COMPARING PERCEPTIONS OF THE SYSTEMS ANALYSIS AND DESIGN COURSE
BRANDI N. GUIDRY University of Louisiana at Lafayette Lafayette, LA 70504 DAVID P. STEVENS University of Louisiana at Lafayette Lafayette, LA 70504

ABSTRACT Information Systems (IS) practitioners and educators have equal interest in the content of the Systems Analysis and Design Course (“SAD”). Previous research has examined instructors’ perceptions regarding the skills and topics that are most important in the teaching of the SAD course and the class time devoted to each. A similar assessment evaluated SAD course content from a practitioner perspective. Both studies used entropy calculations. A comparison of these studies is presented in this paper. For traditional topics, the group (either faculty or practitioner) with greater agreement believes the topic to be deserving of less class time. For structured and object-oriented topics, the group with the greater agreement also believes the topic to be of greater importance. This analysis demonstrates that practitioners and academics agree on approximately 40% of the SAD skills and knowledge areas. Keywords: Systems analysis and design, Structured analysis, Object-oriented analysis, Management Information Systems curricula, Entropy INTRODUCTION It is important that an education in Management Information Systems (MIS) is reflective of practices and techniques that are currently used in industry. Given the pace of technological innovation, there are ever-changing demands of technology workers [19] [30]. The content of each MIS course should be regularly compared to the skills that are required by employers, as an alignment must exist to ensure adequate preparation of students. As a significant component of the MIS curriculum and a primary focus of IS professionals, the content offered in the SAD course has recently received some attention in the literature [10] [25]. Because of the rapidly changing technological landscape, there is a constant need to develop new systems or to analyze existing ones [14]. Since poor project planning was found to be a principal cause of IT project failure in several major studies (see, for example, KPMG [15]), it is critical to ensure the proper alignment of knowledge and skills in the area of SAD, particularly with regard to specific methodological approaches used. The “IS 2010 Curriculum Guidelines for Undergraduate Programs in Information Systems,” established by the Association for Computing Machinery (ACM) and the Association for Information Systems (AIS), outlines specific objectives that are to be used as guidelines for the development of the SAD course [28]. Although it is recommended that this course, “provide some exposure to the structured SDLC, object-oriented analysis and design (some Unified Process variant using UML as a grammar) and agile methods” (page 53), one main objective of the course suggests to, “use at least one specific methodology for analyzing 40

a business situation (a problem or opportunity), modeling it using a formal technique, and specifying requirements for a system that enables a productive change in a way the business is conducted” (page 51). This focus is consistent with the recommendation to use a course project, as only one methodology would be used for its completion. Since the IS 2010 Curriculum guidelines identify SAD as a core component of an IS program, it is important to seek a better understanding of which specific methodology should be the focus when covering this content [28]. The chosen methodology should be suitable and fitting for application in a multitude of real-world IS settings. The differences in IS educator and IS practitioner perceptions pertaining to the required skills and knowledge of IS graduates has been documented in the literature [1] [16]. This study seeks to understand if these same perceptual differences exist specifically regarding the SAD component of an IS curriculum. As such, the importance of certain skills and topic areas in the teaching the SAD course were evaluated by academics and practitioners. This article compares these viewpoints to determine whether gaps exist between the areas academics perceive as important and those areas that practitioners perceive as such. Previous studies examined academic and practitioner views of SAD topic areas individually. The most recent of these studies [10][25]) are forerunners to the current research and form the basis for the comparison presented herein. Thus, this study addresses the following research questions: (1) Overall, are there significant differences between the SAD topic areas that are considered to be most important by academicians and practitioners? (2) Overall, are there differences between which SAD approaches (i.e. structured- or object-oriented) are being taught in the classroom and those used by industry professionals? (3) Are there differences between which structured approaches are given the most focus in the SAD course and those used in industry? (4) Are there differences between which object-oriented approaches are given the most focus in the SAD course and those used in industry? The results of this empirical study should lead to a better understanding of the perceptual similarities and differences between SAD educators and practitioners with respect to particular SAD approaches, tools, and techniques. Such comparisons should be considered in an effort to make more informed curriculum decisions and to ensure adequate preparation of MIS majors for their professional roles and responsibilities. Furthermore, faculty may find it helpful to compare their class time allotments to their colleagues as well as to the practitioner perceptions of allotment for that same topic. This would help Fall 2014

