Posted by arlene

Internet Survey Confidentiality

Coomber has raised the issue of respondents‘ concerns about anonymity and confidentiality, especially when dealing with sensitive topics or illegal practices. In his survey research on illicit drug dealers he found that users were concerned about their responses being obtained by people other than the researcher. Several options are available that avoid users‘ contact details being passed on to the researcher. While email responses will typically carry the respondent’s email address in the reply,’ filling in a Web-based form will not. If respondents are still concerned about confidentiality, they can be encouraged to complete the survey on an anonymous machine (for example, in a library or Internet café) or to print it off and post it. Kaye and Johnson report that in a survey on attitudes to political information on the Web, 276 out of 306 respondents complied in giving their email address details, suggesting that, in some cases at least, anonymity is not a major concern to respondents.

Internet Survey Access control

Stanton has raised concerns about the effects of access control in Web-based surveys, suspecting that attempts to control the sampling frame by requiring participants to, for example, enter a password to access a survey page may arouse concerns about anonymity and confidentiality. This may then lead to reduced response rates, or reduced candidness. To date there is no research to suggest that this effect occurs. Whereas Stanton urges researchers to assume, for now, that implementing access control does reduce response rates and candidness, we would encourage social and behavioural researchers to aim to maximise control over who participates in the study. Keeping track of who has had the opportunity to participate in the study, that is, being able to measure the sampling frame, is essential in learning more about factors such as response rate and non- response bias in IMR. Given current concerns about the skewed demographics of the Internet-user population, such control is vitally important. Kaye and Johnson share our view that unrestricted access in Web surveys should be avoided, suggesting that passwords provide a useful method for restricting access to a select and measurable sample.

Internet Survey Technical issues

Some researchers have noted particular technical difficulties, which are worth documenting here in order that the reader can be aware of similar problems in his or her own research and take measures to guard against these. Smith reports that using the ‘mailto:’ command in the survey form in order to send responses directly to the researcher’s email address, caused a problem in that the Microsoft Internet Explorer browser did not support this command. This meant that respondents using this browser were unable to submit their responses. There would also be problems if the browser was not correctly configured to send mail, for example because its SMTP server address was incorrect. Another problem reported by Smith concerned the ‘Thankyou’ pop-up box (implemented using JavaScript) failing to appear when the respondent pressed the ‘Submit’ button, resulting in multiple submissions of the same data set being sent as the user tried repeated submissions. The latter problem is something that might occur if the user has JavaScript disabled in his or her browser. We suggest minimising reliance on more advanced procedures where feasible. For more complex presentations that do rely on such features it may be useful to specify these requirements in an introductory paragraph, though this must also be weighed up with the possible effect of discouraging some users by sounding too technical. Some users may have never heard of JavaScript, for example, while still using a JavaScript-enabled browser. Careful piloting should be undertaken to avoid such technical difficulties.

We recommend the use of server-side scripts (these are explained in detail below) as a method of data collection and storage. There are several reasons for this: first, problems with browsers not supporting `mailto:’ are avoided; second, users‘ email addresses are not automatically sent back with their data (a barrier to full anonymity); third, responses can be directed to a file, sent to an email address, or both; and, finally, many browsers pop up a warning about security when the ‘mailto:’ command is invoked, which may discourage some users from submitting their responses. We also emphasise the importance of thorough piloting across different platforms to check that the survey works as intended.’

Possibly related posts: (automatically generated)
Internet Survey Confidentiality; Access Control & Technical issues

• Security Facilities in Java
• Adding JavaScript to the HTML Form
• Network and Programming: the .Net framework
• How to Design and Implement an Internet / Web / Web Page / Site Survey (Internet User analysis)
• VBScript and JScript continue...
• An ASP-Based Windows Server Script
• Java Database, J2EE Framework
• Internet Survey Response rates & Presentation Format
• How to Design and Implement an Internet / Web / Web Page / Site Survey (Design Issue & Postal Survey)
• Why your Computers get attacked and how? PC Privacy Security Issues continued