HELP! How to design a test?

[Ghost of the dead horse]

There are some tests I've desired to design a long time. One of them would be a funny "What Civ IV leader are you" and "Warcraft III personality test". Some of the similar tests have given me lots of fun, so I think it's good to make more of the fun, if I'm able.

So, making a test.

First of all, I have got my preconceived ideas.

The likes that Warcraft III test should have a result "noob", "banker", "rushnoob", etc.

The thing is; if I just design a test with the knowledge I have at the moment, the test is going to reflect my ideas of the subject. Ok, I'm extremely good in WCIII, so admittedly, my opinion is an expert opinion. Still, I'm not content with that at all. Considering the subject of designing a test, my knowledge is nothing.

Just nothing.

I want to make a test that isn't just my opinion on something.

What kind of a process would you suggest for making such test?

I'm thinking that MBTI members who are familiar with some of the following topics might be most helpful.

1) Taking tests
2) Statistical methodology
3) Understanding of what's fun in testing
4) Arranging feedback opportunities
5) How to calibrate one's test results to the population

etc.

so.

Okay, game centered fun tests aren't the most demanding. I've just thought that it would be fun to use some knowledge in maths & surveying to produce the most fun test on the subject that's possible.

 

[KDude]

Personally, the first thing I would do is come up with an idea for the test that had some broad appeal, and wasn't too niche (like Civ IV, although I love that game). List some subjects on many people's minds and decide on one. Something with enough humorous elements to it to attribute archetypes with. I don't know.. something like "Which Child star are you?" Are you a trainwreck like Edward Furlong, party girl like Drew Barrymore, "older than you look" like Gary Coleman ..umm.. I want to fit Danny Bonaduce into the picture somehow, but I'm not sure where he fits.

Actually, that's a horrible idea. Your Civ one is better. Point though is it's good if more people recognize their results. As for working out the "types", I think a good enough model would just sort of flesh itself out.

On a sidenote, it's interesting that you went for playing styles in that WC3 idea. I probably would just refer to the different classes or characters in the game.

 

[nozflubber]

ahh civ4.... that'd be an interesting test. Ive logged way too many hours on that game.....

I always tell Napoleon to fuck off. such a demanding bastard, a true ENTJ!!

 

[nozflubber]

Is it better to rule with -

A) a gentle but firm hand, guiding your people toward prosperity (Cyrus, Ghandi)

B) an iron fist of law and power, demanding and commanding what you will (ghengis Khan, Shaka, Montezuma)

C) neither.

also notice how like the leader characteristics already kind of resemble the supposed "real" typing? Like Churchill has been typed as ESTJ before, and he's one of the Protective leaders. it might be the case a lot of the analysis is already done for you.

 

[Ghost of the dead horse]

Nice ideas there, nozflubber. I'll get back to this.

 

[Ghost of the dead horse]

OK. At this moment I've decided to do a test about Warcraft 3 playing style.

Internally, I'm thinking both in scales (game experience, politeness, concentration on teamwork/soloing, concentration on ladder games/custom games etc.) but also in identities.

I'll do a sum-up.

Scales: measuring people's situation on a continuum and giving them a score of that.
Identities: measuring people's likeness with different templates and giving them that identity that has the best match.

There's good and bad with both approaches. Some combinations of values in different scales might not just make sense. Also, there's certain playstyles that just don't combine with any reasonable combination of scales; So, as a compromise, I would test for some very odd specific playstyles, and use a scale as a backup. That way, I would cover both some extreme peculiarities in playstyles, and the average player, without compromising either.

Hmm.

Sounds like I have a plan.

 

[Ghost of the dead horse]

OK. At this moment I've decided to do a test about Warcraft 3 playing style.

Internally, I'm thinking both in scales (game experience, politeness, concentration on teamwork/soloing, concentration on ladder games/custom games etc.) but also in identities.

I'll do a sum-up.

Scales: measuring people's situation on a continuum and giving them a score of that.
Identities: measuring people's likeness with different templates and giving them that identity that has the best match.

There's good and bad with both approaches. Some combinations of values in different scales might not just make sense. Also, there's certain playstyles that just don't combine with any reasonable combination of scales; So, as a compromise, I would test for some very odd specific playstyles, and use a scale as a backup. That way, I would cover both some extreme peculiarities in playstyles, and the average player, without compromising either.

Hmm.

Sounds like I have a plan.