Posted by arlene

Email Surveys Simple text-based email surveys

Implementing a simple email survey requires little technological expertise, and can be done with minimal software requirements. In its simplest form an email survey can consist in simply sending an email containing text questions in the body, and asking participants to hit the ‘Reply’ button on their mailer, fill in their text answers to the questions, then hit ‘Send’ to return the completed survey questionnaire to the researcher’s email account.’ This procedure was used by Hewson who administered an experimental questionnaire-based study via email to investigate the folk psychological concept of belief. The procedure very simply consisted in pasting the text (which included a short story followed by a question) into the body of an email message and sending this to participants. Participants responded by sending back an email that included their answer to the question.

There are many advantages in the email survey methodology. Perhaps the most salient, for many researchers, is the ease and lack of technical expertise required. This will not only prove useful for the non-computer expert researcher, but also increase the pool of available participants. A disadvantage may be the lack of control over response format. Hewson found that respondents replied in a variety of formats. Most included the question materials with their answer below, but some just sent back the answer, which would have caused problems if details of the condition sent to each participant had not been recorded. This was something that had not been anticipated.

As already discussed, when sending surveys by email the researcher needs to first obtain the participant’s email address. We recommended techniques that maximise knowledge of the sampling frame, as opposed to using large commercially available lists of email addresses. Hewson posted participation calls to several newsgroups, asking interested parties to reply to an email address indicating that they would like to take part in the study, or requesting further information. This methodology was considered successful in that it generated 135 responses within a period of approximately two weeks. Response rates were not able to be measured in this case because it was not known how many people had seen the request. The issues of dormant accounts, accounts that are checked infrequently, and users with multiple accounts were also raised. These issues are less likely to be problematic for methods that access respondents from newsgroups (either by posting a request, or obtaining email addresses from postings) than for methods that use large lists of harvested email addresses. Nevertheless, further techniques can be employed to try to avoid some of these problems, which still may occur. Sending a receipt request (an option on several mailers, for example, Netscape communicator) can be useful in helping determine whether the user has actually opened the email message, though a returned receipt cannot guarantee that the message has actually been read, and lack of a returned receipt certainly does not warrant the conclusion that the message has not been read (users can reject receipt requests). The method of posting participation requests on web pages, inviting interested parties to respond by email, is also a possibility, but we have discouraged this general approach due to the problem of lack of knowledge of the sampling frame (though a number of researchers have piloted ways of keeping track of web site visitors, and this approach could become viable in the future).

While quick and easy to implement, the text-based email survey methodology does have drawbacks. First, the limitation to text-based materials makes it unsuitable for some types of research that require more sophisticated graphics. Much survey research is text-based, but even implementing something like a Likert-type response scale is likely to be ungainly via text-based email, as compared with an HTML form, for example. A further consideration with this method is that anonymity is not preserved since emails can be traced to individuals by the header with sender address information. (As mentioned previously, anonymous email services are available, but relying on this technique will place additional demands on respondents and thus limit the accessible respondent pool.) In sum, a text-based email approach in IMR may work well in some cases but prove too limited in others.

Email-based surveys with graphics

Moving beyond email surveys that use only text, there are several possibilities available. One procedure is to email a file as an attachment, which can contain graphics. The participant can open and edit this and send back to the researcher. Most users will have the required skills to open file attachments, which is particularly straightforward on Windows-based systems, for example. However, platform variation may limit accessibility and participation: different platforms, for example Windows, UNIX, will be able to handle different file formats. This issue needs to be carefully considered. HTML files can be sent as email attachments, and mailers such as Netscape communicator will automatically display these as HTML documents in the body of the email message. Thus if an HTML form is sent, users will simply be able to respond to this form within their browser in the same way that they would respond to a WWW form online. The recipient will simply fill in the form and click the ‘Submit’ button while connected to the Internet, which will return the data to a script on the researcher’s server (which will then save the data to file or mail it to the researcher). The approach reduces the effort required by the participant, compared with having to open and edit a file, then email this back to the researcher.

Thus email can be used to send anything from very simple text questions, to sophisticated HTML forms that are processed by a server script. The latter approach allows the most flexibility and sophistication, and we now turn to discuss how to construct such a form and place it on a web server, bearing in mind that it is equally feasible, and in many cases more desirable (for reasons already discussed), to control who receives the form by sending it via email.

WWW Surveys

Creating a simple Web-based survey is actually not that difficult. The reader can use the code given here in order to produce the survey (HTML form) itself, and should have little difficulty in modifying this survey to fit his or her own requirements, with reference to the useful resources on HTML provided. Placing the survey on a suitable server may require more skill, or help from technical staff. However, once this modest barrier has been crossed, it should be easy to produce further surveys and place these on the server in a similar manner.⁹ To produce a program that inputs the data received directly into an analysis package (such as SPSS) requires more programming expertise, and we do not cover this here. Some readers may find commercially available survey construction packages useful, though these are costly and tend to be aimed particularly at market researchers.

Implementing a WWW Survey

This section demonstrates how to set up a simple survey using HTML forms and collect the data using a basic server-side script, then introduces a few JavaScript techniques with which the survey page can be enhanced. The survey page should work with any modern browser, and we describe server scripts that will work on an appropriately configured UNIX server and on a Windows server. While no experience with HTML is required to get the survey page working, no explanation of the non- form elements is given here. Readers who wish to learn more about HTML should look elsewhere, in any of the many books and online sites dedicated to the subject.

Possibly related posts: (automatically generated)
Software Procedures for Implementing Internet Surveys

• VBScript and JScript continue...
• Client-Side Scripting
• Adding JavaScript to the HTML Form
• The Apache Web server, a rich Java Web site
• VBScript and JScript
• The Apache Web server, a rich Java Web site continue...
• Professional Video Design and Presentation
• Security Facilities in Java
• HTML Coding for different E-mail Reader Platforms
• Hot HTML: Stealing Others' Content for a Web Site