Cyberspace: Oxymoron or New Research Frontier

Abstract:
This article explores the research possibilities of the Internet. Although a substantial body of writing focuses on ways of obtaining and using the information available through the Internet, few articles discuss the use of the Internet as a survey tool. This paper examines a new research methodology, wherein participant’s data is analyzed,delivered and monitored via electronic means.Limitations do exist. These limitations include the lack of a targeted population, the lack of random samples and the self-selection of the participants. However, these limitations may be outweighed by such advantages speed, cost efficiency and ease of data collection and manipulation. The use of the Internet as a survey tool is certainly worthy of consideration.

Introduction:
the explosion of the Internet has substantially altered human interaction. With the advent of the Internet, a worldwide network of almost 16 million computers, used by an estimated 30 million people (Network Wizards, 1997) has crossed time and space, creating a virtual space and community. While no one knows the actual number of new users, estimates suggest the numbers have doubled each year over the past five years. This trend will no doubt continue, due to the declining prices of personal computers, the increasing number of schools and universities providing Internet access; the development of faster communication devices and the shrinking access and service fees.

Internet survey:
Companies have been taking advantage of the Internet to promote their products. In addition, they have begun to use Internet surveys to increase their understanding of their customers. This explains why commercial domains account for more than 60% of total domains. (Internet, 1996) While questionnaires have been the most popular research vehicle, this method does have numerous disadvantages. Observation can provide better data, however the test population is limited by resource requirements. Combining these two methods and using the World Wide Web (WWW) and Hypertext Markup Language (HTML) as the communication vehicle, the best of both worlds can be obtained while minimizing the disadvantages.

HTML and WWW are both recent, related technologies. These new technologies have generated extreme interest in everyone from private individuals to large corporations and even governments. WWW originally provided test and graphics with easy point and click hyperlink navigation. WWW has since grown into a multimedia tool that can collect information worldwide and provide this information to anyone with access to the Internet.

System Requirements:
The proposed research methodology is supported by a mid-to-high end WWW server connected to the Internet via a high speed T1 connection (1.56b per second). A web site with a web page is necessary to house the questionnaire. Microsoft’s FrontPage program or any of the other programs for creating web pages can easily create the web page. Fisher, Margolis and Resnick (1996) developed a survey instrument using ASCII language. The survey instrument was posted at ListServ . Unfortunately some subjects changed the instrument. The authors recommended that the instrument be programmed “to discourage or prevent” subjects from changing the questionnaire. Therefore, user questionnaires should be developed in HTML with Common Gateway Interface (CGI) scripts to capture choices electronically and prevent subjects from changing the questionnaires. All research information should be converted to HTML and presented through the use of a WWW browser (i.e.Netscape). Access to the survey would be observed electronically as each subject accessed the information. The respondents would also be able to identify any additional information that was needed.This data would be collected and reported. Types of information, viewer time, frequency of use and the path of access to information would all be recorded by the WWW server for immediate preliminary analysis. Final analysis would be conducted after a significant number of responses were collected.

The power of computers could help separate users into different demographic groups. Each participant could be asked to select their user group from a pull down menu containing a list of test populations. Additional information could be obtained during this process. This additional information could include, but would not be limited to, information regarding age, sex, education and work experience, certifications and associated professional organizations. Participants would receive invitations and information on how to participate in this research through e-mail and newsgroups. Participation would be limited to only those aware of the Universal Resource Location (URL) for the research WWW site. All those accessing the site would be allowed to participate. Without access control to a site available on the Internet, user validation will be difficult. The on-line questionnaire could be used for user validation, as well as to determine possible biases and current decision making processes of the respondents. The numbers of Internet users indifferent groups provides a very large potential population for researchers to draw from.

The response rate for current questionnaire and observation research methods is 10% to 30%. This rate can cause problems due to the low initial test population. If the test population becomes too small, meaningful statistical analysis comes into question. In contrast, test population size would not be an issue with the use of the Internet as a survey tool, due to the expected high response. However, data retrieval and analysis could become monumental tasks. To handle such a large number of participants, observations would be conducted in real-time via monitoring each user’s access to the survey instrument provided to them through the internet site.

Since the data would be collected in real-time, preliminary analysis could be conducted at any time during the test period. If it is deemed necessary to modify any portion of the information presented during the test period, this could be accomplished without any additional requirements to future participants. All data previous to the change could be used in the final results with a disclosure of this change.

Internet research data collections:
There are numerous data gathering techniques for collection information. The most common of these techniques are focus groups and telephone and mail questionnaires ( Hanke, Reitsch 1994). Identification of the study population is crucial, so that representative samples may be chosen from the population. This presents a challenge, as there are no lists of individuals using the Internet. Additionally, not everyone in the study population may have access to the Internet. Finally, an Internet user may have multiple Internet addresses; this makes choosing representative samples even more complicated.

