(no subject)

[mattlistener]

Hey! I'm actually keen to do this one.

I am interested for people who know me to go to the following site and pick 5-6 words from a list that you feel best describe me. Thanks!

http://kevan.org/johari?name=mattlistener

I will do the same for others' that I see on LJ.

Comments

[arfur]

it feels hard to do this accurately, because of primacy and recency effects with the labels, and because they're lumped together alphabetically rather than having clusters or antonyms near each other.

hm...

[beowabbit]

I totally agree with “confident” and “trustworthy”, too (which are currently listed under “Façade”); they’re labels I considered but had to prune in order to get down to six.

Still, I had a much easier time settling on six adjectives with you than I have with the couple of other people I did this for.

I agree with bubblebabble’s observation that the UI is pretty bad and likely to introduce artifacts in the data.

[mattlistener.livejournal.com]

I don't know, it seems to me that "similarity clusters" would introduce more bias than they would remove. Arranging alphabetically is meaning-neutral, which is probably what is wanted for best validity.

Random would also be meaning-neutral, but if each respondent got a different arrangement any location biases would favor different words for different users. (I suspect words around the edge get picked less often, on average.) It's better to have an identical arrangement for each user, so that the biases are consistent.