Chapter4.pptx management principles in action

management

1. Research Designs: Causal Research

2. 2 2 LEARNING OBJECTIVES After reading this chapter, you should be able to understand: The concept of causality The conditions of causality Natural experiments Controlled experiments The sources of experimental errors Experimental versus non-experimental designs Experimentation in marketing research The criteria of research design

3. CONCEPT OF CAUSALITY A causal research design investigates the cause and effect relationship between two or more variables. But, the concept of causality is complex to understand. Let us take for example, the statement ‘X causes Y’. In everyday interpretation, the statement implies that any change in Y is based on X. However, a scientist would interpret it differently and would rather say X is only one of the factors causing a change in Y. Thus, while everyday interpretation implies a deterministic relationship, in contrast, the scientific interpretation implies a probabilistic relationship. 3

4. CONDITIONS FOR CAUSALITY There are three conditions of causality, namely concomitant variation, time order of occurrence of variables, and elimination of other possible causal factors. These conditions are necessary, but not sufficient for causality. In other words, this means that even when all the three conditions are met in a given case, it cannot be said decisively still that there is a causal relationship at the first place. 4

5. Concomitant Variation Concomitant variation, or invariant association, is a common basis for ascribing cause. Concomitant variation is the extent to which a cause X and an effect Y occur together or vary together in the way predicted by the hypothesis under consideration. 5

6. Time Order of Occurrence of Variables A causal relationship between two variables based on the time order of their occurrence is conceptually simple. It implies that a variable cannot be considered the cause of another, especially if it occurs after the other event. A causal factor needs to act as a ‘trigger’ earlier or even simultaneously with the occurrence of an event. 6

7. Elimination of Other Possible Causal Factors Elimination of other possible causal factors indicate that when a causal relationship between two variables is being studied, and there are one or more other possible factors affecting that relationship , it is those factors that need to be eliminated, ensuring thereby that the relationship is exclusively between two factors. 7

8. SOME DEFINITIONS AND CONCEPTS Experiment Treatments Test Units Dependent Variables Extraneous Variables Experimental Design 8

9. SOME DEFINITIONS AND CONCEPTS Experiment - An experiment is executed when one or more independent variables are consciously manipulated or controlled by the person running the experiment, and their effect on the dependent variable or variables is measured. Treatments - Treatments are the alternatives or independent variables that are manipulated, and whose effects are measured. Test Units - Test units are entities to whom the treatments are presented, and whose response to the treatments is measured. 9

10. SOME DEFINITIONS AND CONCEPTS Dependent Variables - These are the measures taken on the test units. Typical marketing examples include sales, preference, awareness, customer satisfaction, etc . Extraneous Variables - These are variables other than treatments that affect the response of test units to the treatments. Experimental Design - An experimental design involves the specification of ( i ) treatments that are to be manipulated, (ii) test units to be used, (iii) dependent variables to be measured and (iv) procedures for dealing with extraneous variables. 10

11. SYMBOLS DEFINED X represents the exposure of a test group to an empirical treatment, the effects of which are to be determined. O refers to the processes of observation or measurement of the dependent variable on the test units. R indicates that individuals have been assigned randomly to separate treatment groups, or that the groups in themselves have been allocated randomly to separate treatments . Movement from left to right indicates movement through time. All symbols in any one row refer to a specific treatment group. Symbols that are vertical to one another refer to activities or events that occur simultaneously 11

12. CAUSAL INFERENCE STUDIES Natural Experiments - There are three classes of designs for natural experiments: ( i ) time-series and trend designs, (ii) cross-sectional designs and (iii) a combination of the above two . Time-series and trend designs After-only without control group Before–after without control group Multiple time-series Cross-sectional designs Combination of cross-sectional and time series design 12

13. CAUSAL INFERENCE STUDIES Controlled Experiments After-only with control group Before–after with one control group Four-group , six-study design 13

14. Controlled Experiments ‘Before–after’ experimental designs without control were subject to certain limitations : History - it may cause the ‘before’ and ‘after’ measurements to differ; Maturation – it signifies biological and psychological changes in the subject, and which takes place with the passage of time; Pre-testing – it may affect the internal validity of the before–after design, and Measurement variation - variation in measurement may cause variations in the before and after measurements and these may be taken as the effect of the experimental variable. 14

15. SOURCES OF EXPERIMENTAL ERRORS History Maturation Pre-measurement Effect Interactive Testing Effect Instrumentation Selection bias Statistical Regression Mortality 15

16. SOURCES OF EXPERIMENTAL ERRORS History - History refers to the effect of extraneous variables because of an event that is external to an experiment occurring at the same time as the experiment. Maturation - Maturation refers to a gradual change in the experimental units arising due to the passage of time. It is similar to history, but it differs, as the actual outcome is usually less evident. Pre-measurement Effect - This error is caused because of the changes in the dependent variable as a result of the effect of the initial measurement. Interactive Testing Effect - This error arises on account of change in the independent variable as a result of sensitising effect of the initial measurement. 16

17. SOURCES OF EXPERIMENTAL ERRORS Instrumentation - Instrumentation refers to the changes in the measuring instrument over time. Selection bias - Selection bias refers to assigning of experimental units in such a way that the groups differ on the dependent variable even before the treatment. Statistical Regression - Statistical regression effect occurs when test units have been selected for exposure to the treatment based on an extreme pre-treatment measure. Mortality - Mortality refers to the loss of one or more test units while the experiment is in progress. 17

18. EXPERIMENTAL VERSUS NON-EXPERIMENTAL DESIGNS Non-experimental designs (exploratory and descriptive research designs) are unable to provide the control necessary to infer that a causal relationship exists. Descriptive studies as non-experimental design satisfy only one condition of causality, namely, concomitant variation. There is no control over the other two conditions of causality – the time order of the two variables, and the elimination of other possible causal factors. 18

19. PROBLEMS OF EXPERIMENTATION IN MARKETING RESEARCH There are three problems, which need to be considered before undertaking any experiment: Cost is a major consideration in marketing experiments . Time is another consideration while deciding on experimentation . Many control problems arise when one think of conducting an experiment, say, in test marketing. 19

20. CRITERIA OF RESEARCH DESIGN Having discussed a number of research designs, and sources of potential experimental errors, we now turn to the criteria, which a good research design should have. The first criterion of a research design is that it must answer the research questions . The second criterion relates to controlling the independent variables—both the independent variables of the study as also extraneous independent variables . The third criterion is generalisability. 20

21. Cases Case Study 4.1: Huawei Technologies Case Study 4.2: Dunkin’ Donuts 21

22. 22