Experimental research in consumer behavior is typically conducted by asking undergraduate students to complete a pencil-and-paper survey administered in a classroom or laboratory. While this allows relatively large numbers of people to participate in any one session, it greatly constrains the sorts of experimental manipulations and controls that can be used. Recently, researchers have started using personal computers to collect experimental data. While this allows researchers to present carefully controlled manipulations and use many different, more interesting and realistic media, few people can participate at any given time.
The UF's Department of Marketing has extensive experience with such controlled research. It has used an experiment laboratory in Bryan Hall for a number of years. It has facilities for the traditional pencil-and-paper tests and surveys administered under controlled physical conditions. The facility has also collected a number of microcomputers over the years. The current collection ranges from 10 year old PS/2 machines running the first Windows platform to two machines running Windows 95.
The World Wide Web (Web) offers researchers the opportunity to merge the best features of pencil and paper tests and computer based approaches. A group of current, similarly configured LAN-based micros with adequate server capacity will permit us to expand our ability to do controlled research via graphical user interface (GUI) modes with a reasonable number of participants. The existing capacity effectively eliminates this mode of testing.
With current, similarly configured LAN-based micros, our researchers can design experiments that present information and collect responses using multimedia Web authoring tools such as Macromedia Authorware, Macromedia Director, or Alliare Cold Fusion. In addition, tests and surveys can be completed over the Web using HTML and a forms-capable browser and in the laboratory. This will enable us to create metrics showing, among other things, systematic differences due to controlled laboratory conditions versus generic Web conditions. This could substantively add to or knowledge of the extent to which controlled experimental conditions attempting to simulate real-world conditions do, in fact, give results similar to those same tests given in participants' residences.
We will also be able to test the extent to which prior knowledge of and experience on the Web affects users with typical and advanced students. A common complaint about computers in general and many software applications is that the new user faces a formidable, often frustrating learning process to be able to use the tools effectively and efficiently. We have to opportunity to assess the extent to which this effect could constrain e-business. We have an opportunity to pick participants for our research with known differences in Web knowledge and skills. We could find that such background differences are systematically associated with major differences in our research outcomes on e-business. Or, that their background differences dot no matter. Either result could provide direction for designing electronic commerce applications.
Such differences may be inextricably bound together. If the new shopping method requiring network access and Web browsing is too complex for the average consumer, he or she may eschew it for traditional shopping. At the same time, those experienced with network access and Web browsing could find the alternative preferable to traditional shopping. This is a possible outcome since many researchers have shown that the decision rules customers use depend on the effort required to collect and process information.
An implication of the outcome would be that substantial development effort is necessary to make network access and Web browsing much more "user friendly" to the average user. Conversely, we might find that significant differences in users knowledge and experience in network access and Web browsing have little impact on the usual information collection effort. Whatever the result, this research could provide valuable insights into the design of the electronic commerce human-machine interface.
The UF College of Business has, effective June 29, 1998,
computer equipment and software literacy minimum
requirements. Our
We will use for our typical student participants individuals
enrolled in UF College of Business'
Marketing 3023. This course is usually taken during the Fall
Semester of the Junior year. Students in the course have the option of
participating in
experiments for extra credit.
In contrast, our advanced students, in addition to having the minimum
computer hardware and software literacy requirements noted above
and Marketing 3023, will be seniors majoring in
Decision & Information Sciences (DIS). They will have completed the
following courses in their
major on computing
languages and those that use software packages.
2 x 2 Factor Design
This mix of participants and locations makes possible a 2 x 2 factor
design.
However, there are complications regarding locations
outside the Marketing Department's experiment laboratory on the fifth
floor of Bryan Hall. These complications must be reflected in
the choice of student participants to assure the factor design is
valid.
Differences in transmission speeds between directly connected
laboratory-based clients and servers as compared to Web-based clients in
the community dialed into those servers over analog phone lines are
very large, and known. So, transmission speed is part of the
location difference in the 2 x 2 factor design noted above.
However, we do not yet know the effective transmission speed
differences between directly connected clients and those connected via
cable and/or ADSL lines. We will eliminate cable
or ADSL line users in our initial research. When a sufficiently large
number of student participants have cable and/or ADSL service, they will
be systematically included in future experiment structures.
The benefits of this approach include:
We hope to have the machines installed and procedures in place by the
beginning of the Fall Semester, 1998. We intend to begin running
experiments by October 1998 and to have the data analyzed by May 1999.
Preliminary reports of the results should be available by July 1999.
Research in Electronic Commerce
We can simultaneously conduct the above methodological research and
explore basic issues in electronic commerce. Below we propose a
few research projects related to electronic commerce.
These projects will be run in a staggered fashion. Data collection
will begin in October 1998 for the first project and run through July
1998 for the last project. Data analysis will begin in January
1999 for the first project and run through October 1999 for the last
project. Research results will be available by August 1999 for
the first project and by March 2000 for the last project.
The URL for this site is,
We will periodically post progress reports to this site.
This research provides an opportunity to extend the already recognized
expertise of the UF marketing department into the area of electronic
commerce. The marketing community is very interested in developing
on-line counterparts to traditional channels. However, it is not
known whether theories of consumer behavior that are valid in traditional
shopping environments will apply equally well to on-line shopping sites.
This research will help marketers understand the fundamental
similarities and differences between traditional and on-line consumer behavior.
This research will also benefit the research community by providing a test
of whether behavioral experiments can be run reliably over the web.
Academic departments expend considerable manpower to run experimental
subjects.
Web experiments, if reliable, could allow researchers to collect data with
much less effort.
The research will also provide evidence on the extent to which extant computing
and Web knowledge of individual users contributes to the level of effectiveness
and efficiency in using the communication media. Evidence that such
extant knowledge does or does not affect users would be valuable information
to builders of applications for e-Business and all other applications on
the Web.
IBM can benefit from this research program in the following ways:
We intend to publish the most theoretically interesting and managerially
relevant results of this research in leading marketing and consumer
behavior journals. These include:
We are requesting equipment to (1) provide a uniform environment for
experimentation in the Marketing Department's experimental laboratory in
Bryan Hall and (2) permit the two faculty that will be working directly
with participating students to demonstrate procedures in their offices on
the same equipment. The following spreadsheet image includes the
detailed IBM descriptions, part numbers, and prices provided by UF's
IBM representative.
NOTE: This will be updated on receipt of cost information from Ms.
Bonnie Clark, IBM.
© 1999 University of Florida, all rights reserved.
It was last updated on 17 April, 1999.
All students must have a notebook computer at the start of their junior
year. The required specifications for these notebooks include,
Many College of Business classrooms will have at each seat an electric
outlet and a 10BaseT socket giving access to the college's LANs.
As of this writing, Room 101 in the building named Business
Administration has this wiring installed. nbsp;Room 101 is located on the
ground level where the letter "A" is shown on the UF map with
the building name "Business Admin" at this
site.
They will be in their senior year and enrolled in the UF College of
Business' ISM 4220, Introduction to Business Datacom, with Dr. Elnicki for their
instructor. They will have previously covered in ISM 4220,
These DIS students may also be concurrently enrolled in up to 3 of 6 additional major course requirements in computing and quantitative methods.
Their participation conditions will be equivalent to the students in Marketing 3023.
Laboratory
Residence Typical Student
Advanced Student
Project Web Site
Value to Marketing and the College of Business
Value to IBM
Plans to Publish
We will also publish working papers related to this project on this website.
Proposal Budget
