SurveyMonkey Audience is a little more expensive per response, and therefore probably doesn’t make sense at too large a scale, but I’ve found that it’s much easier to set up for a small scale. The cost is $1 per response (up to 4 questions and then $0.25 per additional question) in the United States, and the respondent quality is generally decent. For demographic, education, and other filters, the cost per response goes up based on the nature and number of filters.
SurveyMonkey Audience is, however, available in only two non-US countries (and quite expensive in those).
All in all, I’d say that if you are interested in conducting a quick survey of US-only populations, and don’t need too large a sample, SurveyMonkey Audience is a good way to go. You can also later scale up the same SurveyMonkey survey using MTurk (i.e., you can use MTurk to get more participants on the same survey). And you can use the same survey (but a different collector) to get responses for free from friends or groups of people you know, to better do comparisons.
Issa Rice and I did our Wikipedia usage survey research using SurveyMonkey, and we used SurveyMonkey Audience for general population benchmarks:
Another option worth looking at for single-question surveys is Google Consumer Surveys. For single-question surveys, it’s pretty cheap (10 cents per response). Basically Google shows the survey in an ad-like context. Learn more at https://www.google.com/insights/consumersurveys/home
SurveyMonkey Audience is a little more expensive per response, and therefore probably doesn’t make sense at too large a scale, but I’ve found that it’s much easier to set up for a small scale. The cost is $1 per response (up to 4 questions and then $0.25 per additional question) in the United States, and the respondent quality is generally decent. For demographic, education, and other filters, the cost per response goes up based on the nature and number of filters.
SurveyMonkey Audience is, however, available in only two non-US countries (and quite expensive in those).
All in all, I’d say that if you are interested in conducting a quick survey of US-only populations, and don’t need too large a sample, SurveyMonkey Audience is a good way to go. You can also later scale up the same SurveyMonkey survey using MTurk (i.e., you can use MTurk to get more participants on the same survey). And you can use the same survey (but a different collector) to get responses for free from friends or groups of people you know, to better do comparisons.
Issa Rice and I did our Wikipedia usage survey research using SurveyMonkey, and we used SurveyMonkey Audience for general population benchmarks:
http://lesswrong.com/lw/nru/wikipedia_usage_survey_results/
Another option worth looking at for single-question surveys is Google Consumer Surveys. For single-question surveys, it’s pretty cheap (10 cents per response). Basically Google shows the survey in an ad-like context. Learn more at https://www.google.com/insights/consumersurveys/home