Research Methods for Engineers Griffith School of Engineering

Research Methods for Engineers Griffith School of Engineering Lecture 2: Research Design Prof. Sherif Mohamed 2014

What is Good Research? ü Review R ü Rigorous R ü Results R 1

The Research Process q Major components: Research Question Research Design Sampling Design Pilot Testing Data Collection Analysis Reporting Results 2

The Research Process Theory, Literature Review Practical problem Research Question Data collection Exp. , survey, analytical, numeical Solution/Answer Data analysis Results Conclusions 3

Research Design q Blueprint for meeting objectives q The logic that links the data to be collected (and the conclusions to be drawn) to the initial research question. q A research design has two purposes: to help us get answers to research questions to control variance 4

Interactive Model of Research Design EXISTING THEORY PERCEIVED PROBLEM CONCEPTUAL CONTEXT PURPOSE RESEARCH QUESTION METHODS PERSONAL EXPERIENCE DATA & CONCLUSION VALIDITY 5

Research Design q Type of data source Experiment Simulation Survey, q etc Methods and techniques Numerical simulation, e. g. , Finite element analysis Analytical analysis, e. g. , Regression analysis Physical modelling, e. g. , laboratory or field experiments etc. 6

Definitions q Theoretical (Conceptual) Framework q Independent / Dependent Variables q Relationships q Hypothesis q Theory q Model q Empirical Research 7

Theoretical Framework q Following problem definition, you need to establish a framework that represents the relationships among VARIABLES – often presented in the form of a schematic diagram. X Y Do money, career and lifestyle explain the difference in happiness among middle-aged population? Happiness Z 8

Variables A variable is an object, event, or process that can take on or change to different values or qualities, whereas a constant is an object, event, or process that has only one fixed value (Evans, 1985). q A symbol to which numerals or values are assigned. q Independent Variable (IV) Potentially causes a change in the other variable IV causes the DV to occur q Dependent Variable (DV) A variable in which we expect to observe a change q 9

Examples of Variables Variable Name Levels (types, values) q Gender Male or Female q Test Scores #, correct q Room Temp. Hot, Warm, Cold q Room Temp. Degrees, Fahrenheit Nominal Variable (different groups M/F with no numeric properties) Quantitative Variable (amounts or quantities) 10

Relationships Research Questions – establish relationships Does having a smart boss improve productivity? Is gender related to happiness? Do M/Females differ in happiness? Is there a direct relationship between cost and quality of a product? Does exposing this material to extreme cold temperatures reduce its tensile strength? 11

Relationships between Variables IV DV Intensity of traffic will effect your mood whilst driving Mechanized production will be less expensive than manual production. Attention level in students is a function of a lecture duration. Do students learn better when the lecture is short & sweet? ! 12

Let’s Identify Variables D Bolt under loading x 1 D 45 L x 2 Side view Tensile stresses Potential cracks Face view 13

Mediator Variables IV Variables Mediator DV that affect how the IV influences the DV Management aims at evaluating all of our projects by the end of this week (high goal) and this resulted in me working overtime (persistence) and producing a high number of financial reports (performance). The No-Overtime payment (rewards) resulted in bad feelings about my job (satisfaction) and an increasing frustration towards my firm which seems to value its active employees so little 14 (commitment).

Relationships IV DV Positive Relationship Increase in IV Increase in DV Decrease in IV Decrease in DV Negative Relationship Increase in IV No Relationship Decrease in DV 15

Increases in the values of one variable are associated with both increases and decreases of the second variable. Anxiety (or pressure or stress) can help increase performance until it gets too high – then it decreases performance. Performance low Curvilinear Relationship high Relationships low Anxiety high 16

Studying Relationships Experimental method Addresses the problems of inferring cause and effect Manipulate the first variable and observe the effect on the second variable Control for all other variables 17

Effect of IV on the DV q Does distraction cause poor performance on an exam? q Independent Variable “cause” variable Manipulated first q Dependent Variable “effect” variable Measured as a response to the IV 18

Studying Relationships Non-experimental method Behavior is observed as it naturally occurs Variables are measured only (not manipulated) q Direction of cause and effect Does X cause Y? or Does Y cause X? Trust & Business q Uncontrolled third variables Perhaps Z causes both X and Y If so X and Y will appear to be related 19

Controlling other Variables q “Other” = Intervening q Experimental Control Variables are kept constant q Randomization Randomly assign participants to conditions Run the conditions of the study in a random order q Some other variable could not cause the relationship between X and Y. 20

Propositions and Hypotheses q Proposition: statement about concepts that may be true or false. q When a proposition is formulated for testing, it is a hypothesis. q Descriptive hypothesis: state the existence, size, form or distribution of some variables. the current unemployment rate in Australia is 4%. q Relational hypothesis: State relationships between variables. The greater the stress, the lower job satisfaction 21

Role of Hypothesis 1. Guides direction of study. 2. Testable analysis (to accept or reject) 3. Enables clear conclusion to be drawn. Husbands and wives agree in their perceptions of their respective roles in property purchase decisions. Who to study? q What context? q What to study? q 22

Research Question vs. Hypothesis q RQ is posed when there is not enough evidence on the basis of the literature reviewed to predict the nature of the relationship. Does having a smart boss improve productivity? q Hypothesis is a tentative statement predicting the relationship of the variables. Smart bosses represent a crucial ingredient to improve productivity. 23

Hypothesis q One-tailed Hypothesis Predicts a specific relationship between variables e. g. , PG students socially interacts more than UG students do q Two-tailed Predicts Hypothesis a specific relationship without specifying the direction of that relationship e. g. , The level of social interaction among PG and UG students is different 24

Empirical Research q Empirical research: Describe, explain, and make predictions through observation. to represent observations and propositions based on sense and experience by methods of inductive logic, including mathematics and statistics. q Must be guided by Theory or Model. 25

Hypothesis Testing q Prove or disprove? q Minimise q Test bias it as rigorously as possibly – objectivity q In testing you need to provide evidence through results of testing to support or otherwise the hypothesis at an appropriate level of confidence. 26

Theory Not the opposite of fact. q More complex than a hypothesis. q Set of systematically inter-related concepts, definitions and propositions that are advanced to explain and predict phenomena (facts). q Integrated set of principles that explain and predict facts q Example: The number of housing approvals can be used to predict the strength of the 27 economy.

Models q. A representation of a system that is constructed to study that system in part or in whole. q Theory explains, Models represent. Physical Numerical Statistical Simulation Virtual etc. 28

Research Proposal q Written proposal q Provides clear statement of project aim, significance and hypotheses q Incorporates decisions made during early research planning phases. q Usually covers the research question, research design, sampling design and budget. q Provides evaluation of benefits / contribution of research. 29

Research vs. Consultancy Report relevant research question oriented (target oriented) thorough (compact) literature and empirics based reasoning (reference) valid and reliable (subjectively trustworthy) novel, but rooted to the body of knowledge (rooted in best practices) 30