W e l e a d

Transcription

1 1

2 2

3 Developing a Robust Research Framework T. Ramayah School of Management Universiti Sains Malaysia ramayah@usm.my

4 Variables in Research Moderator Independent Mediator Dependent Control 4

5 Caution Diagrams can help organize relationships in a logical order, understand how concepts intervene or moderate and even describe how a process unfolds over time. They can provide structure to amorphous concepts and highlight inconsistency in arguments. However, as useful as they can be, diagrams are only stage props rather than the performance itself (Sutton & Staw, 1995: 376). 5

6 Research This is part of Alice s conversation with the Cheshire Cat. In this Alice asks the Cat (Carroll 1989:63 4): Would you tell me, please, which way I ought to walk from here? That depends a good deal on where you want to get to, said the Cat. I don t much care where, said Alice. Then it doesn t matter which way you walk, said the Cat.

7 Research Problem A situation that needs a solution and for which there are some possible solutions. If a situation has no solutions then it makes little sense or no sense expending resources to researching it. A research problem may be described as an incongruence; a discrepancy between what is and what ought to be. It may be also described as the gap in knowledge that needs to be filled. 7

8

9 Deductive Theory Develop hypothesis Collect and Analyze data Accept/reject hypothesis 3-9

10 Inductive Develop Theory Analyze patterns & themes Formulate Relationship Observe phenomenon 3-10

11 Contribution in Research

12 Conceptual Contributions 1. improved conceptual definitions of the original constructs; 2. the identification and conceptual definition of additional constructs to be added to the conceptual framework (e.g., additional dependent, independent, mediating, and/or moderator variables); 3. the development of additional theoretical linkages (i.e., research hypotheses) with their accompanying rationale; and 4. the development of improved theoretical rationale for existing linkages.

13 Empirical Contribution 1. testing a theoretical linkage between two constructs that has not previously been tested, 2. examining the effects of a potential moderator variable on the nature of the relationship between two constructs, 3. determining the degree to which a variable mediates the relationship between two constructs, and 4. investigating the psychometric properties of an important scale.

14 Methodological Contribution 1. reduce the potential problems with shared method variance through the insightful use of multiple methods of measurement, 2. increase the generalizability of the research through more appropriate sampling procedures, 3. allow the investigation of the plausibility of "thirdvariable explanations" for the results of past studies, and/or 4. enhance the construct validity of key measures through the use of refined multiple-item measures and/or the use of measurement approaches that do not rely on self-reports.

15 Example

16 Contribution Continuum

17 Contribution Continuum

18 Suggested Reading Summers, J. O. (2001). Guidelines for conducting research and publishing in marketing: From conceptualization through the review process. Journal of the Academy of Marketing Science, 29(4), Morgan, N. A., Kaleka, A., & Katsikeas, C. S. (2002). Antecedents of Export Venture Performance: A Theoretical Model and Empirical Assessment. Journal of Marketing, 68(1),

19 Theory A set of concepts, definition and propositions that are inter related systematically which is forwarded to explain or predict a phenomenon

20 Theory of Reasoned Action (TRA, 1980) Theory of Planned Behavior (TPB, 1991) 3-20

21 Theory Levels Types Competing

22 Differences Theoretical Framework: Stimulus elicits response Conceptual Framework: New teaching method improves students academic performance.

23 Use of Theories Based on XYZ theory we predict that A1 will affect B1 or Based on the findings in Often Cited Study K, A1 will predict B1 are inadequate. Simply naming the theories and citing key publications is not the same as articulating the logic that informs the relationship among constructs. 23

24 Use of Theories A thorough discussion of the constructs results in a deepening of the understanding of the constructs but not of the relationship among them. What must be clear is how and why constructs influence each other, the logic of the direction of the relationship specified and under what conditions the relationship is predicted to exist. 24

25 Theory If a theory is directly borrowed from marketing, psychology, or economics and not extended, refined, or limited by boundary conditions, then the contribution to theory is minimal and the manuscript is likely to contain a modest contribution at best (Ladik & Stewart, 2008). 25

