Pages

Wednesday, May 23, 2012

Google Surveys are Here: Implications for Multicultural Research

Google is my favorite tech company these days. It has created and continues to create solutions that are quite outstanding.  That brings freedom of research and expression to the world, literally.

Recently the company launched Google Surveys. That allows users to:

  • Create a survey online, including graphics
  • Respondents answer questions to get free access to content online, one or two questions at the time, only
  • The content publishers get paid for allowing their viewers to answer questions
  • The survey creator finally gets the data tabulated and graphically displayed, and the data is updated continuously as more people answer

The system of Google Surveys is still being developed and eventually it will become a killer app, I think.  Right now the survey creator or researcher pays 10 cents per response, or 50 cents per response if the responses are from specific targets.  So, if you ask 50 questions from 500 Hispanics it costs you $2,500.  Not that bad although not that inexpensive either compared with companies like Research Now that can do similar work via River methodology at very competitive prices depending on the specifics of the project and include many more questions in one survey.  There are several issues that Google Surveys needs to address before their platform is really useful for multicultural research.  Here is why:



  • The fact that each respondent answers only one or two questions at a time makes correlations and multivariate statistics impossible to conduct.  If you have a survey with 50 questions, you need all 50 responses from each individual to be able to compute statistical associations and other relational analysis.
  • They are currently only targeting English language publishers, thus if the researcher wants to target “other language” respondents that is difficult as the incidence may be too low and the data slow to collect.
  • The cost is not really lower from that of competitors that do target “other language” groups, and who get individual respondents to actually complete an entire, let’s say, 50 question survey at one time. That way the researcher can conduct actual segmentations and other complex data analyses.
  • Also, the target of Google Surveys now are people who visit sites that charge for content. That makes the respondents somewhat idiosyncratic and perhaps not representative of the universe of online respondents.

At this time "Google Surveys" is a useful tool for people who want to know the answer to one or two very specific questions. In multicultural research, we usually need to find relationships between variables from the diverse cultural groups in the US.  Google Surveys does not seem to be ready to address these issues at this time. Nevertheless I am hopeful that they will eventually have a very general and powerful tool.