Journal of Computer Information Systems

individual SAD faculty members in their efforts to ensure a better knowledge alignment. This paper continues by providing a review of literature and presenting the methodology and analysis. Discussion and conclusions follow. LITERATURE REVIEW Curriculum gaps have been studied in many academic disciplines including accounting [4] [8] and marketing [3] [24]. Importantly, the examination of academic and practitioner perceptions often yields differing opinions regarding some topics, a recurring theme in most of the previously referenced studies. And, although the literature also reflects a continuing effort to understand the similarities and differences between MIS curricula and IT industry requirements [12][16][17][26], much less research attempts to assess the knowledge alignment in the specific area of SAD. Previous research on this topic revealed that although there is much disagreement amongst IS educators regarding SAD topic areas that are most important, structured methodologies (e.g., data modeling, process modeling, and data-flow diagramming) tend to be a common focus in the classroom (see Guidry et al [10]; Tastle and Russel [27]). Further, while the Tastle and Russell [27] survey results suggested that IS instructors were not yet embracing object-oriented methodological approaches to SAD, more recent research conducted by Guidry et al [10] suggests otherwise. This is also consistent with the survey results reported by Burns et al [5] indicating the majority of IS educators covered both structured and object-oriented approaches in their SAD courses. Some elements of both approaches have been criticized in the literature, with some authors making reference to the structured waterfall methodology and its lack of iteration and others suggesting that object-oriented methods involving the use of Unified Modeling Language (UML), for example, use an extensive amount of rather complex diagramming tools [9][20][21]. As a result, some authors have proposed a hybrid approach to SAD: one that incorporates both structured and object-oriented analysis and design techniques [2][7]. Sircar, Nerur, and Mahapatra [22] propose that although there are some fundamental differences in the analysis and design phases of the two approaches, some commonalities exist in the implementation phase. Thus, the knowledge and skills gained from those that are familiar with structured approaches are not completely different from those learned with more object-oriented methods. When describing the shift from structured to object-oriented approaches, Sircar et al [22] suggest, “The conceptual shift during analysis and design is considered architectural, whereas for programming it is deemed merely incremental” (page 457). Still others believe these two SAD paradigms are completely separate and any attempt to combine the teaching of both would lead to confusion in the educational and practitioner realms (see, for example,[20]). In an effort to better understand the relationship between educator and practitioner perceptions in the area of SAD, Anandarajan and Lippert [1] closely examined the research of both parties. Not surprisingly, their study revealed differences in the academic and practitioner research related to systems planning and development. Academic-oriented research is more focused on those areas of SAD that are related to the IS planning process, including theoretical and conceptual models pertaining thereto, while the research of IS practitioners focus more on experiences and investigations of problems related to SAD issues. Fall 2014

As suggested by the authors, these differences in research focus may help to further explain the apparent knowledge alignment issues that exist between these two groups [1]. Efforts to reconcile differences between educator and practitioner perceptions in the area of SAD are an important step in ensuring that students entering the workforce have the knowledge and skill set necessary to meet employer needs. Student preparedness is critical, as organizational consequences of underdeveloped IT skill sets can potentially lead to lower job performance [29], lower IS success [6], and can ultimately contribute to system failure. With the ever-changing nature of technology and the evolution of development approaches, educators face the challenge of staying abreast of the most widely used methods and techniques. The current study will shed some light on differing perceptions that may still exist today in an effort to bridge the curriculum gap that is evidenced by the literature. METHODOLOGY This paper compares differences in educator and practitioner perceptions regarding the topic areas that are most important in the teaching of an SAD course and the amount of class time that should be devoted to each. These perceptions were gathered from the previous research of the current authors for comparison purposes [10][25]. The educator survey was completed by SAD faculty members at AACSB accredited business schools throughout the United States. There were 124 usable responses from instructors at 64 different colleges and universities in the United States. The responses represent participants with various ranks (37.9% at rank of Professor, 29% at Associate Professor, and 21% at Assistant Professor for a total of 87.9% of respondents in a tenured or tenure-track position). Further, 41.9% of respondents had taught the course 5 years or less, 45.2% had between 6 and 20 years of experience, and 12.9% had taught the SAD course for more than 20 years. The practitioner survey was completed by 98 participants from four countries (Canada, United Kingdom, South Africa, and the United States) and 24 states within the United States. The respondents were from various industries and had a variety of IT job roles and functions. In general, participants had a great deal of work experience, with 32% of respondents having 10-19 years of experience, 31% having 20-29 years of experience, and 28% having 30 years or more of work experience. A survey instrument developed by Tastle and Russell [27] was slightly modified for use in both of these studies. Although the topic areas remained the same for both the educator and practitioner questionnaires, the wording of the questions reflected the particular audience that was being surveyed. See the work of Guidry et al [10] and Stevens, Guidry, and Aiken [25] for full copies of each survey and the number of responses given to each possible answer for each question. Beyond the general demographic questions (which varied slightly for educators compared to practitioners), there were three sets of questions that formed the basis for the surveys: 1) general questions regarding the percentage of time that is spent (educators) or should be spent (practitioners) in the teaching of more general SAD topics (i.e. data analysis, file and database design, process modeling, etc.); 2) questions specifically pertaining to the importance of structured SAD approaches (only if they were taught or used in industry by the practitioners); and 3) questions to identify perceptions regarding the importance of object-oriented concepts (only if such concepts were taught by 41

