- I started with only the overall categories (the colored ones), since it was easier to start by gauging how much a woman's clothes mattered overall, rather than deciding that something as specific as the fit of her clothes, for example, was worth 3 %.
- I broke down the large categories into their various components, and ranked each one individually, ignoring the overall category score (despite the aforementioned difficulty).
- I totaled up the scores of the components in each category and compared that total to the overall score that I had intuitively given the category at the outset. If there was a large discrepancy (more than a few percent), I reconsidered the original overall category score, as well as the scores of the various components, adjusting either or both in order to make them accurate (and match).
- In order to check the result of step three, I then listed all of the sub-categories in descending order of importance, and worked my way down the list, making sure that the hierarchy of importance was accurate. So, for example, if step three had let me to rank hairstyle at 5 %, but this showed up on the ordered list below makeup, which I'd assigned a 7 %, I went back and made adjustments (because I know intuitively that hair is more important than makeup).
- I then iterated steps two through four until everything matched my intuition. The process was actually fairly enlightening, as it made me rethink some of the stereotypes I had assumed about the importance of some things (heels, for example), and reconsider the value of some other things I had thought were more important than they are (hair, for example).
- Finally, I ranked each sub-category for controllability. This was much simpler than ranking the importance of the various aspects of beauty because it is less a matter of opinion, and does not require the consideration of multiple examples in order to reach a conclusion. Some of these rankings are described in the original post.
Feminine Beauty Is Highly Controllable
The post that bears this title contains a table that I have duplicated below, along with some notes about how I generated it.