RESEARCH STATEMENT

RESEARCH OBJECTIVE

My career goal is to discover new facts, procedures, methods, and techniques and provide intellectual support and contribution to the business and the academic community in the area of information systems.

RESEARCH METHODOLOGY

I am interested in multidisciplinary research that has a combination of behavioral and technical elements. A multidisciplinary approach allows me to apply uniquely my diverse educational background and work experience to topics of interest. Specifically, my research interests lie in the areas of Multimedia and decision support systems, developing and managing business computing systems (e.g., decision support systems [DSSs], group decision support systems [GDSSs], and expert systems), and electronic commerce (EC).

My primary research program focuses on multi-media decision support systems. In my dissertation, I am investigating the role of text and image in decision support system. This dissertation is multidisciplinary in nature and involves such megatrends as behavior of decision maker and decision support systems. My secondary research projects include studies in DSS and the impact of time stress and monetary reward system. All these research studies are explained in Summary of Research Activity section.

I do not have a predisposition towards any particular type of research methodology. I believe that the methodology should be dictated by the research question. For instance, if the research question involves investigating causal relationships, then a lab experiment seems appropriate; on the other hand, if very little research exists on a topic, an exploratory study with case studies and/or structured interviews may be a good starting point.

In my dissertation, I am using laboratory experiment by randomly selecting human subjects by using monetary reward systems to accomplish experiment in a specific media (i.e. Text based decision support systems or Image based decision support systems) under a three time windows.

SUMMARY OF RESEARCH ACTIVITY

The following discussion summarizes what I have done so far along the dimensions of my interrelated research interest listed above. These activities resulted in a number of conference proceedings, and working papers to be submitted for publication in near future.

1.         A Relative Assessment of Two Decision Support Systems

A decision maker must frequently acquire information sequentially from a support system, within a given time period, in order to make an ultimate choice among alternatives. Within this time frame, he/she must repeatedly evaluate whether to select an alternative based on current information or to postpone the decision in order to acquire additional information and thus improve chances of making a correct choice. The strategy used may cover a wide spectrum ranging from being completely optimal to completely non-optimal depending on the extent of deviation from optimality.  Given a certain source mode (i.e., a mode of presenting information in text or an image shape), a decision maker may perform differently in the decision-making environment than if given a different mode.

My research interest is examining the relative benefits of two kinds of information modes used for supplying information where the information content could appear at any of four complexity levels. It does so via an experiment exercised within the context of a time-pressured setting very similar to those faced by the decision makers a variety of applications.

To implement the above experiment (and other, similar investigations), an extensive experimental platform was designed, developed, and refined.  This platform contained a training tool set that helps one conduct subject performance-related laboratory experiments under various complexity levels for various information source choices. The platform offers an investigator control over such critical parameters as the information modes, the complexity of information presented for each mode, the matrix of monetary incentives, and boundary values for time windows. The design, implementation, adaptability, and portability of this platform are major contributions of this research.

After finalizing the platform, it was implemented using object oriented programming approach Visual Basic 5.0.

2.                  An Expert System for Advising Students

Using Guru Expert System, design, and develop a prototype for advising students in Morehead State University.

3.                  Decision Support Systems for Career Choices

Using Guru Expert System, design, and develop a prototype DSS for career opportunities for students in Morehead State University.

This was built for students who were interested to see their options in jobs based on their selected criteria.

4.                  Evaluating Of An Advising Support System: An Experimental Laboratory Study

            This was proposed to show the importance of a good advising system in higher education.  An advising system is an example of an unstructured information system which could be solved by the decision maker by acquiring information, developing solution techniques, and/or altering standards of representation (each with their special costs).  A good advising needs to be pointed out in terms of an idiosyncratic knowledge.  A good advising system, also, be treated as a contract which based on a Agent‑Principal theory among four partners society, educational institution, academic advisor and students in which the contract involves incentives and transaction costs.  This proposal discusses various ways academic advising services are currently provided by institutions and examines factors to consider in evaluating and organizing an optimal academic advising system.  An evaluating model of advising system will be presented through a laboratory experimental setting.

5.            Evaluating Of Expert System Tool: Quantitative Method

            Regardless of their styles, today's managers depend on information from reliable sources (eg. human/machine expert system) in support of decision-making.  As we search the functional areas of business, there is a significant amount of interest in expert systems and the shells that are available for developing these systems.  These shells can be used for research, prototyping and developing end-user applications.  It would be wise to establish a mathematical approach that can be used to evaluate these shells.

DISSERTATION SUMMARY

Title:  SEARCHING FOR INFORMATION: EXPERIENCES WITH A TEXT-BASED & AN IMAGE-BASED DECISION SUPPORT SYSTEM

This dissertation examined the potential relative benefits of two kinds of information modes used for supplying information, where the information content could appear at one of the four complexity levels. The experiment was conducted within the context of a time-pressured setting very similar to those faced by the decision-makers in our example scenarios. The experimental task involved decision making under the condition of value-driven strategies for information gathering employed by subjects vis-à-vis “optimal” (but not obvious) strategies built into the experiment by the investigator.

