Unfortunately, Google does not reveal how many sites in Indonesia are members of this network. But for comparison, there are around 52 thousand sites in Indonesia that “rent” space for ads from Google through AdSense, according to
builtwidth.com. Usually sites looking for additional income starting from AdSense, and then some will also enter Google Opinion Rewards for Publishers.
Google Surveys distributes questionnaires to these sites, and randomly selected “selected” visitors will be asked to fill out a questionnaire. The method of “intercepting” site visitors is called web intercept, and is essentially the same as the method of selecting respondents by intercepting people in public places, such as mall intercepts. In 2018, the response rate from Google Surveys on online sites was 25% (meaning, 1 in 4 site visitors who were presented with a questionnaire eventually wanted to complete it).
In other words, essentially the sampling method from Google Surveys is convenience sampling, not random sampling. If the sampling method is convenience-based, how can we know the data collected can represent the population? First, let’s recall that the population of Google Surveys in Indonesia and in any country is internet users, not the national population. According to the Digital 2022 Global Overview Report compiled by We Are Social and Hootsuite, the number of internet users in Indonesia is 74% of the population or ± 202 million people. Second, to reduce bias, Google Surveys will automatically weight the survey results. For example, from the survey sample, the percentages of women and men are 52% and 48%, while in the population of internet users in Indonesia, the proportions are 50% and 50%, respectively. Google Surveys will assign a weight of 52 / 50 = 1.04 to all female respondents, and 48 / 50 = 0.96 to all male respondents. The same weighting will also be done for age.
If the sampling method of Google Surveys is convenience sampling, then how is it different from distributing survey questionnaires to an organization’s followers on social media? Yes, distributing questionnaires to followers is technically a convenience sampling method. However, this method contains a greater selection bias, because for someone who can become a social media follower of an organization, he or she may already be much more interested and understand the issues discussed by the organization. This selection bias will be significantly reduced in Google Surveys.
Three easy steps to create Google SurveysCreating a survey using Google Surveys is not difficult, it only takes three steps. First, write questions (we assume the reader already knows how to formulate the questions in the questionnaire according to social research methodology); secondly selecting the target respondents; then the third to confirm and payment.
Step 1: Write the questionsIf you have ever created a questionnaire in Google Forms, then you can certainly fill out questions in Google Surveys. However, it should be noted that the maximum number of questions in Google Survey is 10 items. If your survey requires screening questions, then you can only enter a maximum of 4 questions, and you can only add 6 more questions.
What does it mean? Some surveys require screening questions to select respondents to fill out the questionnaire, usually based on demographic profiles (gender, age, region), behaviour, interests, or attitudes. In screening questions, respondents who choose the appropriate answer can continue to fill out the questionnaire (screen in), while respondents who choose an inappropriate answer will stop filling out the questionnaire (screen out). For example, the survey was conducted targeting respondents with an age range of 19-39 years. People who chose answers in that age range could continue to fill out the questionnaire, while others could not. If further granular age data is required, demographic (not screening) questions about age can still be made.
Figure 1: Writing Questions