Journal of Computer Information Systems

the educator or used in the workplace by the practitioner). The Likert scales used for each of set of questions were identical to allow for comparisons between surveys. Specifically, an eightpoint scale was employed for the first set of questions, with options ranging from “none” to “= 50%. A five-point Likert scale ranging from “definitely important” to “somewhat important” and so on, up to “definitely unimportant” was used for the second and third sets of questions. More qualitative questions were posed at the end of each survey asking participants to list any CASE tools or other model-based software that was either used in the classroom or in industry. An opportunity to offer additional information that was relevant and noteworthy was provided at the conclusion of each of the two surveys. Survey respondents were assured of strict confidentiality and anonymity. Both surveys were pilot-tested to ensure ease of interpretation and reliability of the questions. Cronbach’s alpha was calculated separately for each of the three groups of questions for both the educators and practitioners. For educators, the Cronbach alpha values were .925 for questions regarding traditional topics, .864 for structured analysis questions, and .830 for object-oriented analysis questions. The internal reliability for the educator survey questions is therefore very strong, given the generally agreedupon lower limit of at least .7 [11]. Similarly, for practitioners, the Cronbach alpha values were .961 for the traditional topics, .790 for structured analysis topics, and .699 for object-oriented analysis topics. Therefore, the internal reliability for the practitioner survey ranges from very strong to acceptable over the three sets of survey questions. DATA ANALYSIS Recent research examining skills and knowledge areas for MIS courses have used a measure of the agreement (or equivalently, disagreement) between survey responses using entropy [27]. Entropy theory was first put forth by Shannon [23] at Bell Laboratories as a more intuitive way to examine differences between probability distributions. Tastle and Russell [26] demonstrated the advantages of using the entropy measure when comparing numerous probability distributions with varying means and standard deviations, but each with the same entropy value. Guidry et al [10] also use this measure, together with skewness, to determine the extent to which educators agree about the importance of and time spent on (i.e. “coverage”) of topics. Stevens et al [25] also use these two measures to determine the extent to which practitioners agree about the same topics. Hutcheson [13] developed a t-test to evaluate differences in entropy values, and so that test is utilized, together with t-tests for differences in mean values, to evaluate the differences in perceptions between educators and practitioners. The earlier studies also reported skewness values for each survey question, and those are repeated here for ease of reference. With the first set of questions, greater positive skewness indicates less time should be spent on the topic. With the second and third sets of questions, more positive skewness indicates the topic is more important. Shannon’s entropy, h(p), is calculated using the natural logarithm function ln(x), according to equation 1: h(p) = - ∑ pi ln(pi), (1)

the number of actual responses for category i by the total number of respondents answering the question. Also, n = the number of possible answers for each question. The minimum value of entropy is always 0 and occurs when all survey respondents give exactly the same answer to a question. The maximum entropy occurs when the answers are uniformly spread across all possible choices. Hutcheson [13] provides the following formulas (labeled (2), (3), and (4)) for the calculation of the t statistic and its degrees of freedom, for testing the null hypothesis that there is no difference in entropy value between two distributions: t = (h 1 – h 2 )/(V â r h 1 + V â r h 2 ) 1/2 (2)

where h1 and h2 are the entropy values, calculated as in equation (1), for the two distributions being compared, and Vâr h1 and Vâr h2 are estimates of the variance of each distribution, obtained from series expansion, and given by: Vâr h = [Σp i ln 2 p i – (Σp i lnp i ) 2 ]/n + (s – 1)/2n 2 + (-1 + Σp i -1 - Σp i -1 lnp i + Σp i -1 Σp i lnp i )/6n 3

