Discussion of Quasi-Experimental Designs - Coursework Example

Discuss Quasi-Experimental Designs Discuss Quasi-Experimental Designs Part I Jackson Even Numbered Chapter Exercises(p.360) Question#2 I would recommend non-equivalent groups design as the preferred quasi-experimental design for the study. This is because the existing groups used for the weekly quizzes experiment are not divided and the research can not assign the participants randomly. Question #4 Some of the possible confounds in the study outlined in the answers above include history, maturation, instrumentation, test, regression, selection bias and mortality/attrition. Question #6 A researcher might choose to use a single-case design due to a number of potential reasons. For example, single-case designs require fewer participants and are therefore easier to conduct. In addition, single case designs are idiographic in nature, allowing for deeper examination of the effects of intervention on the individuals (Jackson, 2012). Question # 8 Multiple-baseline design differs from a reversal design in that the multiple-baseline design measures multiple variables prior to and after a treatment while a reversal design (also known as ABAB design) is a kind of single-case experimental design which can only measure a single case. Additionally, in Multiple-baseline design, treatment is often introduced to a multiple number of people at different intervals. 2. Advantages and Disadvantages of quasi-experiments Quasi-experiments have a number of advantages as well as potential weaknesses. Some of the major advantages of quasi-experiments include greater validity (as it involves more real world conditions) and significantly much more feasibility given logistical and time constraints (Trochim and Donnelly, 2008). On the other hand, the weaknesses of quasi experiments include less controlled variables, and lack of randomness. The absence of randomization is a fundamental weakness as it can lead to weaker evidence by making. This weakness does not always matter as long as it is recognized as quasi- experiment offers a powerful research tool especially in areas where true experiments may not be feasible. 3. Random Assignment of Participants Yes, when participants are randomly assigned to groups, it is assumed that the groups are equivalent at the beginning of the study. This is particularly attributed to the randomness of the selection. However, the participants eventually became nonequivalent at the end due to the researcher’s lack of control over the selected participants for ach group. This may be solved by using pretests and covariance experimental designs. 4. Threats to Validity of Research The particular outcomes of a given study can be used to effectively suggest that a threat is likely to have been present include covariation, temporal precedence as well as the presence of other plausible alternative explanation for the research results may reveal a threat to the validity of the research. 5. Types of Designs a. Non-equivalent control group pretest only: Involves comparing the treatment and comparison group using pretest measures only. The main selection threats addressed by the design include addictive effects with selection and subject mortality. b. Non-equivalent control group pretest/posttest: The non-equivalent control group design involves comparing the treatment and comparison group using pretest and posttest measure. The design does not address selection threats such as self selection maturation, regression, testing, and instrumentation. This is attributed to the lack of randomness of the samples. c. Cross-sectional: The design involves measuring different variables of a given population at a defined time (Trochim and Donnelly, 2008). Despite is relative cost effectiveness; Cross sectional design fails to address selection threats such as non-probability sampling and self selection. This is because cross sectional studies do not often provide a proper basis for establishing causality. d. Regression-Discontinuity: This is a kind of quasi-experimental pretest-posttest design that particularly measures the impact of intervention. RDD addresses selection threats such as threats such as selection regression, maturation and testing through the application of treatment assignment based on a continuous eligibility index that varies with continuous distribution. 6. Why quasi-experimental designs are used more often Quasi-experimental designs are generally used more often than experimental designs due to a number of reasons some of which include being logistically easy to manage as they reduce time and resource requirements and greater validity due to its involvement of more real world conditions (Trochim and Land, 1982). 7. Research Question for single groups post test only design The readings for this assignment have revealed that different researches require design choices and implementation. An example, of a research question that can fairly yield unambiguous findings for a single groups post test only design is: Does smoking marijuana cause cancer? Part II 1. Research Questions Addressed by the Study Goldberg’s research primarily addresses the research questions such as: i. What is the effectiveness of the various TV adverts directed to the children? ii. How do the Canadian French and English speaking children respond differently to television adverts in each of the two languages? 2. Goldberg’s rationale for the study? Goldberg’s rationale for the research is to contribute to the public debate regarding the appropriateness and effectiveness of TV adverts directed to the children particularly given the law passed in Quebec that outlawed television advertising to children. In general, the study was particularly designed to contribute to the theory that English speaking Canadian children tend to respond more positively to television adverts on American children TVs as compared to their French speaking counterparts (Goldberg, 1990). In my opinion, the results contributed to the theory as a number of variables were scientifically tested with little threats to internal and external validity. 3. Constructs Addressed by the Study The main constructs addressed by the study include consumerism, marketing and Parsimony. However, the constructs are not operationalized. Instead the researcher has used the variables studied in the research to measure and operationalize the particular constructs of interest in the study. 4. Independent and dependent variables in the Search Study The independent variables in Goldberg’s study include the children’s cultural affiliation, language spoken (English or French) and the absence or presence of television adverts. On the other hand, the dependent variables include toy awareness level as well as the cereals purchased by children. 5. The type of Design used by the Researchers The researchers have employed a Quasi-Experimental design to access the effectiveness of the TV adverts that are directed to the children (Goldberg, 1990). This is particularly evidenced by the fact that the empirical study primarily seeks to estimate the casual impact of an intervention (children TV advertisement) on the given target population (children). 6. Internal & External validity threats addressed by the Researchers The researchers have addressed a number of internal and external validity threats some of which include history, maturation, instrumentation, test, regression, selection bias and mortality/attrition. However, the researchers have attempted to address many of these threats in their design by undertaking statistical comparisons between the groups in the study. The threats they did not address was the mortality threat. Generally, this failure has minimal effect on the interpretation of the findings. Finally, Goldberg’s findings are convincing because the study involved the use of more real world condition thereby enhancing its validity. References Dinardo, J. (2008). Natural experiments and quasi-natural experiments. The New Palgrave Dictionary of Economics. 6(3): pp. 856–859. Goldberg, M. E. (1990). A Quasi-Experiment Assessing the Effectiveness of TV Advertising Directed to Children. Journal of Marketing Research 27, 4, 45-46. Jackson, S. L. (2012). Research methods and statistics: A critical thinking approach. Belmont, CA: Wadsworth Cengage Learning. Trochim W. M. K., Donnelly, J. P. (2008). The research methods knowledge base. Mason, OH. Thomson Custom.  Trochim, W. and Land, D. (1982). Designing Designs for Research. The Researche Journal, 1(1): pp. 1-6. Read More