To implement the above experiment (and other, similar investigations), an extensive experimental platform was designed, developed, and refined.  This platform contains a training tool set that helps one conduct subject performance-related laboratory experiments under four complexity levels for each of the two information source choices. The platform offers an investigator control over critical source-related parameters like the information mode, complexity of information presented for each mode, matrix of monetary incentives, and time windows

boundaries. This platform not only features portability but also because it is coded in object oriented Visual Basic, one may edit it for various classes of subjects. The design, implementation, adaptability, and portability of this platform are major contributions of this research.

This research possesses the following unique features to set it apart from other media-related studies. It not only incorporates the concept of a time horizon by establishing time bounds on key aspects of subject-system interaction, but it also implicitly captures the notion of “time pressure” through:

(a)     Reward Mechanism: the reward mechanism relates closely to subject performance in terms of speed, and accuracy (i.e., its heavy reliance on Vernon Smith’s work on induced value theory in experimental economics). Driving subject performance by performance-based monetary rewards alone in a strictly controlled experimental setting hinges largely upon Smith’s postulate of non-satiation (refer to Chapter 2). The experiment simulates essential features of a class of realistic systems but without any emphasis on “realism” in our experiments to maximize control over subject performance by the reward mechanism and the reward mechanism alone. Subjects ought not to be motivated to perform for any other reason, context realism included.  This assurance underlies the extensive body of literature that repeatedly validates the principle in a variety of empirical tests in the microeconomic theory and DSS evaluation arenas.

(b)     Task complexity: because a given time bound for decision making may prove satisfactory for one type of task, performing as well with a more complex task within the same time bound proves temporally and cognitively more challenging.

(c)     Search for Information: By adding a feature of searching in sequence for the adequate information, one may test the behavior of decision-makers under various task complexities in a certain mode.  In this procedure one may examine whether the decision-maker is over-acquiring and incurring excessive information, under-acquiring and incurring excessive risk of decisional error, or adequately acquiring information.

(d)     Optimality vs. Non-optimality of Decision-Making: determined based on searching for information.  According to the deviation matrix built up in the experiment, one may structurally analyze how far away the user’s searching and final decisions appear from an optimal one.

(e)     Ex-ante approach: This research adopts the ex-ante approach to system evaluation whereby one deliberately applies a hypothetical system to a set of hypothetical tasks. Many system-related studies use the ex-post approach that requires a fully implemented system.

The experimental results show no significant differences in decision-making performance between those supported by text mode presentations versus image mode presentations.  This is in spite of the fact that persons using the image mode had to learn a new language, while those using the text mode of decision support did not.  The implication is that it is possible to design compact image languages (e.g., for hand held devices) without sacrificing decision performance.  Indeed, factoring out the time for learning the image language, suggests that an image language may be a more effective interface mode for decision support.  Future research can use the platform to explore this possibility.

FUTURE RESEARCH PLANS

My dissertation has initiated what should prove to be interesting and fruitful streams of future research.

Apart from conducting experiments with all complexity levels and time windows (from that used in this study), the existing platform can be applied readily in the following settings as well:

(a)     The researcher may allow a subject to choose an information mode from the existing two sources of his/her choice.  These will be used with all experiments in a session or allow him/her to pick a combination of both sources in the session. The experiment may compare performances with established sources (investigator set-up) against those realized with “variable” (subject-selected) information sources.

(b)     The researcher may apply various reward mechanisms and analyze subject’s behavior under each matrix. Another for a share of a fixed pool of reward money. As one earns more, less money remains for other participants.

(c)     The researcher may establish the decisional tree of memorizing patterns on only a page of a notepad

rather than showing them on a set of cards on display screen.  After collecting data, compare them with existing platform experiment data.

The above examples illustrate a few of the experiments possible with little modification to the existing experimental platform. Extensions may undertake for other interesting experiments.  These extensions involve the following:

(a)     We tested the task in English text.  Researchers may use non-English languages for various cultures. Establishing one common symbolic language (i.e., the image mode we used) gives an opportunity for cross-cultural testing.

(b)     Researchers may establish the task in meaningful scenarios for English and non-English subjects, rather than meaningless and abstract task that we used.

(c)     Researchers may create “composite” sources that mix various modes within a single experiment in a session (e.g., part of the information appears in the Text mode while the remainder appears in an image mode.)

(d)     A researcher may revise the coding so that experiments can be conducted with other types of modes such as moving images and voice.

(e)     Researchers may add color, or change the size of the font, as a control variable to the text and image modes currently in place. This is particularly interesting for the environments where the device display features such as size and memory limitations may be particular concerns. Similarly, variants of the audio mode may be constructed by allowing voice to convey “moods” like seriousness, aggravation, sadness, passion, happiness, excitement, and such.

(f)     Researchers may establish other kinds of information sources than the two used here (e.g., full-motion video, still images, moving images, graphs, charts, music), and various other context-defined task complexity levels.