There are two types of group user environments on the Internet; Listserves and Usenet newsgroups. ListServ is an Internet mailing list for people with shared interests. Anyone can subscribe to the list and exchange information. Subscribers can specify topics of interest. They will only receive messages addressing these interests, thus reducing the amount of junk e-mail. The mailing list is maintained by the list owner/moderator. The Internet address is http://www.neosoft.com/internet/paml. The publicly accessible mailing list receives about 42,242 hits per day according to ListServ statistics. It is quite active.

Usenet newsgroup is another group environment. It is not a formal network, but rather a number of computers that exchange e-mail with pre-assigned subject headers. The e-mail is referred to as an article; the subjects’ headers are news groups. As with ListServ, anyone can subscribe to the lists and post articles in the newsgroups. Newsreader software, such as newsreel and newsreader, is necessary to obtain the news or article posting from the new server.

It is quite easy to obtain comprehensive listing of newsgroups and mailing lists. The search engine is fast and easy to use. The sampling design is very simple. For example, if the research test population definition is accounting professionals, one simply types the keyword “accounting” in the search engine and all available mailing lists under the keyword “accounting” are shown. (See examples in attachment 1) Since the cost of sending e-mail is minimal, one could select all the mailing lists of the targeted population. The next step would be to e-mail the selected list administrator and ask for permission to post an invitation on the survey news, inviting all the subscribers to visit the research web page.

This procedure is in sharp contrast from the survey method suggested by Fisher, Margolis and Resnick (1996). They posted the survey directly to the subscribers’ e-mail addresses. The current research method invites the subscribers to visit the research web page and participate in the survey. Fisher, Margolis and Resnick experienced problems including the alteration of their survey instrument and subscribers forwarding surveys to non-subscribers. Therefore they had difficulties managing the survey data. With the survey housed in a Home Page, Fisher’s problems could be avoided.

Data Retrieval and Analysis:
Survey data is retrieved in real-time. Once a subscriber completes the questionnaire and clicks on the “submit” button at the bottom of the screen, the data is captured and placed in the research database. The associated observation information is also captured in real-time. As each subscriber navigates through the survey instrument, a time stamped data entry is made into the research database. By collecting all the time stamped entries for each user, decision-making profiles can be developed. This profile would show what information was requested, how long they visited each piece of information, and the sequence of the information visited. Repeat subscribers would only be counted once.

Once sufficient data has been collected, Internet access to the WWW server will be terminated. Analysis of the data in the database can then be started. Analysis of the questionnaire data would determine if any participants did not meet the defined criteria. If so, their information access data would be discarded or flagged. Once this task has been completed, analysis of the remaining data could be started. Simple determination of user class profiles, as well as the number of times a piece of information was visited, could be completed quickly. However, determining what class of user visited each piece of information, and the sequence of visits would require additional programming and a statistical analysis application.

Limitations:
The Internet is convenient and has the potential to access a large population. However, the targeted population should be a representative population, so as to validate the research. That is, the characteristics of the Internet subscribers are assumed to be the same as non-subscribers in the same population. This may be difficult to ascertain as there is no list to verify the information.

Finally, there is no way to control the number of survey responses a subscriber can send from different e-mail addresses. If one subscriber has more than one e-mail address, he or she can send one response for each address.

Conclusion:
The Internet is a valid and promising survey tool. While there are some limitations, the potential benefits far outweigh the drawbacks. The speed and ease of data collection and manipulation are considerable advantages. The Internet is the research tool for the 21rst century.

References:

Elliott, H. (1995, June) “More distributors set up shop in the cyberchannel.” Electronic Business Buyer, 21, 16-18

Fisher, Margolis and Resnick (1996) “Breaking ground on the virtual frontier: surveying civic life on the Internet.” American Sociologist, Spring 1996

Fouss (1996) “Faster and Smarter, Technology and globalization are defining the researcher of tomorrow.” Marketing Research; Winter 1996

Hanke and Reitsch (1994) Understanding Business Statistics 2nd edition. Irwin Publishing.

Hoffman & Novak (1996) “Marketing in Hypermedia Computer-mediated Environments: Conceptual Foundations” Journal of Marketing, July-September

Internet 1996, Third Edition, Sams Net Publishing

ListServ Homepage, Public Accessible Mailing Lists-Frequent ask questions. Download from Internet address Http://www.neosoft.com/internet/paml/answers.html

Network Wizards (1997), Internet Domain Survey. Menlo Park, CA:

Rao (1996), “Nielsen’s Internet Survey: Does it carry any weight?” Forturn, March 18, Vol. 133 Issue 5, P. 24