26 Literature Review Comprehensiveness Current Ritual citations and Citation bombing 26

27 Literature Summary 27

28 Concept Map Performance Impacts Task Characteristics 18,19,20,21,29 17,18, 20,21,25,44 Perceived Usefulness Technology Characteristics 6,18,19,26,28 19, 28,36,37 Task- Technology Fit 28,45 18,19,20,22,23,27,28,29,30,31 18,26,33 Information Quality 25,38,39,44 12,13,14,15,16,17,25,44 End User Computer Satisfaction 1,2,3,4,8,9,10,11 Perceived Ease of Use 5,6,7,8,9,10,11 5,6,8,9,10 Attitudes Towards Usage 5,10,23,24 40,41,42,43,44 System Usage 28

29 Heterogeneity Observed Unobserved 29

30 Model of a Moderator (Condition) Independent Dependent Moderator Who did it work for? When does it work? 30

31 Model of a Moderator (Condition) Satisfaction Loyalty Gender Who did it work for? When does it work? 31

32 Goal Setting Theory (Locke et al., 1981) Setting Difficult goals Performance Task Ability Who did it work for? When does it work? 32

33 Moderator Variable A moderator specifies the conditions under which a given effect occurs, as well as the conditions under which the direction (nature) or strength of an effect vary. Baron and Kenny (1986, pp. 1174, 1178) describe a moderator variable as the following: 33

34 Moderator Variable A qualitative (e.g., sex, race, class) or quantitative variable... that affects the direction and/or strength of a relation between an independent or predictor variable and a dependent or criterion variable... a basic moderator effect can be presented as an interaction between a focal independent variable and a factor (the moderator) that specifies the appropriate conditions for its operation...moderator variables are typically introduced when there is an unexpectedly weak or inconsistent relation between a predictor and a criterion variable. 34

35 Model of a Moderator (Condition) Satisfaction Loyalty Image 35

36 Hypothesis Should I Hypothesize the form of My Interactions in Advance? YES, not only should the existence of an interaction effect be predicted, but also its form. In particular, whether a moderator increases or decreases the association between two other variables should be specified as part of the a priori hypothesis (Dawson, 2014). 36

37 Hypothesis H1: The positive relationship between satisfaction and loyalty will be stronger when perceived image is high. H2: The positive relationship between satisfaction and loyalty would be stronger for male compared to female. 37

38 Hypothesis H3: Body Mass Index (BMI) moderates the relationship between exercise and weight loss, such that for those with a low BMI, the effect is negative (i.e., you gain weight - muscle mass), and for those with a high BMI, the effect is positive (i.e., exercising leads to weight loss) 38

39 Suggested Reading Sharma, S., Durand, R. M., & Gur-Arie, O. (1981). Identification and analysis of moderator variables. Journal of Marketing Research, 18(3), Dawson, J. F. (2014). Moderation in Management Research: What, Why, When, and How. Journal of Business and Psychology, 29(1), Chin, W. W., Marcolin, B. L., & Newsted, P. N. (2003). A partial least squares latent variable modeling approach for measuring interaction effects: results from a Monte Carlo simulation study and an electronic-mail emotion/adoption study. Information Systems Research, 14(2),

40 Mediation 40

41 Research Model 41

42 Basic Requirement Despite the extensive use of complex statistical modeling in the behavioral sciences, the quality of a research project is largely determined by the design decisions that are made before any analysis is done and even before the study is conducted. The conceptualization of a mediation analysis requires forethought about the relationships between the variables of interest and the theoretical meaning behind those relationships. (McKinnon et al., 2012) 42

43 Mediator Variable (Mechanism) A mediator specifies how (or the mechanism by which) a given effect occurs (Baron & Kenny, 1986; James & Brett, 1984). Baron and Kenny (1986, pp. 1173, 1178) describe a mediator variable as the following: 43

44 Mediator Variable The generative mechanism through which the focal independent variable is able to influence the dependent variable of interest... (and) Mediation... is best done in the case of a strong relation between the predictor and criterion variable. How did it work? Why did it work? 44

45 Suggested Reading Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical Mediation Analysis in the New Millennium. Communication Monographs, 76(4), Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York: Guilford Press. MacKinnon, D. P., Coxe, S., & Baraldi, A. N. (2012). Guidelines for the Investigation of Mediating Variables in Business Research. Journal of Business and Psychology, 27(1), Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36(4),

46 Methods for testing Mediation Variation 1. Implicit Procedures 2. Explicit Procedures

47 Implicit Procedures The more traditional and familiar methods such as the BK Method and the JMB Method. They infer the presence (or absence) of mediation processes by examining the joint significance of various constituent paths connecting X to Y (Cohen & Cohen, 1983; MacKinnon et al., 2002). The mediation effect of M1 is deemed to be established when the direct path from X to M1 (i.e., a1 ) is significant and the direct path from M1 to Y is significant (i.e., b1) is also significant.

