Thanks for your kind words and question, Esteban. Our method for measuring ascribed color differs from standard questionnaire-based approaches. Briefly, there are three main steps:

1. Determine the culturally appropriate categorization for a standardized set of facial portraits.

2. For survey respondents, select which standardized facial portrait most closely resembles each respondent.

3. Use the culturally appropriate categorization of the matched facial portrait as a proxy for how each respondent would be perceived by other Puerto Ricans in terms of color.

The key to this process is the formative ethnographic research about the local model of ethnic classification. For more about this phase of the research, see ref. 75 in our PLoS ONE paper:

Gravlee CC (2005) Ethnic classification in southeastern Puerto Rico: the cultural model of “color”. Soc Forces 83: 949–970.

You can obtain a copy of the paper from my website: http://www.gravlee.org/pu.... The link between ethnography and survey measurement of "color" is cultural consensus analysis of the identification task described in the Social Forces paper. We use the consensus estimates of color for the portrait that most closely resembled survey respondents as a proxy for social classification.

Note that you can download the full set of standardized facial portraits here:

http://www.gravlee.org/re...

Last, I should mention that we also measured self-identified "color," using the categories that emerged from ethnography. But for a variety of reasons, the consensus-based estimate of ascribed color is the key variable of interest.

I'd be glad to answer any additional questions and hope that this approach will be useful for your work in Puerto Rico.

Clarence C. Gravlee