Addressing data collection problems in web-mediated surveys

Web-based approaches are increasingly being used for carrying out surveys, for example in research or to obtain user feedback in product and systems development. However, the drawbacks of web surveying are often overlooked. Errors in web surveys can be related to sampling, coverage, measurement, and non-response issues. Low response rates and non-response bias are particularly important for web-based surveys. This paper reports on a web-based survey in an international engineering consultancy, aimed at eliciting feedback on the development of systems to support sustainable engineering, that produced a low response rate. To investigate the reasons for this, a follow-up survey was conducted by telephone. The majority of those questioned were unaware of the original survey. The telephone survey showed that reasons for non-completion by those who were aware may be categorized as resources issues, relevance, and fatigue. Differences between those who were aware of the original survey and those who were not are explored and a gap is found between action and intention, i.e. good intentions to complete a survey are very unlikely to translate into action and completed surveys. The paper concludes with practical guidance for administering web-based surveys and observations on the merits of telephone surveys.