Five Years: The April Fools' Postmortem ExperimentOver the years I've run a few of April Fools' Day gags on the site. They generally revolved around a silly idea and included some modifications to the presentation of the site. The first one I did, back in April 2005, was 'The Super Milk Chan Experiment' which was a simple redesign of the site pretending that it had switched over to covering news on the extremely bizarre anime that was airing on Adult Swim at the time. The following year I reused the same joke except as 'The 12 Oz. Mouse Experiment' for a different bizarre show.
ARG, though I did throw a few jokes in there for good measure.
The General IdeaDevelopment of the entire thing took me perhaps around a month. I didn't have a super clear idea what I wanted from it, just that it was going to be a series of ridiculous quizes. The original idea was to make it some kind of Guild of Calamatous Intent recruiting site. That idea basically stuck but I slowly moved away from explicitly stating that it was a Guild operation. It went from being Guild-branded to only revealing itself at the end to containing only very vague hints. In retrospect that change may not have been the best since it removed a familiar context from what was happening. If people don't know that they are supposed to be acting one way they they may be more inclined to be honest.
It also implements a rudimentary mail queue system which I did to make it seem more like applicants were being pre-screened. The queue was cleared by me manually for the entire day. I wrote a quick and dirty script that told me the number of emails in the queue. I checked it several times a day and ran the send script whenever I noticed that there was mail waiting to be sent.
Most of my day looked like this.
Reception & ParticipationI wasn't entirely sure what would come of this whole thing since I didn't really run it by anyone prior to launching. Obviously the biggest fear was that nobody would participate and if people did participate then nobody would complete it. The first thing ended up not being true, but the second one sadly did come true.
Just The FactsThere were a total of 72 sign ups from the initial form, however only 62 of those were legimate email addresses and 53 were unique emails (with a few people entering two or three times). So 53 people isn't too shabby for a fairly ominous looking form put up on April Fools' Day.
However, here comes the sad stats. Whenever you use the sign up form an account is created for you, with the username and password (both randomly generated) emailed to you later on. Of the 72 accounts created, 33 of them never got beyond the initial registration, meaning they never logged in an accepted the terms of service.
Chart denoting the status of all accounts.
The Specifics of the TestsSo what was the trick to the tests? Well, basically it came down to extremes and some basic trickery. The first test was a simple scored self-evaluation. For the maximum score every question had to be answered with either a 1 or a 5. The more villainous you are (or at least answering from a villian's perspective) the higher your score will be.
A score of 30 or higher was required to pass the test with a 100 being the highest possible score (five points per question). My choice of 30 seemed to be just about right since the average score was 27.84.
Here's a breakdown of how people answered coupled with the expected correct answers:
Light green is the highest possible scoring answer (always 1 or 5), dark is the average answer given.
Here's a breakdown purely for each question and how many times it was answered correctly:
Number of times a question was answered correctly
The third exam, 'Situational Reactions' scored you based on how agressive you were. You passed the test if you scored either very high (for crazy psycho agression) or very low (denoting attention to bureocratic rules). If you answer D to every question then you will pass, likewise if you answer B to everything you would get similar results. Mixing in too many of the other answers would put you into neutral territory and you would fail.
The fourth and final test was meant to show that people should be deceitful and take advantage as much as possible. The site 'pretended' to be broken and then presented you with an opportunity to set your test score. The only requirement for passing was if you set your score above 100 (denoting that you were thinking big).
The data for those two tests isn't really relevant since only two people took the third test and one person took the final test.