Internetiküsitlus Eesti Rahvusringhäälingu auditooriumiuuringute näitel: metodoloogiline analüüs
Title: Methodological Analysis of Web Surveys: The Example of Estonian Public Broadcasting Audience Surveys. The Internet, and technology in general, are shaping several aspects of our daily lives — including the way surveys are conducted. We can see that different research methods have found their way to the Internet; what is more, the data presented in this thesis indicates that web surveys are widely used. The thesis at hand sets out to explore the methodological point of view of web surveys. Importance of this topic is enhanced by the fact that web surveys are often used by people with little survey experience. A web survey’s sampling can be unforeseeably large, often creating an illusion of very accurate and good quality results. Web surveys offer a large number of options that can be beneficial for those who know how to utilize them. For example, one advantage appears already in the course of questionnaire creation: a surveyor can choose between numerous audiovisual elements and use them to simplify the realization of their survey goal. For example, Estonian Public Broadcasting was able to carry out a survey where the respondents were asked to assess audio clips. On the other hand, it is important to identify the line beyond which the option to add elements might have an unwanted effect on the results. Web survey designs are made up of several important and specific nuances: what is the proper response format, how to avoid multiple submissions, what is the most appropriate color solution etc. The best part of methodological researches that tackle the subject of web surveys investigate sampling creation strategies. The Internet provides a surveyor with a multitude of methods for respondents’ recruitment, both probability-based and non-probability-based. In general, web surveys are not considered to be an appropriate method of research; however, if the surveyor wishes to represent the whole population, a web survey can be part of a mixed-mode survey. This kind of survey was used, for example, in the most recent population and housing census in Estonia. Nonetheless, the target groups’ access to the Internet is imperative to the success of a web survey. In Estonia, ca ¾ of the population uses the Internet, net surfing and browsing is more popular among younger age groups, but the users’ age difference is diminishing. Though the response rate is an important aspect of quantitative surveys, this paper does not give a clear overview of this particular data. Using literature as basis for such assessments is complicated, because researchers define the term differently; neither can one rely on options provided by surveys, because these are not homogenous. What is more, response activity (rate) is affected by numerous other factors, making it difficult to identify a specific method’s influence. Various influencing factors have been summarized in the chapter about response rate. It is simpler to obtain data about response rates of web panel samplings, because recruiters have access to survey invite recipients’ figures. For the most part, web surveys function as cheap and swift means of research; however, people can easily neglect the matter of results’ quality. The chapter about quality looks into this issue more thoroughly — examples from the USA show that providers of web survey panels have been exposed to external pressure, which, in the end, has undermined the quality of survey results. It is important to emphasize that web survey clients (the party ordering a survey) should learn more about the specifics of web surveys and related samplings. Also, survey clients and providers should work hand in hand to uplift web panel participants’ interest and will to partake. Abuse of e-mail addresses that have been added to databases for purposes unrelated to research can cause negative attitude towards web surveys and ultimately damage the practice’s success.