In response to: Cantrell M. & Lupinacci P. (2007) Methodological issues in online data collection. Journal of Advanced Nursing60(5), 544–549.


The Cantrell and Lupinacci (2007) article was the subject of a recent Journal Club discussion at the University of Pittsburgh School of Nursing, with teleconference attendance by faculty members from several partner schools. The journal club participants agreed that a web-based approach was an appropriate methodology for research in the relatively small and widely disbursed population of cancer survivors, although the paper omitted important details that are useful in evaluating the rigour of web-based research (Eysenbach 2004). The group believed that the following issues deserve further discussion.

The authors’ single recruitment method – advertising on websites used by cancer survivors – clearly targeted the desired populations, but the authors did not identify their criteria for website selection or list the websites utilized. Websites must be chosen carefully both to maximize the study’s exposure and to avoid recruitment bias. Partnering with academic treatment facilities to access e-mail distribution lists maintained by cancer research registries would have entailed obtaining additional Institutional Review Board authorizations, but it would have been a highly targeted recruitment method for this population.

Recent advances in software capabilities, coupled with increases in the speed of computer processing and data transmission, suggest that a number of recommendations for website design (Table 1) might be reconsidered. Graphics and audiovisual content are now less limited by bandwidth constraints and a visually attractive presentation may increase traffic to the study website (Silver 2005). Web 2.0 technologies (O’Reilly 2005) such as Asynchronous Java Script and XML (Ajax) can provide dynamic, individualized and interactive web presentations (McLaughlin 2005). These newer approaches would allow randomization of data collection tools and accommodate complex questionnaire branching patterns. The system could ‘remember’ the tasks that need to be completed and provide prompts to review skipped or incomplete questionnaire items. Subjects would not necessarily need to complete all components of a lengthy protocol during a single website visit because the system, using secure password protected individual accounts, would recognize the user and resume the study protocol at the appropriate point. These new capabilities would allow individualized presentations that may reduce the subject’s perception of participation burden resulting in higher instrument completion rates and reduced attrition.

The authors recommended a compulsory response format whereby respondents cannot skip questions, but must either answer each question in turn or discontinue the survey. This approach raises ethical concerns related to the voluntary nature of research participation mandated by the Belmont Report (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). An alternative strategy would be to provide respondents with the option of selecting ‘not applicable’ or ‘decline to answer’ which respects the autonomy of research participants.

Cantrell and Lupinacci’s (2007) paper illustrates the challenges of keeping abreast of new developments in a rapidly changing technological environment. The explosive growth of computer capabilities challenges researchers to find the most effective way to use web-based technologies while dealing with the associated ethical issues inherent in research with human subjects.