( 3)

This t is compared to a critical t with degrees of freedom, d.f., given by equation (4): d.f. = [Vâr h 1 + V â r h 2 ] 2 /[( V â r h 1 ) 2 /n 1 + ( V â r h 2 ) 2 /n 2 ] RESULTS Comparing educator and practitioner perceptions of traditional topics Using the entropy values from the previous two research articles [10] [25], together with equations (2), (3), and (4), we obtain the data shown in Table 1. The values in Table 1 are grouped based on whether entropies between faculty (F) and practitioners (P) are equal, as well as whether their mean scores µF and µP are equal. Discussion of comparison of traditional topics The data in Table 1 are divided into 4 groups. In the first group are those survey questions that have answers differing in entropy between faculty and practitioners, as well as a difference in mean values of answers. These include survey questions 8–Data modeling, 9–Entity relationship diagram, 10-Normalization, 15-UML, 16–Class diagramming, 21–Interface design, 22–File and DB design, and 23–Program design. For example, survey question 8 asks about the amount of class time that should be spent on data modeling. The entropy for faculty respondents to this question is 1.720 and the practitioner entropy is 1.913. The t-value of -2.113 is significant with df = 211.5 (not shown), and so the amount of agreement between faculty and practitioners is statistically different and not due to chance. The smaller faculty entropy indicates they have greater agreement than practitioners regarding the amount of class time that should be spent on data modeling. “Skew, F” is the skewness for faculty responses to this question, while “Skew, P” is the skewness for practitioner responses to this question. The last two columns in Table 1 are the significance value for the difference in means, and the estimate of the difference in mean values, respectively. Degrees of freedom Fall 2014

( 4)

Where 0 < pi < 1, and i = 1, 2, 3, …, n. pi is the discrete empirical probability of the answer in category i, and is obtained by dividing 42

Journal of Computer Information Systems

TABLE 1: Entropy Values and Comparison Statistics for Traditional Topics

Practitioners believe more class time should be spent on these topics.

values for the test of entropy differences are not reported in the table since all are large (ranging from 179 to 222), making interpretation of the value of t basically the same as a standard z score. The second group have different entropies but their means are not statistically different at the alpha = .10 level. (In this research, alpha = .10 is chosen because the questions regard personal opinions (see, for example, [18]). This group includes survey questions 4–Overview of SA, 5–Project initiation, data, 6–Project management concepts, and 19–Cost-benefit and payback. The third group contains questions that do not differ in entropy, but do differ in mean value. These include survey questions 2–Structured analysis and 3–Object-oriented analysis. The fourth and final group contains those questions that do not differ in entropy or mean value, and include survey questions 7– Overview of methodologies, 11–Process Modeling, 12–Data flow diagramming, 13–Decomposition diagramming, 14–Use case, 17–Sequence diagramming, 18-State-transition diagramming, and 20–Systems design concepts. In the first group, it is evident that practitioners believe more class time should be spent on the topics represented in questions 8–Data modeling, 9–Entity relationship diagram, 10Normalization, 21–Interface design, 22–File and DB design, and 23–Program design as compared to educators. Practitioners, Fall 2014

however, feel that less class time should be spent on the topics represented in questions 15-UML and 16–Class diagramming as compared to faculty. Importantly, as it pertains to the first and second grouping (and many of the topics of the other groupings as well), there is always greater agreement amongst the group that believes less class time should be spent on a particular topic. It is also interesting to note that the difference in mean scores for questions 10-Normalization, 15-UML, 22–File and DB design, and 23–Program design each correspond to an approximate 5% difference on the Likert scale. Further, the difference in mean scores for questions 8–Data modeling, 9–Entity relationship diagramming, 16–Class diagramming, and 21-Interface design each correspond to an approximate 2.5% difference on the Likert scale. Entropy calculations for the second grouping indicate that faculty members have more agreement on all topics in this section. And, although a significant difference was not found when comparing means, it should be noted that practitioners believe more class time should be spent on each of the topics in this group. The third group of topics consists of only 2 survey questions: the amount of class time that should be spent on 2-Structured analysis and 3-Object-oriented analysis. Because the entropy values do not differ, it can be said that faculty and practitioners 43