48 Explicit Procedures Methods that conduct formal statistical tests for the presence (or absence) of mediation effects and include: 1. The Sobel Test (Sobel, 1982); 2. The Distribution of Product Method (DOP Method) (MacKinnon et al., 2002, 2007; Williams & MacKinnon, 2008); 3. Bootstrapping (Bollen & Stine, 1990; Shrout & Bolger, 2002); 4. Monte Carlo Simulation (MacKinnon et al., 2004; Preacher and Selig, 2002); and 5. Bayesian Estimation (Yuan and MacKinnon, 2009).

49 Measurement 49

50 Bacharach s (1989) We want to know about this, but we can only test here. SEM may be useful here

51 Reflective (Scale) Versus Formative (Index) Operationalization of Constructs A central research question in social science research, particularly marketing and MIS, focuses on the operationalization of complex constructs: Are indicators causing or being caused by the latent variable/construct measured by them? Indicator 1 Indicator 2 Indicator 3 Indicator 1 Indicator 2 Indicator 3? Construct Construct Changes in the latent variable directly cause changes in the assigned indicators Changes in one or more of the indicators causes changes in the latent variable 51

52 Measurement Models Reflective Formative TIMELINESS LIFE STRESS X1 X2 X3 X1 X2 X3 X1 = Accommodate last minute request X2 = Punctuality in meeting deadlines X3 = Speed of returning phone calls X1 = Job loss X2 = Divorce X3 = Recent accident Indicators must be highly correlated (Hulland, 1999) Indicators can have +, - or 0 correlation (Hulland, 1999) 52

53 Hair et al. (2014) Reflective Measurement Model Formative Measurement Model 53

54 Problems in Specification Reflective measurement is most commonly used but in many cases a formative measurement would be appropriate Should be reflective Should be formative Total Modelled as reflective Modelled as formative 947 (65%) 17 (1%) 455 (31%) 41 (3%) 1403 (96%) 58 (4%) 32% of the constructs have been measured incorrectly Total 964 (66%) 497 (34%) 1461 (100%) Jarvis et al. (2003) Journal of Consumer research, Journal of Marketing, Journal of Marketing Research, Marketing Science ( ): N = 1,192 Fassot (2006): (X 2003) N =

55 Uni or Multi Dimensional 55

56 Sampling 56

57 Generalizability Two distinct types of generalizability are identified in consumer research. 1. One entails the application of specific effects, whereas 2. the other entails the application of general scientific theory. Effects application and theory application rest on different philosophical assumptions, and have different methodological implications.

58 Effects Application The first type of generalizability, effects application, maps observed data directly into events beyond the research setting. That is, the specific effects obtained are expected to mirror findings that would be observed if data were collected for other populations and settings in the real world.

59 Theory Application The second type, theory application, uses only scientific theory to explain events beyond the research setting. Effects observed in the research are employed to assess the status of theory. But, it is the theoretical explanation that is expected to be generalizable and not the particular effects obtained.

60 Knowledge Claim 1-60

61 Hulland et al. (2017) When the primary aim of the research is to test the veracity of proposed theoretical effects, the use of a convenience sample may suffice. If the goal is to test theoretical hypotheses of interest, as is usually the case, the most important consideration is to select measurement objects and a research context in which the hypotheses can be meaningfully tested, using sources of data (both primary and secondary) that yield accurate information about the units studied. 61

62 Hulland et al. (2017) However, if the researcher is interested in a particular target population to which generalizations are to be made, an explicit sampling frame has to be specified and the manner in which the sample was drawn needs to be clearly and explicitly described. 62

63 Hulland et al. (2017) First, survey researchers follow certain established survey practices ritualistically, even when they are not directly relevant to the research in question or when they are not particularly meaningful. For example, a focus on coverage, sampling, and non-response error seems misplaced when there is no real target population to which the researcher wants to generalize the findings. However, in typical academic marketing surveys there is often no obvious target population to which the researcher wants to generalize the findings, the sample studied is arbitrary (e.g., chosen based on ease of access), and it is difficult to talk about selection bias when the sample is one of convenience (although nonresponse will lead to loss of power). 63

64 64

65 65

66 66

67 Thank you for listening 67