http://www.threadless.com/ Fri, 01 Jun 2012 17:18:53 -0500 Thu, 29 Jan 2009 20:06:16 -0600
Because the opportunity cost is high, I chose to sample a small percentage of the full catalogue and make my conclusions from the pool of data collected. My search has led me to conclude that there are not, in fact, any t-shirts that I cannot picture on both male and female counterparts of our species. (NB. The Unicorn shirt comes very close to being girly.)

I bet the wheels behind this machine carefully plan the pickings every couple of weeks, and they must know more than just the final average score. After all, the final set of scores has not only a mean, but also a standard deviation.

Example. Scoring ends, and Design X gets an average score of 1.89 (purely hypothetical, and not my Duck Season design, really). If the standard deviation is small, then all genders/ages/races agree that Design X deserves 1.89 as a score. HOWEVER, if the standard deviation is large, we may end up with a bimodal distribution. If all of one gender's scores are separated from the other gender's scores, and separate means are taken, we might find that Gender 1 thinks Design X deserves a 0.59 and that Gender 2 thinks Design X deserves a 3.19. The average of these two scores is the final 1.89, but it is CLEAR that a subset of the population thinks that Duck Season... er... Design X... should be printed.

In conclusion, score my next design (preferably something higher this time), so I won't have to be sad about it afterward.

Thank you.]]> Wed, 28 Jan 2009 21:06:55 -0600
No. such. luck.

We study the past to prevent the same mistake from happening in the future. I feel enlightened after having analyzed the mistakes I made in the last running. ...And yet, I just entered my second design and am already checking it periodically to see if it's up and running.

I should remember that Threadless promises to send me an email when the design is in the competition. So far, they haven't let me down. And so, I go back to work. At least for the next two minutes.]]>