Journal of Computer Information Systems

have the same amount of agreement or disagreement about these. However, faculty answers are approximately one-half point (.573) higher for structured analysis and almost one point (.834) higher for object-oriented analysis. These differences indicate practitioners believe less class time should be spent on each: about 2.5% less time on structured analysis, and almost 5% less class time on object-oriented analysis. There are no significant differences in entropies or means for each topic represented in the fourth grouping. This, therefore, suggests that SAD faculty members and practitioners are mostly in agreement with respect to the amount of time that should be spent on 7-Overview of methodologies, 11-Process modeling, 12-Data flow diagramming, 14-Use case, 17-Sequence diagramming, 18-State-transition diagramming, and 20-Systems design concepts. The lack of significance for difference in entropy between faculty and practitioners for these questions indicates that the amount of agreement regarding these topics is basically the same. Comparing educator and practitioner perceptions of structured analysis topics Like Table 1, Table 2 utilizes data from the previous 2 research studies, together with values obtained from equations (2), (3), and (4) to show the statistical significance of the difference in entropy and mean values between faculty (F) and practitioners (P). Like Table 1, Table 2 is also divided into groups. As indicated in Table 2, the results of survey questions 24–Data modeling concepts, 25–Entity relationship diagram, 27–Draw complete data flow, and 34–Team projects show a difference in entropies between faculty and practitioners and also a difference in mean values. The second group consists of questions 28–Balance data flow diagram, 32– Project management skills, and 35–Interviewing techniques, the answers of which do not result in a difference in entropies, but do result in a difference in means. Finally, the third grouping consists of questions 26–Normalize data model, 30–Activity dependency diagram, 31–Both data and process modeling, 33–Data collection

and survey results, and 36–Use CASE tool. The answers to these questions do not differ in entropy or mean value. Discussion of comparison of structured analysis topics In looking at the results of the survey questions pertaining to structured topics, it is important to note that a lower entropy value indicates greater agreement. Further, if the mean difference is positive (µF - µP > 0), practitioners perceive the topic to be more important and vice versa. Thus, it is clear that practitioners think 24-Data modeling concepts and 25-Entity relationship diagrams are more important than faculty do, and practitioners are also more in agreement about the importance of this topic than faculty. This may be because data modeling concepts and ERDs are typically introduced in a separate MIS course. According to the IS 2000 Model Curriculum, data modeling techniques should be initially covered in the “data and information management” core course offering [28]. Thus, most MIS students are exposed to this material prior to taking the SAD course and faculty members may only feel the need to quickly review these topics as a part of the design phase of the systems development methodology. Conversely, educators perceive 27–Draw complete data flow diagram and 34Team projects to be more important than practitioners do, with faculty also having greater agreement about the importance of these topics. Significant differences in means were recorded for the second grouping. Faculty perceive 28-Balance data flow diagram and 32-Project management skills to be more important than practitioners, while practitioners feel that requiring students to actually perform 35-Interviewing techniques is more important than faculty do. Interviewing techniques are an important element of the requirements gathering phase of systems development. From a faculty member’s perspective, however, it may be quite difficult to teach such skills within the confines of one semester. It is often not possible to offer students the opportunity to participate in a real-world project because of time constraints, so mock interviews associated with a fictional SAD project may

TABLE 2. Entropy Values and Comparison Statistics for Structured Analysis Topics

Practitioners believe these topics to be more important than do faculty. 44 Journal of Computer Information Systems Fall 2014

be the only exposure students have. Further, although there was no statistically significant difference in the amount of agreement amongst faculty and that of practitioners in the second grouping, it is interesting to note that faculty members were more in agreement with respect to the results of questions 28–Balance data flow diagram and 32–Project management skills, while practitioners were more in agreement when looking at the results for question 35–Interviewing techniques. This consistency was evident throughout most of the groupings. In fact, for all questions pertaining to structured analysis topics except 36-Use a CASE tool (to implement a business model) , regardless of whether statistical differences were noted with entropies or means, the group that believes the topic is more important also has greater agreement about that topic. There are no significant differences in means in the third grouping, which means that faculty and practitioners alike view these topics (26-Normalize data model, 30-Activity dependency diagram, 31-Both data and process modeling for a project, 33Data collection and survey skills, and 36-Use of CASE tool) as being equally important. There is also no real difference regarding the extent to which faculty members agree about the importance of these topics as compared to practitioners. Thus, an overall knowledge alignment exists for these topic areas. Comparing educator and practitioner perceptions of object-oriented analysis topics Table 3 repeats the analysis conducted for traditional and structured topics, and summarizes the corresponding values for the object-oriented questions: survey questions 37 through 42. Discussion of comparison of object-oriented analysis topics Among these questions, there are no statistically significant differences in entropy value, which indicates that the amount of agreement regarding these topics is basically the same. Only questions 31–State-transition diagramming and 41–Cost-benefit and payback have different means, indicating a difference in perceptions of importance of these topics: practitioners believe 39-State-transition diagramming to be more important, while faculty believe 41-Cost-benefit and payback to be more important. The remaining questions which include 37-Class diagramming, 38-Sequence diagramming, 40-Completing an object model using project management skills, and 42-Use model-based tools to implement a design have alignment in terms of importance between faculty and practitioners.

As with the two previous groups of questions, whenever µF - µP > 0 (see question 39-State-transition diagramming), the practitioners believe the topic to be more important. In these cases, practitioners also have more agreement (although the amount of their agreement is not statistically different from that of faculty). Whenever µF - µP < 0 (as with question 41-Cost-benefit and payback), faculty believe this topic to be more important and faculty have greater agreement for this question. These two relationships are true for all of the object-oriented questions, with the exception of question 42–Use model-based software for design. The difference in mean scores for questions 39-State-transition diagramming and 41-Cost-benefit and payback each correspond to one-half point on the Likert scale. This is equivalent to halfway between somewhat important and definitely unimportant. This is evidenced by the fact that 16.8% of faculty believe 39-Statetransition diagramming to be unimportant (either “somewhat unimportant” or “definitely unimportant”), compared to 5.2% of practitioners. Conversely, 80.5% of faculty believe question 41-Cost-benefit and payback is important (either “definitely important” or “somewhat important”), compared to 55.3% of practitioners. CONCLUSIONS Overall, the results of this comparison study indicate practitioners believe traditional topics, and especially structured analysis topics (such as data modeling, normalization, file and DB design, and program design) are more important and worthy of more class time than most object-oriented topics. Among the 12 traditional questions with the most statistically significant differences, practitioners think more time should be spent on 10 of 12 topics, especially those which involve more managerial skills. This result appears to clearly answer the first 2 research questions. Overall, practitioners value traditional approaches significantly more than faculty, and also value object-oriented approaches significantly less than faculty. Regarding questions specifically dealing with the importance of structured analysis topics, faculty believe four skills (2 modeling skills and 2 managerial skills) of the twelve to be more important while practitioners believe three skills (2 modeling skills and 1 managerial skill) of the twelve to be more important. It is understandable that practitioners value managerial skills less because students are generally not hired as managers out of school. Experience is generally considered a prerequisite for management, as indicated by project management certifications,

TABLE 3. Entropy Values and Comparison Statistics for Object-Oriented Analysis Topics

Practitioners believe these topics to be more important than do faculty. Fall 2014 Journal of Computer Information Systems 45

such as PMI. These results appear to clearly answer the third research question. Namely, there are significant differences between which structured approaches are given the most focus between faculty and practitioners. Regarding questions specifically dealing with the importance of object-oriented analysis, there was only disagreement in terms of average response among two questions of the six. As was the case with the structured analysis topics, practitioners valued a technical skill more than they did a managerial skill. Since faculty and practitioners agreed on the remaining 4 of 6 questions, the fourth research question appears to be unresolved. There is no statistically different agreement (in terms of entropy) between faculty and practitioners, although there is a clear difference in perception regarding the importance of two of the six topics. Overall, there was agreement regarding class time for traditional topics on 8 of 22 questions, or 36%. There was agreement for the importance of topics (for the combined questions in structured analysis and object-oriented analysis) for 9 out of 18 topics, or 50%. The 9 topics of agreement consisted of 5 of 12 of the structured analysis topics and 4 of 6 object-oriented topics. Hence, cumulatively, there was agreement on 17 of 40 topics, or 42.5%. What are the reasons for differences in faculty and practitioner perceptions of these topics? First, as noted earlier, practitioners generally do not expect students to leave college with managerial skills. Therefore, practitioners value topics such as 34–Team projects, 32-Project management skills, and 41–Cost-benefit and payback analysis less than do faculty. In addition, the practitioner respondents had considerable job experience in their field: 91% of respondents had 10 or more years’ experience, 59% had 20 or more years, and 28% had 30 or more years’ experience. It is therefore not surprising that they valued traditional topics such as 21–Interface design, 22–File and DB design, and 23-Program design more since these skills are relevant to both traditional and object-oriented approaches. This is significant since objectoriented tools and newer methodologies (such as Agile) have seen their greatest growth in the past 10-20 years, after many of the practitioners were already employed. In the authors’ experience, the complexity and knowledge required to implement an object model throughout an enterprise is directly proportional to the size of the enterprise, making traditional approaches preferable for most project teams in government agencies and large corporations. Because there is disagreement on approximately 60% of the topics in the survey, faculty are encouraged to answer the survey questions (see Guidry et al [10]) themselves and then compare their answers to those given by practitioners (see Stevens et al [25]. In this way, individual faculty can better align their class time allotments to the various topics with the expectations of those in industry. This is necessary because there is no single “best answer” for every individual. Some are already spending considerably more time than practitioners would recommend, while others are spending considerably less, and still others approximately the same. Although the results of this study alone are not meant to be a prescription for curriculum modification or development, they should be reviewed and considered in ongoing efforts to bridge the gap between educator and industry expectations. Continued research should be conducted in this area, as the discipline of information systems always has and continues to rapidly change. IS scholars, practitioners, and students will benefit from this research focus. One method that is commonly used for keeping 46

communication lines open between faculty and practitioners is the use of departmental advisory boards [24]. Academic-practitioner research and consulting collaborations are also a means of aligning expectations. As a result of differences in perceptions, graduates should be aware of various methodologies that exist in order to appreciate and apply appropriate models to a variety of systems development projects. Although IS students cannot be expected to master the use of every tool, technique, and methodology during their undergraduate education, they should possess the knowledge and skills necessary to adapt or to apply the tools to which they have been previously exposed. Survey comparisons such as the one presented here provide educators with information to make appropriate decisions regarding the topics covered in the SAD course, as well as how much time should be devoted to each. REFERENCES [1] Anandarajan, M. and S.K. Lippert, “Competing Mistresses? Academic vs. Practitioner Perceptions of Systems Analysis,” Journal of Computer Information Systems (46:5), 2006, pp. 114-126. [2] Bateveljic, P., M. Eastwood, and H. Seefried, “An Approach to Teaching Object-Oriented Analysis and Design,” Journal of Information Systems Education (17:3), 2006, pp. 267272. [3] Brennan, R., “Should We Worry about an “AcademicPractitioner divide” in Marketing?,” Marketing Intelligence & Planning (22:5), 2004, pp.492-500. [4] Bain, C.E., A. I. Blankley, and L. M. Smith, “An Examination of Topical Coverage for the First Accounting Information Systems Course,” Journal of Information Systems (16:2), 2002, pp. 143-164. [5] Burns, T., “Defining the Content of the Undergraduate Systems Analysis and Design Course as Measured by a Survey of Instructors,” Information Systems Education Journal (9:5), 2011, pp. 4-17. [6] Byrd, T.A., and D.E. Turner, “An Exploratory Analysis of the Value of the Skills of IT Personnel: Their Relationship to IS infrastructure and Competitive Advantage,” Decision Sciences (32:1), 2001, pp. 21-54. [7] Carte, T.A., J. Jasperson, and M.E. Cornelius, “Integrating ERD and UML Concepts When Teaching Data Modeling,” Journal of Information Systems Education (17:1), 2006, pp. 55-63. [8] Crawford, D.L., “Practitioner and Educator Preferences Regarding Accounting Curriculum Meeting the 150-Hour Requirement,” Academy of Educational Leadership Journal (15:4), 2011, pp. 47-66. [9] Dobing, B. and J. Parsons, “How UML is used,” Communications of the ACM (45:5), 2006, pp. 109-113. [10] Guidry, B.N., D. P. Stevens, and M. Totaro, “The Systems Analysis and Design Course: An Educators’ Assessment of the Importance and Coverage of Topics,” Journal of Information Systems Education (22:4), 2011, pp. 333-348. [11] Hair, J., W. Black, B. Babin, and R. Anderson. Multivariate Data Analysis (7th ed.), Upper Saddle River, NJ: Pearson Education. 2010 [12] Havelka, D. and J. W. Merhout, “Toward a Theory of Information Technology Professional Competence,” The Journal of Computer Information Systems (14:3), 2009, pp. 106-116. Fall 2014

Journal of Computer Information Systems

[13] Hutcheson, K., “A Test for Comparing Diversities Based on the Shannon Formula”, Journal of Theoretical Biology (29), 1970, pp. 151-154. [14] Kohli, R., and J.N.D.Grupta, “Effectiveness of Systems Analysis and Design Education: An Exploratory Study,” Journal of Organizational and End User Computing, (14:3), 2002, pp. 16-31. [15] KPMG, “The Global IT Project Management Survey,” 2005, retrieved March 5, 2013 from http://www.kpmg.com/CN/ en/IssuesAndInsights/ArticlesPublications/Documents/ Global-IT-Project-Management-Survey-0508.pdf. [16] Lee, S, S. Koh, D. Yen, and H. Tang, “Perception Gaps between IS Academics and IS Practitioners: An Exploratory Study,” Information & Management (40:1), 2002, pp.5161. [17] Lee, D.M. S., E.M. Trauth and D. Farwell, “Critical Skills and Knowledge Requirements of IS Professionals: A Joint Academic/Industry Investigation,” MIS Quarterly (19:3), 1995, pp. 313-340. [18] Lind, D., W. Marchal and S. Wathen (2012). Statistical Techniques in Business & Economics, 15th edition, McGrawHill/Irwin: New York, NY, p.337. [19] Medlin, B.D., D.S. Dave and S.A. Vannoy, “Students’ Views of the Importance of Technical and Non-Technical Skills for Successful IT Professionals,” Journal of Computer Information Systems (42:1), 2001, pp. 65-69. [20] Rob, M.A., “Dilemma Between the Structured and ObjectOriented Approaches to Systems Analysis and Design,” The Journal of Computer Information Systems (46:3), 2006, pp. 32-42. [21] Siau, K., J. Erikson, and I.Y. Lee, “Theoretical vs. Practical Complexity: The Case of UML,” Journal of Database Management (16:3), 2005, pp.40-57.

[22] Sircar, S., S. P. Nerur, and R. Mahapatra, “Revolution or Evolution? A Comparison of Object-Oriented and Structured Systems Development Methods,” MIS Quarterly (25:4), 2001, pp.457-471. [23] Shannon, C.E., “A Mathematical Theory of Communication,” Bell System Technical Journal (27), 1948, pp. 379–423. [24] Stern, B.L. and L.P.D. Tseng, “Do Academics and Practitioners Agree on What and How to Teach the Undergraduate Marketing Research Course?,” Journal of Marketing Education (24), 2002, pp. 225-232. [25] Stevens, D.P., B.N. Guidry, and P. Aiken, “The Systems Analysis and Design Course: A Practitioners’ Assessment of the Importance and Coverage of Topics,” International Journal of Innovation and Learning, forthcoming. [26] Stevens, D.P., M. Totaro, and Z. Zhu, “Assessing IT Critical Skills and Revising the MIS Curriculum,” Journal of Computer Information Systems (51:3), 2011, pp. 85-95. [27] Tastle, W.J. and J. Russell, “Analysis and Design: Assessing Actual and Desired Course Content,” Journal of Information Systems Education (4:1), 2003, pp. 77-90. [28] Topi, H., J. Valacich, R. Wright, K. Kaiser, J. F. Nunamaker, Jr., Sipior, J. and G. J. deVreede, “IS 2010: Curriculum Guidelines for Undergraduate Degree Programs in Information Systems,” retrieved January 5, 2013 from http://www.acm.org/education /curricula/ IS%202010%20ACM%20final.pdf. [29] Wade, M.R., and M. Parent, “Relationships between Job Skills and Performance: A Study of Webmasters,” Journal of Management Information Systems (18:3), 2001, pp. 7196. [30] Weber, R., “Some Implications of the Year-2000 Era, Dot-Com Era, and Offshoring for Information Systems Pedagogy,” MIS Quarterly (28:2), 2004, pp. iii-xi.

Fall 2014

Journal of Computer Information Systems

47

Copyright of Journal of Computer Information Systems is the property of International Association for Computer Information Systems and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.