Web-based surveys in educational research

Abstract

Selcuk University School of Foreign Language, Konya, (TURKEY) E-mails: arifbakla@yahoo.com, ahmetcekic42@hotmail.com, onurkoksalmeb@hotmail.com DOI: 10.7813/2075-4124.2013/5-1/B.1 ABSTRACT Survey research is a challenging and time-consuming process, but current technology offers educational researchers ample opportunities to design and administer time-efficient surveys. With the help of ubiquitous web-based technologies, it is now possible to design and administer a full-length survey on an online platform at no cost at all. The use of such open-source software or free web services makes it possible to design web surveys which are virtually impossible for the non-expert to design and publish online through coding oriented web design software. This study aims to get to the basics of efficiently designed and scientifically validated surveys. It basically discusses various significant points related to survey methodology such as web-based survey design, sampling, data analysis, validation and ethical issues. Key words: Survey design/administration, web-based surveys, survey sampling and validation 1. INTRODUCTION Web-based surveys have traditionally been used in market research, but recently they are also seen in social and educational research. However, most educational researchers seem to have fallen behind technological developments in web survey design and administration because they have little or no information about how to construct and administer scientifically valid and reliable surveys by using web technologies. Still worse than that, they have no awareness of such opportunities. A brief overview of electronic survey methodology might help these researchers consider designing and administering electronic surveys to carry out educational research. Web-based surveys are highly effective if they are designed and administered in scientifically proper ways. Though it might be challenging for those who lack computer literacy to construct web-based surveys, collecting data from participants in different geographical locations is relatively easy when compared with traditional surveys and transferring the collected data into data analysis software is highly time efficient and error free. Though there are significant differences between web-based surveys and their traditional counterparts in terms of such issues as delivery, cost, integration of multimedia elements, etc., both traditional and web-based surveys make it necessary to pay attention to item construction, sampling, validation and administration processes. This paper aiming to explore some basic aspects of web-based surveys is organized into several sections beginning with the basis of survey methodology and electronic surveys. Building upon the background information, the researchers discuss some critical issues such as sampling, representativeness, validation and ethical considerations. 2. BACKGROUND TO THE STUDY Although some censuses were carried out in the ancient world by authorities to explore some aspects of economic and military life, more scientific-like attempts were made in the seventeenth and nineteenth centuries by Graunt (1666) and Laplace (1812). Basically calculating average numbers, English merchant John Graunt tried to count the number of people in London based on the data he gained from a specific area. Then, he generalized his findings to the whole population. The problem with his approach was that taking averages may not give the same results across time and space. A similar attempt made by Laplace (1812) was to find out the population of France. He used central limit theorem, but he made a sampling error since he failed to use random sampling; he instead used cluster sampling (cited in Bethelem, 2009, p. 6). At the end of the nineteenth century, a statistician named Anders Kiaer introduced his Representative Method, and later in 1906 Bowley put forward random sampling method, which enabled him to reach normal distribution through randomly selected samples (Bethelem, 2009, pp. 7-8). His idea formed the basis of scientific sampling methods in survey research. A survey is a solution to many of the problems of a census. In other words, it is a shortcut to the data gathered through a census which involves all members of a given population. Unlike censuses, surveys collect information about only a small part of the population to eliminate problems related to high cost and time (Bethelem, 2009). Bethelem also has the following to say about the seemingly magic side of survey research: