Report Print

ATAL BIHARI VAJPAYEE INDIAN INSTITUTE OF INFORMATION TECHNOLOGY AND MANAGEMENT, GWALIOR

A B Tech. Dissertation on

A Fuzzy Based Personality Assessment Technique Submitted In Fulfillment of the Requirements Of Five Year Integrated Post Graduation Programme

Prepared under the Guidance of:

Submitted By:

Dr. Joydip Dhar

Abhishek Vaid

(IPG200407)

Dr. Naval Bajpai

Animesh Tohan

(IPG200410)

Divij Chandwani

(IPG200424)

CERTIFICATE This is to certify that the dissertation entitled “A Fuzzy Based Personality Assessment Technique”, being submitted to ABV-Indian Institute of Information Technology & Management, Gwalior by Abhishek Vaid, Animesh Tohan & Divij Chandwani, is in partial fulfillment of 5-year Integrated Post Graduate degree. The present dissertation is a record of bonafide work under our supervision and further certified that the work reported in this dissertation has not been submitted earlier to any University or Institute for the award of any degree or diploma.

Dr. Joydhip Dhar

Dr. Naval Bajpai

Associate Professor

Associate Professor

ABV-IIITM

ABV-IIITM

ACKNOWLEDGEMENT We hereby express our gratitude to our supervisors, Dr. Joydip Dhar and Dr. Naval Bajpai for their invaluable guidance throughout the course of our project. It was indeed a great honour and a great learning experience to work under such unerring professors and having received their ever helping cooperation and wisdom. We shall be forever indebt and grateful for all the support provided and the resources that were made available to us. We dedicate our project to their immense knowledge and expertise in their respective areas without which the project would not have been satisfactorily completed. We would also like to thank all the individuals who devoted their precious time in responding to our survey.

Abhishek Vaid 2004IPG07

Animesh Tohan 2004IPG10

Divij Chandwani 2004IPG24

Contents PART 1 ■ Overview 1.1 Project Outlin e 1.1.1 1.1.2

Abstract Introduction

1 1

1.2 Executive Summary 1.2.1 1.2.2 1.2.3 1.2.4 1.2.5

Problem Statement Project Goal Scope of use Complexities Method ology 1.2.5.1 1.2.5.2 1.2.5.3 1.2.5.4 1.2.5.5

1.2.6 1.2.7

Phase 1 – Phase 2 – Phase 3 – Phase 4 – Phase 5–

2 2 2 4 4

Method ology Method ology Method ology Method ology Method ology

Results & Conclusions Future Prospects

6 6 6 8 8 8 9

PART 2 ■ Project – Planning 2.1 Project – Initiation 1.2.8 1.2.9 1.2.10 1.2.11

Project Project Project Project

– – – –

Idea Scope Feasibility Definition

2.2 Literatur e Survey 2.3 Project Phases

10 11 13 14 15 17

PART 3 ■ Methodology 3.1 3.2

Phase 1 : A – Case Study 19 Phase 2 : A Fuzzy Based Working Model 3.2.1

Language 3.2.1.1 3.2.1.2 3.2.1.3

3.2.2 3.2.3 3.3

19 20 23 23

Assessing – Personality: An Extended Methodology Assumptions and Conclusions 26

24

Phase 3 : Industry Level Data Collection 3.3.1 3.3.2 3.3.3

3.4 3.5

SET T Trait Vector SET P

19

Questionnaire 27 Data And Interpretation Assumptions and Conclusion

28 32

Phase 4 : A Real W orld Application (Theory) Phase 5 : Assumptions And Conclusions (Theory)

33 33

PART 4 ■ Results & Conclusion PART 5 ■ Future Prospects References Appendices A List of Figures B Phase 1 methodology C Questionnaire D Snapshots of Excel Sheets

Survey Declarations Glossary

34 37

PART 1 ■ Overview

1.1

Project Outline

1.1.1

Abstract

There is a lot of literature focused on the job of assessing an individual’s personality; more than several theories exist as we speak. These theories emerge from a common foundation and all are co-related with one another in an interesting fashion. Two such most successful theories are 5 Factor Model, and MBTI Type Indicator. Both these theories are much acclaimed in the field of psychological modeling and share some similarities, but having said this, both these theories are also fundamentally very confined in modeling the behavior. Here we propose a fuzzy logic technique to exte nd these (MBTI to be specific) methodologies to greater boundaries and increasing their utilities.

1.1.2

Introduction

Currently the Standard MBTI test is used to classify the participants in one of the 16 MBTI classes. As of now the MBTI based tests are focused to classify the participants in one of the 16 MBTI classes. These Classes are mentioned Figure 1.-1.

I I

ST ISTJ ISTP

SF ISFJ ISFP

NF INFJ INFP

NT INTJ INTP

J P

E E

ESTP ESTJ

ESFP ESFJ

ENFP ENFJ

ENTP ENTJ

P J

Now the problem with such an approach is that more than one individual could fall into same class (any out of 16) and thus a very important question remains unanswered: How do we further distinguish between the people to the same MBTI class? To mediate this inherent shortcoming, we proposed a model which is very straightforward and essentially helps in resolving the scope of unattended ambiguities currently prevailing in MBTI theory of personality.

Part 1 Overview

Page 1

1.2 Executive Summary 1.2.1

Problem statement

In this section we threw some insight into the problem at hand and went on to define our problem statement. We understand the importance of defining the problem statement in any project; whether inclined technically or managerially. We associated paramount importance to the task of defining the problem statement, as we considered it to be a very vital phase in any project and thus in our project [13]. We made sure that our problem statement was easy, simple and complete and thus, defined our problem statement underneath: Problem Statement: “To do an exploratory research in understanding the scope and usage of personality inventories obtained using MBTI theory and further improve the applications of these inventories by improving the present personality assessment methodology with the help of Fuzzy based questionnaires”

1.2.2

Project goal

Firstly, just like we mentioned in the above section, we decided to perform an exploratory research in the field of personality inventories and their applications. Personality Inventories are by definition “An assessment tool used to determine which of these personality types a person falls into: extroverted, introverted, thinking, feeling, sensing, intuitive, judging, and perceptive. It is used as part of a self assessment done for career planning purposes“. [11] Personality inventories are used extensively in various kinds of organizations to facilitate the process of decision making regarding HRM Therefore we aimed and channelized our project towards improving the applications of personality inventories by enhancing the accuracy and flexibility currently prevalent in personality assessment techniques. Secondly, for the purpose of our study, we decided to consider theory of MBTI. MBTI theory is an acclaimed theory used in the field of personality assessment. There are of course other theories aimed at assessing personalities but we chose MBTI because of its ease of use and vast applicability. So another vital goal of our project was to understand and improve the application of MBTI theory. Thirdly we decided to choose a real time application of personality inventories and work to improve the scope of application. We decided to improve the method by which these Inventories are currently created [3] i.e. various ways of assessing an individual’s personality through MBTI Theory. Therefore in broader sense our main project goals were following: T o improve the accuracy and flexibility of currently prevalent personality inventories. To understand the theory and application of MBTI theory in making personality inventories. To improve (If possible) the ways of accessing personalities using MBTI theory, thus improvi ng the quality and usage of personality inventories.

1.2.3

Scope of use

There are overly optimistic researchers in academic and corporate arena who think that personalities are good ways of clustering people. These clusters may represent a project team, particular group of customers who show similar taste for products or a bunch of people who show similarities in behavior and lifestyle. Clustering people according to some performance or preference criterion has always been beneficial, commercially and productively. Here we provide some of the market scenarios where our project could improve the commercial viability of personality inventories and reap greater benefits for various genres of business.

Part 1 Overview

Page 2

Web Portals : Imagine the growing market of web portals. Today web portals come in variety of flavors; our particular concern is Social Networking portals, Matrimonial portals and Business Networking Portals. These portals have become very rapidly the cyber hub for almost anyone who at least understands the basics of Internet. Now realizing the fact that huge numbers of internet users are attached to one of more such portals, we could realize the concept of automated personality assessment. Let us try to explain the concept to its grave. Imagine, for example a matrimonial portal; now it’s very important for these portals to provide to its customers the right kind of people to get in touch with. Today portals just put up the profile of an individual, as mentioned by them (with certain level of authentication) and leaves the mature decision of personality assessment to a user for her to interpret on her own. Now many a times a user can’t probe into potentially thousands and thousands of members of a portal. Therefore a need for an automated personality assessment module immediately arises as a consequence. Designing such a portal is a tedious and commendable achievement in itself, but what I’ d like emphasize upon is that in current context, such a module is sure to utilize an approach which is non trivial and may benefit from concepts of Fuzzy Logic. The scope of Fuzzy Logic is easy to realize because running a personality assessment test which is in form of an automated survey would have to handle intuition intensive opinions on part of a user. Now such a data has a huge scope for vagueness which we need to model accurately as the efficiency and accuracy of a module is in direct relation with the proficiency of such a modeling technique. Therefore such vagueness could be handled with incorpo rating the fundamentals of Fuzzy logic in data collection and process of analyzing and translating that data at later stage.

Marketing : Today marketing has become a complex job at hand. Specialized marketing firms rely on several factors for digging out marketing opportunities to provide its clients with the right kind of people they want. As per the current methodology these specialized marketing firms prefer to divide the customers ( population) into clusters based on information such as income, age, occupation, housing and race [ 12]. All of these parameters are important in their own sense but the list is far from complete. It suffers fro m the absence of a factor of vital significance – Personality. If only there exist a way to track down the personality of the individuals, these companies would make a swift effort to make it count for the purpose of marketing. Obviously such a thing doesn’t exist as of now, that is why this list is deprived of it, but we believe that with these project is capable of mediating this problem and make personality a critical parameter to make important marketing decisions and strategies.

Software Engineering : Today 66 % of the projects fail. [14] Some of the experts in the field of Software Engineering attribute such a high amount of failure to various factors like lack of user involvement, long or unrealistic time scales, poor or no requirements gathering, scope creep, no change control system, poor testing, but most believe that lack of team coordination and poorly assigned roles is one of the most significant ones. Some researchers even believe that roles reflect personality [3]. Since our project aims at improving the accuracy, scope of usage & applicability of personality inventories, it is bound to help in making proficient project teams. There are many more niche areas for personality inventories to be applied to serve the cause and enhance the productivity.

Part 1 Overview

Page 3

1.2.4

Complexities

Just like any project development lifecycle there were challenges and complexities associated with our project. Since our project relies on new foundations and subtle assumptions about personalities we need to be over vigilant against these challenges, as they may kill the very purpose of our project. We based on our understanding found following challenges to be extremely vital:1. It has to be a standalone module compatible with most web based platforms so that it can be easily integrated with any one of them. 2. It need not be hundred percent accurate, as It just needs to work with suffiecient level of accuracy and precision. However, with more and more usage and data collection the System should be capable to learn and produce more realistic and accurate results. 3. It has to be simple so as to be easily understood by people who really don’t understand the discipline of psychology, behavioral sciences and other related fields. It should be fairly accessible to almost all kinds of people. 4. It needs to have a fun element attached to it. As the module is supposed to make fairly accurate predictions about patterns of personalities in people, it needs to achieve this with an element of excitement and fun. 5. It needs to provide a good, relative and comprehensive feedback. Keeping in mind all the points which just mentioned, and assuming that we comply with them. We can vision a web module which is capable of assessing personalities automatically We could ask users to run there personality assessment tests and compare their assessments against each other. This process can be done proactively and results can be computed at the backend level. This process will be fully automated and without need of anyone at any stage. A lot many further enhancements can be done at the time of deployment to achieve further levels of customization. We tried implementing a small portal ourselves with our limited technical skill set but couldn’t because of lack of time and resources; However this adds good future prospects to our project.

1.2.5

Methodology

As per the description above, our project looked strenuous and demanding. There were various factors which were absolutely indispensible for this project to prove its worth and achieve its claims. Some of the most important issues we needed to sort out as to ensure the healthy and productive handling of the project are following: 1. Strong Understanding of basic fundamentals in disciplines like Behavioral Sciences, Basic Psychology & Organizational Behavior. 2. A Strong web development skill set, for development of online module of personality assessment. 3. A continuous access to good databases for the purpose of literature survey. 4. A good financial funding in case we need to invent primary data for our research. 5. Availability of good in-house experts in the fields mentioned in first point. These experts may include the members of academic community or of commercial community. 6. An extended project development time of around 6 – 8 months, which may even stretch up to a year if we try to consolidate our findings with good amount of primary data. There are other important issues as well, but they are of little significance when compared with the ones mentioned above. Out of all the issues 2nd, 4th & 6th. were more serious in terms of threats they offer to our project development scenario. Due to these threats we felt that it was necessary to view our entire project development effort to be divided among various phases which are sequential in nature. This is analogous to waterfall model of project development in software engineering parlance. The phases of our project are summarized in Figure 1-1.

Part 1 Overview

Page 4

Phase 1 (Independent) • In this phase We propose a case study to be conducted in an IT level under graduate course in our institute to improve upon the present methodology

Phase 5

Phase 2

• In this phase, we try materialize our working model which is manual till now, into a web module. This phase is meant to be a purely technical phase and could even be done outside the scope of our BTP.

• In this Phase we prepared a theoretical Model and underlining methodology aiming itself at improving upon present assessment techniques, by incorporating concepts of Fuzzy Logic.

Phase 3

Phase 4

• In this phase we try to implement the theoretical model from previous phase, manually. This is arguably the most exciting and challenging phase in our project.

• In this Phase we try to attach our working model [A deliverable from phase 4] to improve the status and proficiency of IT project teams.

Figure 1-1 – Phases in our Project

Part 1 Overview

Page 5

1.2.5.1 Phase 1 – Methodology Since Phase 1 was an independent stage and didn’t relate sequentially to other phases in the project, its mention here is irrelevant. For details about Phase 1 kindly refer Section 3.1

1.2.5.2 Phase 2 – Methodology In Phase 2 we prepared a mathematical model for modeling our concepts and methodology. The model used the definition of Fuzzy Logic and was created by integrating techniques which provide enhancement over present methodology. The details of the model can be found in Section 3.2. The critical features are mentioned below: A data structure we call Trait Vector. It maps the significance of a trait in user in floating point numbers in increasing fashion. Every user is initially assigned an empty trait vector [all values set to 0.00] to begin with. For the purpose of responding to our questionnaire, user draws graphs. User respond by drawing graphs, on a 2 dimensional space provided to them. The patterns they draw, populate the values in their trait vector and help in modeling the responses When a User is finished, we have a Trait Vectors ready corresponding to each respondent. We analyze this trait vector against 16 standard trait vectors, we have defined already. The analysis is done by two algorithms to resolve any scope of ambiguities and uncertainties which we may have arisen in the above process. Finally we show the user feedback in terms of individual traits. We also assign him/her a unique class which is one of the 16 MBTI classes with highest level of participation he/she showed during analysis. The underlining methodology on which our model is based in compared against the trivial methodology of personality assessment in Figure 3-2.

1.2.5.3 Phase 3 – Methodology Since our phases were sequential in nature, so once we completed Phase 2 we had a working model which was capable of assessing an individual’s personality in Fuzzy Based approach. In Phase 3 we took a step further and tried to implement our model and underlining methodology inherited from Phase 2 with real time industry level data. For the purpose of Data collection we conceived a whole new questionnaire according to our needs. The questionnaire was our primary source of data collection, on which we could understand and interpret the results of our model. The questionnaire is given in Appendix C. The questionnaire was based upon a new concept of Fuzzy Logic based graph plotting [7]. These helped us to model vagueness and improve results and findings. For the purpose of Data Collection we distributed over 70 questionnaires in 7 organizations including NTPC, HCL, Idea Cellular & American Express. We choose people of various domains ranging from IT, Insurance, Consultancy, Banking and Academia. The data collected was interpreted according to a scheme developed and was organized and analyzed using Microsoft Excel® 2007 as a tool. For details kindly refer to Section 3.3 Appendix

Part 1 Overview

Page 6

Personality Assessment Methodology - A Comparison Conventional Methodology

Our proposed Methodology

Step 1 A user Reads instructions and is assigned a value initialized to value set to 0.

User reads instructions and is assigned a Trait Vector which is initialized with all values set to 0.00.

Step 2 A user start to answer questions based on multiple choice format .

A user analyzes the question, understands it and starts to compile his/her response in form of a graph.

Step 3 For each question entertained the respondent's score keeps on getting populated.

With each graph drawn, a particular column (corresponding to trait) gets appropriate value.

Step 4 At the end of the questionnaire a score reflects the user's personality

At the end of the questionnaire we have all columns correponding to all the traits with definite values.

Step 5 Based on a specific implementations user is assigned to any 1 of 16 MBTI Classes.

2 algorithms analyze the Trait Vector and assign a user to all 16 MBTI classes with different levels of participation.

Step 6 User is shown his MBTI type mnemonic and some verbal description.

User is provided with a comprehensive analysis and comparitive feedback.

Figure 3-2– Comparison of trivial and our methodology for personality assessment

Part 1 Overview

Page 7

1.2.5.4 Phase 4 – Methodology We are not personality experts; neither are we researchers in the domain of psychology, human behavior and human personality. So it really doesn’t make any sense if we don’t apply our project findings to some real world application. Once we tested and consolidated our model in phase 3 we needed some real world application where these findings could be applied and results can be verified. For the purpose of which we chose the problem of assigning roles in a project team. This is a well known problem in project management domain and is usually solved with help of legacy personality inventories [5] we tried to solve this problem according to our personality inventory. Unfortunately this application needed much more time and knowledge and proved to be out of scope of our project, but still we very much look forward to do it in near future and further optimize our model and its workings.

1.2.5.5 Phase 5 – Methodology Though we rate our model and methodology as something evolutionary but still it is cumbersome to the very purpose it is designed for – the task of comparing individuals on the basis of personality traits. The inner working of the methodology relies on two algorithms which are tedious to be worked upon manually [i.e. with a pen and a paper and a physical form of questionnaire], therefore we decided to make an automated web module for catering to this limitation. [See Section 1.2.3 Scope of Use] The web module which we vision should be something which does all the work and compile results to provide feedback to the user. All the tasks from user logging in to know his/her personality type to providing detailed feedback, everything should be automated and user friendly.

1.2.6

Results and Conclusions

All throughout our project we tried to reinforce our approach. We build a model in Phase 2, and followed it by real industry level data collection. It took us nearly about 17 weeks to get sufficient data to derive important results. Though we had more data coming but we decided to compile whatever data we managed to get at the end of 17th week. The Gantt. Chart summarizing our project activities throughout these 18 months are given in Figure 1-3.

Figure 1-3 – Gantt. Chart

Part 1 Overview

Page 8

We had no hypothesis to propose and no theorems to prove. We always believed that with our proposed methodology we could seriously improve the scope of and application of personality inventories. The data we collected could be interpreted from many dimensions. We ourselves analyzed the data along some of the dimensions ( details of which are given in Section 4.1) mentioned below: Ideal Trait Vectors: With the data we had we could clustered Trait Vectors according to their standard MBTI type and then we took their average. We believe that this new averaged Trait Vector represents in close proximity the class it belongs which is one of the 16 standard types. Comparison: With the help of Trait Vectors we could no compare any two people along the lines of different traits. Statements like : A is 5% more Flexible than B B & C together are more Confident than D & C Can be inferred from data itself Clustering: Clustering a group of people was now easy; I can choose people based on their Trait Values to cluster them according to my preferences. Sub Classification: All the applications of MBTI could now be refined. Suppose we know that for a particular job we need 2 ESTJs and 3 INTJs, now with the help of the Data we could pick the best 2 ESTJs and best 3 INTJs. These are just some the dimensions we earlier talked about. With the personalities now in form of values we could even do complicated statistical analysis to discover more complicated results. Since we weren’t able to apply our model and methodology to some real time application because of lack of time, we leave it as an interesting and exciting opportunity for a future endeavors. As far as conclusion is concerned, our project concludes with following note: “We were able explore the realms of human personality, its assessment mechanisms and its applications successfully. We also successfully developed a new model and methodology over and above the present MBTI methodology to facilitate the task of personality assessment. Finally we successfully tested and checked the complexities of our methodology with real time industrial level data.”

1.2.7

Future Prospects

We believe that our theory since was a novel approach has great future prospects. Some of the people which it’s bound to impress are from among following domains: Academicians currently working in the discipline of Behavioral Sciences, psychology and human resource management. Industry level Managers having a portfolio under Human Resource management. Web developers, especially Web 2.0 developers and associated entrepreneurs Experts in the field of Data Mining and its applications Survey Industries for Marketing Research.

Part 1 Overview

Page 9

PART 2 ■ Project Planning

2.1 Project – Initiation 2.1.1

Project – Idea

In the last half decade, there have been tremendous advancements and deployments in two important technologies. One of them is the concept of “Fuzzy Logic” and other one is that of an “Online web portals”. While the former is a theoretical branch closely related to that of set theory and field of AI, the latter has invaded the internet to prove itself as the hub of millions of users online. Imagine the growing market of web portals. Today web portals come in variety of flavors; our particular concern is social community portals, Matrimonial portals and Business Networking portals. These portals have become very rapidly the cyber hub and pub for almost anyone who at least understands the basics of internet. Now realizing the fact that huge numbers of internet users are attached to one of more such portals, we could realize the concept of automated personality assessment. Let me try to explain the concept to its grave. Imagine, for exa mple a matrimonial portal, now it’s very important for these portals to provide to its customers the right kind of people to get in touch with. Today portals just put up the profile of an individual, as mentioned by them (with certain level of authentication) and leaves the mature decision of personality assessment to a user for her to interpret on her own. Now many a times a user can’t probe into potentially thousands and thousands of members of a portal. Therefore a need for an automated personality assessment module immediately arises as a consequence. Designing such a portal is a tedious and commendable achievement in itself, but what we’d like to bring to your attention in current context is that such a module is sure to utilize a fuzzy logic approach in finding personality trait of its potential users. The advantage with fuzzy logic is easy to realize because running a personality assessment test in form of an automated interview would have to handle intuition intensive opinions on part of a user. Now such a data has a huge scope for vagueness which we need to model accurately as the efficiency and accuracy of a module is in direct relation with the proficiency of such a modeling technique. Therefore such vagueness could be handled with incorporating the fundamentals of Fuzzy logic in our data collection technique and process of analyzing and translating that data at later stage. Once such a module is ready, we could easily integrate it with any web portal without much difficultly. We could ask users to run there personality assessment tests and then automatically dig out various users showing similar personality analysis against to a particular user. This way a person would be able to come in contact with people or group of people which in a good probability shows similar personality behavior and traits to her. This process is fully automated and without the intervention of user at any stage. A lot of further customization could be achieved at the time of deployment to grant a further level of customization. The broader scheme of our main idea is depicted in Figure 2-1

Project Planning

Page 10

Trivial MBTI Theory defined in terms of Personality Traits

Fuzzy Logic scope in Qustionnaring

A New fuzzy web based personality assessment system using enhanced methodology over trivial MBTI Theory

Web Deployment

Figure 2-1 Schematic diagram of our main Idea

2.1.2

Project – Scope

There are overly optimistic researchers in academic and corporate arena who think that personalities are good ways of clustering people. These clusters may represent a project team, particular group of customers who show similar taste for products or a bunch of people who show similarities in behavior and lifestyle. Clustering people according to some performance or preference criterion has always been an active area of research and is beneficial from a commercial point of view. Here we provide some of the market scenarios where our project could improve the commercial viability of personality inventories and reap greater benefits for various genres of business.

Web Portals : Refer to Section 2.1.1

Project Planning

Page 11

Marketing : Today marketing has become a complex job at hand. Specialized marketing firms rely on several factors for digging out marketing opportunities to provide its clients with the right kind of people they want. As per the current methodology these specialized marketing firms prefer to divide the customers ( population) into clusters based on information such as income, age, occupation, housing and race [ 12]. All of these parameters are important in their own sense but the list is far from complete. It suffers from the absence of a factor of vital significance – Personality. If only there exist a way to track down the personality of the individuals, these companies would make a swift effort to make it count for the purpose of marketing. Obviously such a thing doesn’t exist as of now, that is why this list is deprived of it, but we believe that with these project is capable of mediating this problem and make personality a critical parameter to make important marketing decisions and strategies.

Software Engineering : Today 66 % of the projects fail. [14] Some of the experts in the field of Software Engineering attribute such a high amount of failure to various factors like lack of user involvement, long or unrealistic time scales, poor or no requirements gathering, scope creep, no change control system, poor testing, but most believe that lack of team coordination and poorly assigned roles is one of the most significant ones. Some researchers even believe that roles reflect personality [3]. Since our project aims at improving the accuracy, scope of usage & applicability of personality inventories, it is bound to help in making proficient project teams. One may argue as to why the present methodology of personality assessment doesn’t have such a widespread scope in the areas mentioned above. This is so because the compare to traditional methodology our approach does benefit from definition of Fuzzy Logic and is more inclined and efficient in modeling the vagueness in ones thoughts, perception and intuition. These captures of an individual’s vagueness give our approach a competitive edge over the traditional methodology. Besides assessing an individual’s personality, our approach also scores high compare to present methodology when it comes to comparing two people on the basis of their personality. With the traditional methodology (MBTI Standard Type Indicator) one can only compare two individuals by the virtue of classes they belong to. Two individuals belonging to same class can’t be further compared. This is a serious limitation of MBTI Type, Since in a large population sample ( lets say 2000) we will have at least 125* individuals in a class and there is no way to compare these people further which is a limitation since these may have further differences on various parameters. Figure 2-2 on the next page summarizes the advantages our approach has over Existing Methodology

* Normall y we’ll ha ve more than 125 indi viduals in classes like ENTJ, ISTJ whi ch a re more common types than others . Project Planning

Page 12

Existing Methodology

Our Approach

Figure 2.2 Scope-comparison between existing methodology and our approach

2.1.3

Project – Feasibility

Feasibility Analysis is an important state project development lifecycle. By feasibility analysis one questions the availability and feasibility of issues like Resources, Time, finances, Human Resource and other important critical issues related with project development. It is fairly important to address all of the issues that may jeopardize and hamper the running of a successful project development lifecycle. Here we defined in tabular format what we thought were bottlenecks which could paralyze the progress of our project. We also analyzed the feasibility of our project (our approach) along the dimensions of basic parameters like Budget, Time, Resources, Domain level knowledge & Human Resource and then transformed our Approach into a more formal shape of Project Definition in next section. Figure 2-3 on next page summarizes the feasibility analysis of our project.

Project Planning

Page 13

Parameters

Feasibility

Consequences

Time

4 Months at maximum

•Web Module may not be feasible. •The Data Collection if any need be done as soon as possible. •Literature Survey shall not be more than 4 weeks at maximum.

Budget

Unknown at this stage

•Most of budget shall be concentrated on Data Collection Stage and Web Deployment Stage (If feasible)

Resources

Availaible - Basic

Human Resource

2 Academic Mentors 3 undergradute students.

Resources

•General Resources are availaible, like Databases access for literature survey, General Stationary etc. •Web Related Resources may be a problem

• Lack of an Industry level Alliance may hinder the scope of Data Collection & real time application.

•Availaible

Expertise

•Organizational Behavior •Mathematical Modeling

•Unavailaible •Advanced Web Devleopment

•Lack of an expertize in the realms of web devleopment may be a problem.

Figure 2-3 Figure summarizing the feasibility analysis of our Approach

2.1.4

Project – Definition

After doing the feasibility analysis of our Approach in the previous section we finally defined our project. The project Definition is stated under Project Definition “To do an exploratory research in understanding the scope and usage of personality inventories obtained using MBTI theory and further improve the applications of these inventories by improving the present personality assessment methodology with the help of Fuzzy based questionnaires”

Project Planning

Page 14

2.2 Literature Survey Once we defined our project definition our Project Initiation stage was over. The next thing was of course to do literature survey on the topic. Since literature survey is crucial and time consuming we thought of a novel approach to dig out literature of our need. We laid down some rules of thumb and certain guidelines to help us not get carried away and get over with the stage of Literature Survey in reasonable amount of time so that we can get things started. We chose to adhere ourselves to following guidelines: A maximum of 4 weeks duration. Maximum of 12 Base papers and minimum of 6 including all the topics and sub topics of interest. Figure 2-4 summarizes our options.

S No.

Topic

Minimum Required

Maximum Allowed

Final Count

1

Personality Assessment

1

4

3

2

Fuzzy Based Questionnaires

1

2

1

3

Applications of Personality Inventories

2

4

2

4

Project Teams

2

4

2

5

Web Assessment of Personality Assessment

0

1

1

Figure 2-4 Figure shows the choices we made during Literature Survey stage. Since there was a lot of material available on the topics of our interest, we decided to stick to certain well known Databases only. Our last choice was to stick to following four world known Databases from which we picked out our base papers. EBSCO® IEEExplore 2.7® Release 2.4 ABI/INFORM – ProQuest® ACM® Digital By Adhering to following guidelines we were really able to find out relevant and concrete literature for our project. Our final selections of papers were based on many trivial factors like Relevance, Complexity, Date of publishing etc. Our final inclusion in no specific order is mentioned in Figure 2-5.

Project Planning

Page 15

Rutherford, R.H. Using Personality Inventories to Help Form Teams in Software Engineering. Proceedings of the 6th annual conference on Innovation and technology in computer science education. 73-76 Jung, C.G. Psychological Types The Collected Works of C.G Jung, Harcourt Press, 1923 Karn, J.S. and Cowling, A.J. A Follow up Study of the Effect of Personality on the Performance of Software Engineering Teams. Proceedings of the 2006 ACM/IEEE International symposium on empirical Software Engineering. 232-241 Devito Da Cunha, A. and Greathead, D. Does Personality Matter? An Analysis of Code -Review. Communications of the ACM Vol. 50 (2007), 109-112 Gehring, D.R. Applying Traits Theory of Leadership to Project Management. Project Management Journal Volume 38, 44-54 Zuser, W. and Grechenig, T. Reflecting Skills and Personality Internally as Means for Team Performance Improvement. Proceedings of the 16th Conference on Software Engineering Education and Training (CSEET’03), 2003 Adrian, A.D. The potential for Fuzzy Logic Questionnaires in Management Studies. Systems, Man, and Cybernetics, 1998 IEEE International Conference, 2144-2149 Rutherfood, R.H. Using personality inventories to form Teams for Class Projects - A Case Study. Proceedings of the 7th conference on Information Technology Education (2006), 9-14 Choe, Bae, Kim and Lee. Work in Progress - The Study of Web-Based Adaptive Feedback based on the Analysis of Individual Differences. Frontiers in Education, 2004. FIE 2004. 34th Annual, (2004)

Figure 2-5 List of our base papers.

Project Planning

Page 16

2.3 Project Phases As per the description above, our project looked strenuous and demanding. There were various factors which were absolutely indispensible for this project to prove its worth and achieve its claims. Some of the most important issues we needed to sort out to ensure the healthy and productive handling of the project are the following: 1. Strong Understanding of basic fundamentals in disciplines like Behavioral Sciences, Basic Psychology & Organizational Behavior. 2. A Strong web development skill set, for development of online module of personality assessment. 3. A continuous access to good databases for the purpose of literature survey. 4. A good financial funding in case we need to invent primary data for our research. 5. Availability of good in-house experts in the fields mentioned in first point. These experts may include the members of academic community or of commercial community. 6. An extended project development time of around 6 – 8 months, which may even stretch up to a year if we try to consolidate our findings with good amount of primary data. There were other important issues as well, but they were of little significance when compared to the ones mentioned above. Out of all the issues 2nd, 4th & 6th are more serious in terms of threats they offered to our project development scenario. Due to these threats we felt that it was necessary to view our entire project development effort to be divided among various phases which are sequential in nature. This is analogous to waterfall model of project development in software engineering parlance. The phases of our project are represented in Figure 2-6. A summary of each phase is given underneath. Phase 1*: An Independent stage where we propose a semester long study to improve the process of project s distribution in an undergraduate course for an IT enabled course. Phase 2: Preparing a Model for the methodology aiming itself at improving upon present assessment techniques, by incorporating definition of Fuzzy Logic. Phase 3: Here, in this phase we try to implement the model from previous phase manually. This is arguably the most exciting and Challenging phase in our project lifecycle. Phase 4# Here, we try to attach our working model ( A deliverable from Phase 4 ) to improve the status and proficiency of IT project Teams. Phase 5# Here, we try materializing our working model [manual by now] into a web module. This phase is meant to be a purely technical phase and could even be done outside the scope of our BTP. Figure 2-6 Brief description of phases in our project All the phases mentioned are sequential in manner, However Phase 1 is independent. For the purpose of this summary short description of phases 2 and phase 3 are important. Figure 2-7 shows the phases in a sequential manner.

# These phases ma y be outside the s cope of our B Tech. Project beca use of thei r dema nding feasibility.

Project Planning

Page 17

Phase 1 (Independent) • In this phase We propose a case study to be conducted in an IT level under graduate course in our institute to improve upon the present methodology

Phase 5

Phase 2

• In this phase, we try materialize our working model which is manual till now, into a web module. This phase is meant to be a purely technical phase and could even be done outside the scope of our BTP.

• In this Phase we prepared a theoretical Model and underlining methodology aiming itself at improving upon present assessment techniques, by incorporating concepts of Fuzzy Logic.

Phase 3

Phase 4 • In this Phase we try to attach our working model [A deliverable from phase 4] to improve the status and proficiency of IT project teams.

• In this phase we try to implement the theoretical model from previous phase, manually. This is arguably the most exciting and challenging phase in our project.

Figure 2-7 – Phases in our Project

Project Planning

Page 18

PART 3 ■ Project Methodology After dividing the entire project into phases and defining them, we dealt with each phase one by one. Every phase is designed so that it produces certain deliverables which are useful in next phase, this sequential nature of phases were very helpful indeed as it always helped us know what is to be done when and how. In the foll owing section we discuss the specific methodology employed by us in each phase.

3.1 Phase 1 : A – Case Study Personality help make better project teams [3]. Owing to this fact we decided to design a semester long study in our institute to help allocate projects in an undergraduate IT enabled course on the basis of personality inventories. We planned the timely structure of events to be included in the study and took care of other important factors. Since this phase is a supplement and is not related to other phases or to central idea of our project we decided to mention its detail in Appendix B NOTE: This phase is independent i.e. It doesn’t match with the sequential nature of other phases mentioned in Section 2.3

3.2 Phase 2 : A Fuzzy Based Working Model Till this point our discussion was good. But this is from here onwards when we started building upon our approach. At this phase we were already through with perusing all the literature and we already had a project definition in hand. This is the phase where we started laying foundations of our working methodology. At this point we built a mathematical model and defined its inner workings. This section talks about our model. Since, we were looking to redefine things our way so we needed a consolidated theory at work. This model puts our theoretical concepts and ideas into actions through mathematical parlance. This model directly incorporates our approach and defines necessary structuring and architecture needed to achieve results by our approach. This section has following elements in order: Language – Defines and describes the details of the Language and terminology we use in our model Assessing – personality: An Extended Methodology – Discusses the main workings of our model. Assumptions and Conclusion – Discusses all the assumptions we carry from this Phase to next and concludes the model description.

3.2.1 Language Each of the 16 MBTI personality types is defined by certain personality traits which participate to different extents to finally make these 16 types distinguishable among themselves. Here in our study we short listed 10 personality traits to represent personalities. Next section contains details about each and every Language component.

Methodology

Page 19

3.2.1.1 SET T SET T contains Personality Objects. Each such Personality Object corresponds to a personality trait that we use in our model to represent personalities. The I th object corresponds to I th personality trait and there are in total 10 such Objects in the SET T. Each Object abstracts the following fields whi ch define a personality trait

SET T = { { T1, ID , T1, descri p tion }, ... , {Ti, ID , Ti, descri p tion }, {T16, ID , T16, descri p tion } } A trait ID like T1, ID .—An Integer A description or a small narrative T1, descri p tion to explain the trait itself – A String.

How did we shortlist Traits? As we defined SET T to contain personality traits, a genuine argument would be as to how do we got these traits? Obviously we didn’t shortlist on the consent of our intuition. It was really very important for us to include only those traits which are most relevant. In our literature survey we came across one of the publications titled “Applying Traits Theory of Leadership to Project Management” By DEAN R. GEHRING, PMP, Kennecott Utah Copper. [5].In this paper the author mentions about various MBTI Types and traits which are found in each MBTI Class. Figure 3-1 shows a snapshot from the paper where Author has mapped 16 MBTI Types against 13 personality traits, but not all of these traits are relevant in our context. We decided to pick the 10 most relevant traits based on our purpose of application.

Figure 3-1 – A mapping from MBTI Types to Personality Traits.

Methodology

Page 20

In our literature Survey we also came across another publication “Using Personality Inventories to Form Teams for Class Projects – A Case Study” By Rebecca H. Rutherfoord, Southern Polytechnic State University [8]. In this publication the author maps the roles in a project team against the standard MBTI Types. The snapshot is given in Figure 3-2

Figure 3-2 – A Table mapping Standard MBTI to Roles in a Project Team Now we combined the two mappings given in the two figures on the basis of their common scale – Standard MBTI Types – to come with a third mapping where Standard MBTI Types are mapped to Roles in a project team. The resultant mapping in shown in Figure 3-3* Now from the figure we can straightforwardly deduce that the traits ‘Impact & Influence’, ‘Interpersonal Understanding’ and ‘Concern for Order’ have least counts ( Trait Self-Confidence & Self-Control was addition from our side based on our own understanding). Hence we chose to ignore them on the basis of their relatively less participation in defining roles in a project team, which is our prime application in Phase 4.

* An a ddi tional trai t Sel f-Control & Self-Confidence was a n addition from Figure 3-1 hence i ts column is empty.

Methodology

Page 21

Figure 3-3 – A Table mapping Standard MBTI Types to Traits

Methodology

Page 22

3.2.1.2 Trait Vector Trait Vector(s) play very important role in our study and inner workings of our proposed methodology. If simply put together in words, a trait vector is nothing but a vector having floating point values (within certain lower and upper bounds) which shows some interpretations of a personality type. The value in Ith column of the vector corresponds to the Ith trait in SET T. The magnitude of the value signifies the level of participation of a trait in defining a personality. Therefore these trait vectors define personalities in a broken down fashion. We believed that by breaking a personality into it’s constituent traits we could gain a better understanding of it’s type and to model the same effect we introduced the concept of Trait Vectors. These trait vectors are also ways by which we directly deploy the concept of Fuzzy Logic in our methodology. Figure 3-4 Shows some Trait Vectors.

Figure 3-4 Some Trait Vectors. Now once we defined our personalities in terms of trait vectors we could compare two different vectors along the dimensions of various traits. The main question now was how we could populate such a trait vector? We mediated this problem in upcoming sections.

3.2.1.3 SET P Just like SET T, SET P is also a collection of objects. But this time objects correspond to 16 standard MBTI Types. This implies that SET P only contains 16 elements at max. Set P looks like SET P = { {P 1, ID , P 1, description , P 1, T.V }, ... , {P i, ID , P i, description , Pi, T.V }, {P 16, ID , P 16, description , P 16, T.V } } The Ith object corresponds to I th Standard MBTI Type and there are in total 16 such objects in the SET P. Each Object abstracts the following fields which define a Standard MBTI Type in detail. A Personality ID (P1, ID )—An Integer A description or a small narrative to explain the MBTI type itself (P 1, description )– A String. A Trait Vector corresponding to a standard MBTI Type (P 3, description ) – A Vector of Float Values

Methodology

Page 23

3.2.2 Assessing – personality: An Extended Methodology In this section we present the details of the methodology we incorporated for the purpose of assessing personalities. After understanding and making ourselves familiar and comfortable with the Language , we needed a mechanism to derive useful results from the concepts of Trait Vectors. Since our model is mathematical we incorporated the algorithmic approach and hence, defined 2 algorithms working at the heart of our methodology. The first Algorithm is mentioned under.

function Fuzzy-Personality-Comparison (TV, P) returns a structure* inputs: TV, TV of a participant as was populated in answering the questionnaire. P, Set P containing 16 Personality Objects. declare: d+ (vector), d- (vector), count (vector) initialize : d+ (vector), d- (vector), count (vector) – to zero value vectors for each P i in P : i => 1 : | P | ( cardinality of SET P ) repeat for each jth column in TV : j => 1: | T | ( cardinality of SET T ) repeat

Pij ) increment (count i, 1) d+ i += Mod (TVj - Pij)

if ( TVj >=

else

d- i += Mod (TVj - P ij) End-repeat end-repeat

return ( Resolve-Personality (d+, d-, count ) END * structure: it comprises of three things d+, d- and an identifier ( like ESTJ, ENTJ etc.)

Now one very natural question is: - How does algorithm Resolve-Personality work? In simple words Resolve-Personality analyzes the three data structures passed to it and then resolves any conflicts or ambiguities to return object corresponding to standard personality types and corresponding deviation. These deviations are of important significance as they correspond to partial participation of an individual ( to whom these deviations corresponds to) in a personality class. The working of Fuzzy-Personality-Comparison is shown is Figure 3-6. The working of Resolve-Personality is mentioned below.

Methodology

Page 24

Start

Algorithm takes A SET of Trait Vectors ( P ) and a Trait Vector as input

Then three parameters are declared and which are following: •d+: it’s a vector which stores the positive deviation of comparison between two trait vectors. Since in this algorithm we are comparing a single TV against a set of TV’s so we need a vector to correspond to each and every comparison.

•d-: it’s a vector which stores the negative deviations of comparison between two trait vectors. It works analogously to d+ vector. •count: it stores the number of traits whose magnitude overshot the corresponding value in another trait vector. Simply put together it tells us how many columns in one trait vectors had the values larger than their corresponding counterparts in another trait vectors .

These three parameters are then initialized to zero value vectors.

The TV under observation is checked against each and every trait vector in SET P. the comparison of two trait vectors always yielded three values a positive deviation, a negative deviation and a count. These were appropriately noted down in their respective data structures.

After the comparison is complete we’ll have our three data structures filled with floating point values. Now we pass these Data Structures to another Algorithm named Resolve Personality which returns a structure comprising of two floating point values corresponding to +, - deviations and identifier.

This value received from Resolve-Personality is simply returned.

Stop Figure 3-6 A Flowchart view of Algorithm Fuzzy-Personality-Comparison

Methodology

Page 25

firstly Resolve-Personality tries to find the count having maximum value; since a count corresponds to a comparison between respondent’s TV and any one of 16 standard TV ( and thus a standard personality type in SET P) it having a larger value would directly imply the extent of satisfaction a TV showed against any one of 16 Standard TV’s. ( Kindly go through the explanation on Trait Vectors earlier in the tex t ). If more than one such count exists (possibility of a tie) then the algorithm favors the one having higher +ve going deviation ( degree of closeness), if still a tie persists it resolves it by favoring the one with lower –ve deviation ( further degree of closeness). After such a process there is negligible scope of any further ambiguity remaining, but if there is then any such ambiguous count is chosen ( since all are equally likely to the extent of preciseness the system could allow). This way we get the best count corresponding to a personality type in SET P. The identifier of this personality type along with two deviations is then returned by the algorithm . We don’t emphasize upon the inner workings of algorithm Resolve-Personality yet because we believe that it may be implemented in lot many manners having different implementations. Implementation give above is just one many possible.

3.2.3 Assumptions and Conclusions In Phase 2 we built a mathematical model and designed certain algorithms to produce desired results. We also defined language which we are using in context of our project. Finally we showed how our algorithms works on data structures defined in the language to produce desired results. In Section 2.3 – Project Phases, we mentioned how we designed our phases in a sequential manner i.e. one phase produces certain deliverables for the upcoming phase. Here we conclude d our Phase 2, and mentioned the deliverables it provides for upcoming Phase 3.However, with the deliverables we also made some assumptions in this phase which will be resolved in further phases. Therefore besides giving certain deliverables this phase assumed certain things as well, more relevant and important of these assumptions are given below: Assumptions In Algorithm Fuzzy-Personality-Comparison, we assumed the availability of TV corresponding to a respondent. In Algorithm Fuzzy-Personality-Comparison, we assumed the availability of 16 standard Trait Vectors which makes up the SE T P.

As for the deliverables, this phase provided us with following deliverables for the upcoming Phases. Deliverables A Mathematical Model and a language to define personalities in the context of our Approach Algorithms to work on the language defined and produce unambiguous results 10 Standard traits on basis of which personalities will be defined and compared. This phase also made sure that Traits chosen are in coherence with the context of application taken up in Phase 4, which now may be somet hing like using personalities to assign roles in a project team.

Methodology

Page 26

3.3 Phase 3 : Industry level Data Collection After having a working model at our disposal it was crucial for us to put this model to test. The challenge was to feed the model with appropriate and sufficient data so as to see its workings. But since our model boasts to work on a new approach, it was highly unlikely for us to find the data which we can use in off the shelf manner. Therefore the immediate need of the hour was to make our own data. Once we ascertain to create our own data, we were confronted with the problem of how to collect data. Data collected should not only be reliable but sufficiently large and diverse as well. These added the dimension of quality to our data collection task. Finally we decided to weave together a questionnaire for the purpose.

3.3.1 Questionnaire We made a questionnaire for data collection. The Questionnaire is given in Appendix C. The questionnaire was based on concept presented in a paper titled “The Potential of Fuzzy Logic questionnaires in Management Studies” by Allyson D. Adrian, Georgetown School of Business, Georgetown University. Following are main features of our questionnaire. There are 2 sections with examples on usage given prior to the section. Section 1 contains 20 questions with 2 questions focusing on each trait. All the questions in Section 1 are Fuzzy Based Graph plotting type. Section 2 focused on assessing an individual on Standard MBTI Type with a trivial true/false type questionnaire. We also made sure that our questionnaire looked professional and interesting enough. To cater to this attributes we employed following techniques to make it look more professionally appealing and easy to respond to: Efforts put in to make the questionnaire look more appealing and professional. We put a sticker on sheet to hide the staple pins. We put our institute stamp to attach a sense of authenticity. We also put a serial number on the front page to assist the distribution and handling. We provided sample instructions to facilitate the understanding of the questionnaire We use a low contrast theme to make it more visually appealing.

Methodology

Page 27

Stamp and our Institute’s Logo Efforts put in to make questionnaire easy to use. We provided ample instructions and examples wherever needed to facilitate the understanding of the underlying concept behind the questionnaire. We encouraged the users to engage in an online version of Standard MBTI test instead of responding to our questionnaire because it provided less work and instant results. We kept an option to skip entire section 2 if one already knew about his/her standard MBTI type . For Section 2, we relied on the questions given on humanmetrics.com []. However to make questions more appropriate for Indian audience we tweaked the language of questions but retained the context. Besides these things we made certain strategies regarding distribution of the questionnaire to companies, since it was vital so as to convince the person in-charge of the essence and worth of the whole exercise he/she will be engaging his employees in for the sake of our project. Our main policies in this regard were: We decided to take a strong recommendation signed by both are mentors and printed on our institute’s letterhead. We decided upon not to distribute more than 15 questionnaires in a particular organization. We decided to bring this to the notice of managers of the organization where we took our questionnaires that we encouraged respondents to fill the questionnaire at their spare time, thus asking him/her to distribute it further to employees and asking them to fill it at their homes . Thus saving a good amount of company time. We decided upon to deliver the questionnaires preferably on Friday nights as we understood that employees would be relaxed during weekends and can engage their time filling up the questionnaire. We made it a policy to contact the person in-charge of an organization once every 3 days to press him in cooperating well. We decided to only contact organization in Gwalior and Delhi, since going elsewhere was expensive. There were also issues regarding what companies to distribute our questionnaire to, since usually the human resource differs in different sectors. Following were the sectors we thought we could entertain: IT Sector Insurance Sector Banking Sector Academic Sector Consultancy

Why did we have 2 sections in our questionnaire? Since we mentioned in Section 3.2.2 – Assumptions and Conclusions that one of the assumptions that we need to mediate in this phase is to define the standard trait vectors for 16 standard MBTI types, therefore we felt a need to also know the standard MBTI type of a respondent whose TV we just calculated. By knowing the standard MBTI type we could anytime group TV’s having same MBTI type and take their mean (or some other Statistical tendency) to reflect the TV corresponding to the MBTI type. Hence we always asked respondents to fill in their standard MBTI types or engage in Section 2 because it helped us consolidate our theory and subdue the assumption.

3.3.2 Data and Interpretation Our questionnaire which we designed for data collection was fuzzy based i.e. it didn’t use the trivial unary choice scale that questionnaires normally operate upon. Since our questionnaire was different, we needed an all together different scheme to interpret the responses. Users used to draw graphs based on their understanding but it was still not very clear (though we had an Intuition) to us as to how we can interpret these graphs to fill in trait vectors.

Methodology

Page 28

Interpretation Scheme– From Graphs to Trait Values Our major concern at this stage was to somehow convert graphical responses into floating values which can be filled up in corresponding column in Trait Vector and ultimately we get a trait vector corresponding to each respondent based on patterns of graph he/she drawn. Our major concern at this stage was to somehow convert graphical responses into floating values which can be filled up in a corresponding column in Trait Vector and ultimately we get a trait vector corresponding to each respondent based on patterns of graph he/she drawn. For the sake of which we developed an Interpretation scheme. Figure 3. 7 shows the steps involved in obtaining a trait vector corresponding to a user from his/her responses in the questionnaire. Next section exemplifies these steps on a scanned section of an actual questionnaire.

For Section 1 of Each Questinnaire •After obtaining a questionnaire we note down the Y values corresponding to each of 7 metrics given along X-axis with the help of a pencil. We repeated this activity for all 20 graphs in Section 1

For All Graphs in Section 1 •Then corresponding to each graph, we labeled 3 extreme left values (in order from left to right ) as N 2.5, N 1.66, N 0.84 respectively. •Similarly, corresponding to each graph, we labeled 3 extreme right values ( in order from right to left ) as P 2.5 , P 1.66, P0.84 respectively. •We always ignored the middle value.

For both graphs of each Trait we defined following things: •Net -ve, 1 as = - ( 2.5 x N 2.5 + 1.66 x N 1.66 + 0.84 x N 0.84 ) / 10 •Similarly, Net -ve, 2 for second graph of same Trait. •Net + ve, 1 as = ( 2.5 x P 2.5 + 1.66 x P 1.66 + 0.84 x P0.84 ) / 10 •Similarly, Net + ve, 2 for second graph of same Trait.

For Each Trait •Total -ve = (Net -ve, 1 + Net -ve, 2 ) /2 , for each Trait •Total + ve = (Net + ve, 1 + Net + ve, 2 ) /2 for each Trait •Total = Total -ve + Total + ve for each Trait

For All Traits in Section 1 •We now have a value Total corresponding to each and every trait in Section 1, we simply stored this value in appropriate column of Trait Vector corresponding to the current respondent.

Figure 3.7 Interpretation Scheme. Converting Graph patterns to Trait Values

Methodology

Page 29

Interpretation Scheme.—Exemplified Now we’ll explain the Interpretation Scheme given in Figure 3.6. We provide a scanned section containing a graph of one of our questionnaires in Figure 3.8.

Figure 3.8 A scanned section of a graph from one of our questionnaires. Step 1st We marked all the Y values corresponding to each of the 7 metrics along X-axis.

Step 2nd For the graph above we computed N values and P values. N 2.5 N 1.66 N0.84

Methodology

0.06 0.90 1.00

P 2.5 P 1.66 P0.84

0.25 0.49 0.57

Page 30

We stored these values in an our Excel sheet

Step 3rd For the graph above we have

Net - Ve

-( 2.5 x 0.06 + 1.66 x 0.90 + 0.84 x 1.00 ) / 10

- 0.38

Net + Ve

-( 2.5 x 0.25 + 1.66 x 0.49 + 0.84 x 0.57) / 10

0.19

These values gets calculated in our excel sheet.

Step 4th For the graph above we have

Total - Ve Total + Ve

( -0.38 + -0.14)/2

-0.26

(0.19 + 0.17)/2

0.18

Total

(-0.26 +0.18 + 0.50)

0.42

These values gets calculated in our excel sheet.

Step 5th Now we stored this value in to the Trait Vector corresponding to this User. In our Excel file Trait Vector looks like one below

Our major concern at this stage was to somehow convert graphical responses into floating values which can be filled up in corresponding column in Trait Vector and ultimately we get a trait vector corresponding to each respondent based on patterns of graph he/she drawn.

Significance of (2.5, 1.66, 0.84) The metrics on the X axis (like –Never, Rare etc.) reflect the intensities in decisions, for e.g. Never is more severe negation than Very rare which in turn is more severe than rare. Therefore their participation in overall

Methodology

Page 31

significance will vary, which is modeled by assigning weights to these metrics. The more severe a metric the more weight it has, it’s a concept of weighted mean compare to normal mean which is just average of values.

3.3.3 Assumptions and Conclusions In this phase we put our model to work. We collected data from organizations through a custom made questionnaire which was based on fuzzy based graph plotting. After collecting the data we made an Excel sheet to store all the data and do necessary computation on it to discover new information. Finally we got our Trait vectors ready corresponding to each respondent. We also kept track of standard MBTI type of each individual so as to map MBTI types Trait Vectors so as to obtain 16 standards Trait Vector. All in all the data we collected had following configuration.

Number of Questionnaires distribut ed

70

Number of questionnaires evaluated

51

Number of organizations taking part in Survey

8

The list of major companies that took participation in Survey is: Organizations

Type

Number of participants

American Express

Bank

7

NTPC

Thermal Power Generation

5

Netgables Pvt. Limited

IT Solutions

6

Idea Cellular

Cellular Services

5

HCL Infotech

IT Services

6

Unicon*

Investment Solutions

10

This is all about Phase 3 of our project. This is was the most strenuous of exercise that we have to do in our project. It took us nearly 2 months to capture all the data, compile it and then interpret it to derive results which act as deliverables for next Phase. Though the data we collected was less but still it was good, genuine and complete. As high as 95 % of data was complete this meant missing graphs Section 1 and missing selections in section 2. With the help of data we collected we mediated both the assumptions we mentioned in Section 3.2.3. The main deliverable from this phase was the model which we introduced in section 3.2 with initialized language to put it to work i.e. now we had some data to match any new coming entry against. Though still this initialization is weak because paucity in data but still it’s enough to roll things on. This was the last Phase that we completed in our Project. The upcoming phases are not yet implemented and the description is purely theoretical. * These companies ha ven’t returned copies a t the time of compilation of this report.

Methodology

Page 32

3.4 Phase 4 : A Real world Application (Theory) We are not personality experts; neither are we researchers in the domain of psychology, human behavior and human personality. So it really doesn’t make any sense if we don’t apply our project findings to some real world application. Once we tested and consolidated our model in phase 3 we needed some real world application where these findings could be applied and results can be verified. For the purpose of which we chose the problem of assigning roles in a project team. This is a well known problem in project management domain and is usually solved with help of legacy personality inventories. We tried to solve this problem according to our personality inventory. Unfortunately this application needed much more time and knowledge and proved to be out of scope of our project, but still we very much look forward to do it in near future and further optimize our model and its workings.

3.5 Phase 5 : Assumptions and Conclusions (Theory) Though we rate our model and methodology as something evolutionary but still it is cumbersome to the very purpose it is designed for – the task of comparing individuals on the basis of personality traits. The inner working of the methodology relies on two algorithms which are tedious to be worked upon manually [i.e. with a pen and a paper and a physical form of questionnaire], therefore we decided to make an automated web module for catering to this limitation. [See Section 1.2.3 Scope of Use] The web module which we vision should be something which does all the work and compile results to provide feedback to the user. All the tasks from user logging in to know his/her personality type to providing detailed feedback, everything should be automated and user friendly. Fig C describes the broad level functioning our web—module , since it’s a whole new project developing such a state of an art web module so we left it outside the scope of our project and as possible future prospect.

Methodology

Page 33

PART 4 ■ Results and Conclusion Phase 1: A – Case Study All throughout our project we tried to reinforce our approach. We build a model in Phase 2, and followed it by real industry level data collection. It took us nearly about 17 weeks to get sufficient data to derive important results. Though we had more data coming but we decided to compile whatever data we managed to get at the end of 17th week. The Gantt. Chart summarizing our project activities throughout these 18 months are given in Figure 1-3. of Section 1.2.6. The distribution of MBTI types in our data was interesting; the final distribution is summarized in Figure 4-1 below:

MBTI Distribution 7

No. of Respondents

6 5 4 3

2 1 0 ISTJ ISTP ESTP ESTJ ISFJ ISFP ESFP ESFJ INFJ INFP ENFP ENFJ INTJ INTP ENTP ENTJ

Figure 4-1– Standard MBTI distribution on data collected. Right from the start of our project, we focused on our theory, neglecting the end results which we might achieve. We concentrated on the theoretical part of the theory in Phase 2 and then on its practical viability by collecting data in Phase3. We planned the application in Phase 4 and Phase 5 but couldn’t because of lack of time and resources. Since we were not able to apply our theory to any real world application (Phase 4), we didn’t have any tangible results. However Phase 3 did provide us with numerous dimensions on which could interpret our data We ourselves analyzed the data along some of these dimensions to discover some facts which were earlier impossible to even realize. These interesting patterns are enumerated below pictorially: Trait wise distribution of MBTI classes i.e. showing relative participation of a Trait in a standard MBTI class. We show this distribution for traits, “Initiative” and “Flexibility”.

Results and Conclusion

Page 34

Figure 4-2– Trait 1 distribution on Standard MBTI types

Figure 4-3– Trait 6 distribution on Standard MBTI types Simultaneous comparison between more than 2 respondents belonging to any MBTI classes on the basis of any number of traits. To exemplify in Figure 4-4 we compared three of our professors who took part in the survey. 0.85 0.80 0.75 0.70 Dr. Sahu

0.65

Dr. Dhar 0.60

Dr. Patwardhan

0.55 0.50 0.45 Trait 1

Trait 2

Trait 3

Trait 4

Trait 5

Trait 6

Trait 7

Trait 8

Trait 9

Trait 10

Figure 4-4– Comparison of respondents belonging to any MBTI type

Results and Conclusion

Page 35

Simultaneous comparison of several MBTI classes based on their Ideal trait vectors which are obtained by averaging the Trait Vectors of their respective occupants 0.9 0.85 0.8 0.75

INFJ

0.7

ESTJ

0.65

ENFJ

0.6 0.55 0.5 Trait1

Trait2

Trait3

Trait4

Trait5

Trait6

Trait7

Trait8

Trait9

Trait10

Figure 4-5– Comparison of MBTI classes through their respective Trait Vectors Beside the ones mention above there are plenty of more dimensions available. This is made possible by transforming personalities into values which can be easily compared and analyzed compare to verbal descriptions. As far as conclusion is concerned, our project concludes with following note: “We were able explore the realms of human personality, its assessment mechanisms and its applications successfully. We also successfully developed a new model and methodology over and above the present MBTI methodology to facilitate the task of personality assessment. Finally we successfully tested and checked the complexities of our methodology with real time industrial level data.”

Results and Conclusion

Page 36

PART 5 ■ Future Prospects

We believe that our theory has huge future prospects since it’s a novel approach in a widely known area of research. There are some areas where it (our theory) can apply itself directly to produce great results. Our theory may appeal more to specific group of people below: Academicians currently working in the discipline of Behavioral Sciences, psychology and human resource management. Industry level Managers having a portfolio under Human Resource management. Web developers, especially Web 2.0 developers and associated entrepreneurs Experts in the field of Data Mining and its applications Survey Industries for Marketing Research. As far our project is concerned it has tremendous scope of future work. There are a lot of areas where an individual can contribute some of the possible lines of improvement are mentioned below: Efforts can be made to consolidate the theory itself. More results can be obtained by various statistical inference mechanisms on the data. Improvement in the model by enhancing the Algorithms and theory. Efforts to standardize the traits based on the domain of use and study. Efforts can be made to automate the methodology with the help of concepts in the realms of Web 2.0

Results and Conclusion

Page 37

References [1]

Rutherford, R.H. Using Personality Inventories to Help Form Teams in Software Engineering. Proceedings of the 6th annual conference on Innovation and technology in computer science education. 73-76

[2]

Jung, C.G. Psychological Types The Collected Works of C.G Jung, Harcourt Press, 1923

[3]

Karn, J.S. and Cowling, A.J. A Follow up Study of the Effect of Personality on the Performance of Software Engineering Teams. Proceedings of the 2006 ACM/IEEE International symposium on empirical Software Engineering. 232-241

[4]

Devito Da Cunha, A. and Greathead, D. Does Personality Matter? An Analysis of Code-Review. Communications of the ACM Vol. 50 (2007), 109-112

[5]

Gehring, D.R. Applying Traits Theory of Leadership to Project Management. Project Management Journal Volume 38, 44-54

[6]

Zuser, W. and Grechenig, T. Reflecting Skills and Personality Internally as Means for Team Performance Improvement. Proceedings of the 16th Conference on Software Engineering Education and Training (CSEET’03), 2003

[7]

Adrian, A.D. The potential for Fuzzy Logic Questionnaires in Management Studies. Systems, Man, and Cybernetics, 1998 IEEE International Conference, 2144-2149

[8]

Rutherfood, R.H. Using personality inventories to form Teams for Class Projects - A Case Study. Proceedings of the 7th conference on Information Technology Education (2006), 9-14

[9]

Choe, Bae, Kim and Lee. Work in Progress - The Study of Web-Based Adaptive Feedback based on the Analysis of Individual Differences. Frontiers in Education, 2004. FIE 2004. 34th Annual, (2004)

[10]

Keirsey, D. Please Understand Me II. Prometheus Nemesis Book Company, (1998)

[11]

The Minnesota Multiphasic Personality Inventory http://en.wikipedia.org/wiki/Minnesota_Multiphasic_Personality_Inventory

[12]

Alex Berson, Stephen Smith, and Kurt Thearling, Building CRM Applications

[13]

Whitten, Bentley, Dittman. Systems Analysis & Design Methods, 6th ed. Reading, TATA McGRAW-HILL, 2003.

[14]

Lewis, Bob, Survival Guide, http://www.infoworld.com/articles/op/xml/01/10/29/011029opsurvival.html, 2004.

Appendix A List of Figures Caption

Page No.

Figure 1-1 Figure 1-2 Figure 1-3 Figure 2-1 Figure 2.2 Figure 2-3 Figure 2-4 Figure 2-5 Figure 2-6 Figure 2-7 Figure 3-1 Figure 3-2 Figure 3-3 Figure 3-4 Figure 3-6 Figure 3-7 Figure 3-8 Figure 4-1 Figure 4-2 Figure 4-3 Figure 4-4 Figure 4-5

5 7 8 10 13 14 15 16 17 18 20 21 22 23 24 29 30 34 35 35 35 36

Title Phases in our project Comparison of trivial and our methodology for personality assessment Gantt Chart showing project activities Schematic diagram of our main Idea Scope-comparison between existing methodology and our approach Figure summarizing the feasibility analysis of our Approach Figure shows the choices we made during Literature Survey stage List of our base papers Brief description of phases in our project Phase descriptions in our project A mapping from MBTI Types to Personality Traits. A Table mapping Standard MBTI to Roles in a Project Team A Table mapping Standard MBTI Types to Traits Some Trait Vectors. Flowchart view of Algorithm Fuzzy-Personality-Comparison Interpretation Scheme. Converting Graph patterns to Trait Values A scanned section of a graph from one of our questionnaires Standard MBTI distribution on data collected Trait 1 distribution on Standard MBTI types Trait 6 distribution on Standard MBTI types Comparison of respondents belonging to any MBTI type Comparison of MBTI classes through their respective Trait Vectors

Appendix B Phase 1 Methodology

Role of Personality Inventories in Distribution of University Level IT Projects and Team Formation - A Case Study

Abstract It is now long known that Personality plays an important role in Team Playing. To understand the ex tent to which personality affects the work quality and work flow is still a very active area of research in management community. A lot of literature is being written and a lot of case studies are being performed just to consolidate this idea. Based on some of the work which has already being done in this regard we ourselves thought of implementing a semester long study at our institute to understand, validate and further propose important concepts, ideas and conclusions in this area .

General Terms

Methodology

Management Project – Teams

Keywords Personality, Teams,

Personality—Inventory,

Normally Teacher introduces the concept of a Project and then explains various details regarding the projects. However, most faculties in charge feel it more comfortable to leave it onto students to come up with a team and possible team structure. This way some students take all the benefit and others get deprived of the very essence th e project was suppose to grant. In our study we try to consolidate various ways to alleviate this problem. We propose a methodology with a timely schedule which could be followed in a semester long IT course to study the effects of this problem and finding possible solutions based on the general theory of personality inventories.

Project

Introduction We as a students of Information Technology and being a student of IT, we thoroughly understand the important and vital role of active projects which are part of our curriculum. Not only these projects provide active and practical ways of learning and understanding concepts, but they also prove helpful in mediating the gap from theory in books to active work in a corporate. No wonder that every student especially in the discipline of IT should look forward to these projects as they provide very rich grounds for learning and personality building. However, one of the problems, which we ourselves realized having been engaged in quite a few of such projects are ‘Assignments of Role in a Team’.

The methodology is simple and straightforward. We first plan to decide upon a course (preferably an IT course) and then plan to contact the faculty in charge regarding the course. We then look forward to explaining her various events and tasks we’ll be following to ensure that project assignment is done in accordance with criteria we claim are essential for providing a healthier learning environment. Having decided our course which is let say “Course X”, we (our team and faculty in charge) would then be introducing the project to the batch under consideration for study. Once the project is made fully understood to the students, the students themselves will be asked to submit teams on their own preferenc e. For the sake of convenience the projects assigned shall be similar to the maximum extent in their requirements, specification and structure and the number of participants allowable per team should be constant and should not be less than 5. These two constraints are meant to bring certain amount of uniformity and coherence in various stages of project development and end project deliverables. These requirements will also help in evaluating the end deliverable with a single common perspective.

Having taken care of things mentioned above, we shall have a list of teams provided to us by students on their own understanding of the project requirements and self preference.

The project activity plan is given below in the table to better vision the timely relevance of various events and activities crucial in the project.

Next task is to help each and every student assess his/her personality on a standard MBTI scale. Then we plan to split the entire batch into three groups. Let’s for now call these Groups GA, GB, GC. The characteristics and dynamics of each of these groups are mentioned below.

After the semester we shall have collected some interesting observations. Now we need to interpret these observations to conclude certain critical results. The guidelines to interpret are mentioned in the nex t section.

Group

Size*

Characteristics / Dynamics

GA

X/3

This group contains the teams specified by the students on their own preference.

X/3

This group contains the teams which are made sure to possess the participants of same personality types. E.g. A team of all EST Js

Based on the final assessment of teams, one of the groups would eventually have the highest aggregate then the other two. By probing the results in the light of various project planning aspects we could derive following critical conclusions:-

X/3

This group contains teams where each and every role is carefully allotted as per the personality to role assignment criterion.

Observations and Conclusion

GB

GC

* Size is the number of teams allotted per group. # X is the total number of teams in the class.

Now having understood the dynamics of various groups under study let us focus on the scheduling on how to put up such a study in a methodical manner.

The Study – A Timely Plan Up till now we learned how to structure the requirements of our study. First of all the project was explained to all the students eradicating any doubts whatsoever regarding a smooth operation. Next groups wer e carefully formed according to dynamics mentioned in the table above. We also made each and every student take a standard MBTI test to know about an individual’s personality type. Then groups were formed and the knowledge was made available to students themselves.

If group GA has the highest score this means that, teams which were based on self preferenc e which impli es natural compatibility are most successful. However the possibility of such a things is highly unlikely as it is a proven fact that such teams hardly work their way up the quality assurance. If group GA has the highest score this would essentially mean that when the members of a team are homogeneous in their persona, they yield better results. However according to current management concepts this also has a little likelihood. It is the group GC which according to us has the maximum chance of showing the most promising results. In group GC the roles were assigned based on compatibility criteria which map a personality type to a certain role in a project team. Therefor e these teams are most likely to be compatible and thus successful.

The conclusion of this study was to strengthen the belief that heterogeneous teams where the roles are differentiated on the basis of personality types produce better results then those teams which are made randomly or without any effort. Another strong motive of this study was to propose a methodology which could be followed by faculty members to assign projects in an academic environment. It should be clear by the end of this study that a team where roles are well distributed based on MBTI compatibility types provide a rich learning experience than where no such effort is made. We also suggest that such a methodology should be applied to various academic teams other than IT teams in particular.

A Word of Caution We strongly feel that such a study though is very effec tive for any student but at the same time could also lead to dissatisfaction among many especially among those teams which were shuffled by us and allotted to group GB, GC. Some students could even complain on the circumstances being unfair to them since they were put in a team where they don’t feel uncomfortable. To alleviate all these problems we’ll have to grant certain incentives to teams of group GB, and GC, but have to make sure this incentive is not too much.

References Rutherford, R.H. Using Personality Inventories to Help Form Teams in Software Engineering. Proceedings of the 6th a nnua l conference on Innovation and technology in computer science education. 73-76 Jung, C.G. Psychological Types The Collected Works of C.G Jung, Harcourt Press, 1923

Keirsey,D.,”KeirseyTempermentWebsite”, Internet. http://keirsey.com, 2006

Online.

Keirsey, D., and Bates, M., Please Understand Me, Del Mar, California: Prometheus Book Company, 1984 KeirseyTests, http://honolulu.hawaii.edu/intranet/committees/FacDe vCom/guidebk/teachtip/kerisey.htm, 2006

Appendix C Questionnaire

Appendix D Snapshots of Excel Sheets

Details of the Respondents

Evaluation of Trait Values

Trait Vectors corresponding to each Respondent

Summary of Trait Values for each MBTI class.

Survey Declarations

Letters From Companies

Glossary

Important Terms

Term

Description

MBTI

A methodology and an instrument for identifying an individual's personality type based on Carl Jung's theory of personality preferences.

Personality Inventory

Fuzzy Logic

Exploratory Research

HRM

Software Engineering

Data Structure

Gantt. Chart

An assessment tool used to determine which of these personality types a person falls into: extroverted, introverted, thinking, feeling, sensing, intuitive, judging, and perceptive. It is used as part of a self assessment done for career planning purposes. A system of logic dealing with the concept of partial truth with values ranging between “completely true” and “completely false Exploratory research is a type of research conducted because a problem has not been clearly defined. Exploratory research helps determine the best research design, data collection method and selection of subjects. Human resource management (HRM) is the strategic and coherent approach to the management of an organization's most valued assets - the people working there who individually and collectively contribute to the achievement of the objectives of the business Software engineering is the application of a systematic, disciplined, quantif iable approach to the development, operation, and maintenance of software. In computer science, a data structure is a way of storing data in a computer so that it can be used efficiently. Often a carefully chosen data structure will allow the most efficient algorithm to be used. A Gantt chart is a popular type of bar chart that illustrates a project schedule. Gantt charts illustrate the start and finish dates of the terminal elements and summary elements of a project

Web Portals

A web portal is a site that functions as a point of access to information on the World Wide Web. Portals present information from diverse sources in a unified way.

IEEExplore 2.7® Release 2.4

Provides access to full-text articles from all journals published by IEEE and IEE (The Institution of Electrical Engineers)

ACM Digital®

Algorithm

Group of 54000 on-line articles from 30 journals and 900 proceedings of the Association for Computing Machinery. In mathematics, computing, linguistics, and related disciplines, an algorithm is a finite list of well-defined instructions for accomplishing some task that, given an initial state, will proceed through a well-defined series of successive states, possibly eventually terminating in an endstate.

ABI/INFORM – ProQuest®

One of the world's first electronic databases, a premier source of business information for more than 30 years.

EBSCO®

An electronic journals service available to both academic and corporate subscribers. It aggregates access to electronic journals from various publishers.