations was possible. We immediately mailed this to iVersity without getting any response. Here is our mail to iVersity:
“Ich habe meine Zweifel, ob die Abstimmung wirklich „echt“ ist: Man kann schließlich mit „gefakten“ E-Mail-Adressen abstimmen, da offenbar keine Verifizierung durchgeführt wird. Wir haben das einmal versucht (und uns 1 Stimme gegeben). Theoretisch könnte ich so meine Hilfskräfte hinsetzen und sie tagelang mit Fantasie-E-Mail Adressen für uns abstimmen lassen. Damit ist die Abstimmung, so wie sie zur Zeit läuft, nicht gerade sehr aussagekräftig. Die 1 Stimme sollten Sie bei unserem MOOC-Vorschlag übrigens wieder abziehen, die falsche E-Mail-Adresse ist oem@oem.com.”
And on 24 May (after the voting phase), we mentioned the YouTube/voting discrepancy to iVersity:
“Viele Projekte haben weit weniger Video-Views bezogen auf das Application Video zu verzeichnen als Votes in der Abstimmung. Was heißt das? Aus meiner Sicht ganz klar: Es wurden massiv “blind votes” vergeben, ohne sich das Projekt überhaupt anzuschauen. Für die zukünftige Nutzung eines MOOC durch iVersity heißt das, dass Sie sich sehr genau anschauen sollten, wie viele Male das Video aufgerufen wurde und darüber hinaus, wie viele LIKES und DISLIKES es dabei gab. Hier ein Beispiel: Das Projekt Internationaler Masterkurs Agrarmanagement hat ca. 7.000 Votes aber nur 1800 YouTube Views, davon ganze 7 LIKES und sogar 2 DISLIKES. Und nebenbei bemerkt – ein Masterkurs entspricht nicht dem Profil eines MOOCs. Dass es allerdings auch Gegenbeispiele gibt, wissen Sie selbst. Tausende von Video Views mit vielen LIKES und keinen DISLIKES, dazu noch mit aufschlussreichen Kommentaren, die zeigen, wo die wahren Interessen liegen.”
… and of course we did not receive any reply.
In addition to this, I don’t think that an open vote system is not relevant fat all. It does not show the quality of an application but tells us how many supporters the applicants could activate through their networks. Whether they will eventually take part in the MOOC is a completely different question. And then, some applications had the advantage of being mentioned in the media (spiegel online) whereas others weren’t ~ no equal chances for all.
By the way, our own proposal got 591 votes but the video was viewed 2.500 times with a record of 65 LIKES no DISLIKE and 22 overwhelmingly positive comments.

By: René Pickhardt

René Pickhardt — Mon, 27 May 2013 10:37:18 +0000

interestingly after I contacted iversity I did not receive any reply. but the two courses ranked first again in the top courses. Especially they now have less votes than they did before (according to my crawled results)
Also the current result list has a lot of courses with less scores than before.
Especially after closing the votes there must have been re rankings by iversity! I kept my crawler turned on and on May 25th 4:15 the web science mooc had 436 votes and we where listed on the 8th result page.
One hour later the web science mooc had 428 votes and was listed on the 7th result page.
So for some reason we lost 8 votes but gained several rankings (probably because others also lost votes)
All of this is happening without any transparency and without any information being propagated.

By: René Pickhardt

René Pickhardt — Sun, 26 May 2013 18:00:50 +0000

Well that is what I am trying to say. There are some processes going on and they are hard to understand if one does not have access to the full log files. The user interface of the website which I crawled changed frequently displaying different data. That is the reason why the number of votes can only be extracted from my crawl for a view dates.

By: pat

pat — Sun, 26 May 2013 13:18:15 +0000

actually they removed votes from basically everybody, compared to your list…

By: pat

pat — Sun, 26 May 2013 13:13:58 +0000

yes, it would be interesting to see all the data.
Anyway iversity changed the displayed ranking, in many cases (e.g. number 2 or 12 in your ranking list) taking away lots of votes for some reasons and in some instances changing radically their position in the list (e.g. 22 or 23)

By: René Pickhardt

René Pickhardt — Sun, 26 May 2013 11:29:02 +0000

Yes that is true but the problem is that while crawling the data I decided against crawling Youtube data so taking Youtube into the full analysis will not make sense here.
Anyway my crawled data set is only a small snapshot of what has really happened. I have send an email to Iversity asking for logfiles which would help a lot but I didn’t get any response yet.

By: pat

pat — Sat, 25 May 2013 15:36:08 +0000

concerning youtube views, you should consider that some videos have a “no public” view setting