The hispanic ethnicity is not generally considered to be tied to a specific race. In various forms I have seen and completed recently, race and hispanic ethnicity are two separate questions. This is more accurate because it does not exclude/ignore, e.g., black hispanics who may live in or descend from Caribbean or Central American nations.
The question about children should have an option “0, and unsure about having some in the future”.
It would help to provide lists of “hard sciences” and “soft sciences” so that people know what they are selecting.
There is a typo in the Liberal answer for the Political question: “moire redistribution of wealth”.
Some people may come from families of mixed religious background. This question should have either a multiple-answer option (more accurate) or specify that responders should choose based on some criteria (vague, open to interpretation).
ACT scores have already been mentioned as an addition to the SAT scores. I think a category for GRE general scores would also be worthwhile; the GRE has higher resolution than the SAT or ACT at the high tail. Going further on these questions, splitting up the scores into the different subject areas (math/verbal) would be nice. Of course, the GRE scoring system has been recently changed, which would necessitate two possible response areas like the SAT question has now. (There are (questionably) accurate conversions between the different scoring systems which could be used for the survey analysis and comparisons.)
Another question which might be interesting: ask responders to take the AQ test. It’s not long, and it provides an inexact but standardized measure which is correlated to Asperger’s/high-functioning autism. Probably better than relying on self-diagnoses, which is common.
In the Less Wrong Use question, there is a typo: “but never a top-level psost”.
The hispanic ethnicity is not generally considered to be tied to a specific race. In various forms I have seen and completed recently, race and hispanic ethnicity are two separate questions.
Correct - to be standard, they should be separate questions. Sometimes they are analyzed together on the back end.
The hispanic ethnicity is not generally considered to be tied to a specific race. In various forms I have seen and completed recently, race and hispanic ethnicity are two separate questions. This is more accurate because it does not exclude/ignore, e.g., black hispanics who may live in or descend from Caribbean or Central American nations.
The question about children should have an option “0, and unsure about having some in the future”.
It would help to provide lists of “hard sciences” and “soft sciences” so that people know what they are selecting.
There is a typo in the Liberal answer for the Political question: “moire redistribution of wealth”.
Some people may come from families of mixed religious background. This question should have either a multiple-answer option (more accurate) or specify that responders should choose based on some criteria (vague, open to interpretation).
For the IQ tests, two which came up in the comments after the last survey were iqtest.dk and sifter.org/iqtest. My scores on both tests were consistent. In a reply to the previously-linked comment, gwern linked his list of online IQ tests.
ACT scores have already been mentioned as an addition to the SAT scores. I think a category for GRE general scores would also be worthwhile; the GRE has higher resolution than the SAT or ACT at the high tail. Going further on these questions, splitting up the scores into the different subject areas (math/verbal) would be nice. Of course, the GRE scoring system has been recently changed, which would necessitate two possible response areas like the SAT question has now. (There are (questionably) accurate conversions between the different scoring systems which could be used for the survey analysis and comparisons.)
Another question which might be interesting: ask responders to take the AQ test. It’s not long, and it provides an inexact but standardized measure which is correlated to Asperger’s/high-functioning autism. Probably better than relying on self-diagnoses, which is common.
In the Less Wrong Use question, there is a typo: “but never a top-level psost”.
Correct - to be standard, they should be separate questions. Sometimes they are analyzed together on the back end.