1Do Object-Oriented Methodologies Communicate More Effectively Than Traditional Methodologies Dale Lunsford High Point University Donna F. DavisUniversity of Southern Mississippi ABSTRACT An experimental situation was utilized in which novice end-users were asked to study information systems documentation, produced either by the Structured Analysis methodology or the Object-Oriented Analysis methodology, and then to recall features or to answer questions related to their understanding of the information system. The study found no differences in the performance of the subjects on either dimension of performance, thus refuting those claims that the Object-Oriented Analysis methodologies are more natural, more closely related to the structure and functions within the scope of the information system, and thus more effective at depicting the information system structure and functions. 1. Introduction It is becoming increasingly important that auditors, managers, and other end-users of information systems understand the manner in which data is processed through those systems. This need comes about as a result of both the rapidly expanding influence of information systems in organizations, and of recent legislation expanding the responsibilities of management and auditors for the accuracy of financial statements. This paper investigates the ability of certain categories of end-users of information systems to understand the components and behaviors of information systems when utilizing models produced by two methodologies. These two methodologies are Object-Oriented and Structured Analysis. Previous studies of the effectiveness of these2methodologies have focused on the performance of systems analysts, who are skilled in information systems modeling. The study reported herein is concerned with the ability of non-information systems professionals to understand the documentation, including the models, of an information system. The Sarbanes-Oxley Act of 2002 (SOX), passed by the Congress of the United States in 2002, explicitly describes the responsibility of management for the accuracy of financial statements as well as the responsibility of auditors to attest to the adequacy of controls over transaction processing. Managers must certify that the information system controls are adequate, and thus must understand the processes which feed data to the financial statements. SOX authorized the creation of the PCAOB, which in turn issued Auditing Standard (AS) No. 2 requiring the auditor to obtain sufficient knowledge of the process and procedures that process transactions in order to determine that the controls are adequate to preclude material misstatements (PCAOB 2004). As the controls are likely to be contained partly or wholly within computer-based processes (PCAOB, 2004, para.40.), auditors must be competent in the interpretation of information systems documentation. A major tool for fulfilling the responsibilities of both auditors and management is a set of models which represent the activities, data and relationships among entities required to achieve the desired outcomes of the organizations information processing activities. While a number of information system modeling paradigms exist, the two most like to be encountered by users are Structured Analysis (Gane and Sarson [1979],3DeMarco [1979], Yourdon [1989]) and object-oriented systems analysis using Object-Oriented Analysis (Coad and Yourdon [1991a]). While the originators of object-oriented methodologies claim that the models produced using these methodologies provide a more natural representation of the information system and, as a result, are more effective communication media leading to better comprehension of the information system, there have been few, if any, empirical studies conducted to evaluate differences in the various types of models’ability to communicate knowledge to end-users. This paper investigates the ability of certain categories of end-users of information systems to understand the components and behaviors of information systems when utilizing models produced by two methodologies. These two methodologies are Object-Oriented and Structured Analysis. Previous studies of the effectiveness of these methodologies have focused on the performance of systems analysts, who are skilled in information systems modeling. The study reported herein is concerned with the ability of non-information systems professionals to understand the documentation, specifically the models, of an information system. 2. Background There are several categories of information system models of interest to users, those depicting behaviors and those describing the data structures utilized by the behaviors. These will be referred to in this paper as function and activity models respectively. The behavior models encompass both those that provide a high level view4of the overall goal-directed activities of the systems, and those that depict the low level individual activities. Structure models are those that describe the entities of interest in the information system and the relationships among those entities. These models are also known as data models. Prior research, summarized in table 1, investigated the effects of various models and representations upon the performance and satisfaction in selected tasks and are not thus generalizable to the entire range of tasks faced by those who interact with information. In addition, prior research has focused on the examination of specific structure models or behavior models in isolation with no attempt made to examine the influence of complementary models that are used in practice. This is especially troublesome for two reasons. First, in practice behavior and structure models are generally presented as sets; because of this, presenting models separately presents an unnatural setting. Second, there is a possibility that there will be differences in the recall and understanding that are achieved when models are presented as sets.5Table 1 Prior Research Related to Performance and Satisfaction with Information System Modeling Constructs Author Treatments Findings Smith (1989) Narrative only and narrative with graphical representation. Graphical representation omitted a critical component. Performance in narrativ Larsen and Naumann (1992) Physical Data Flow Diagrams (concrete model) vs. Logical Data Flow Diagram (abstract model). Subjects better able to