Quantitative research Designs

1. Quantitative research Designs

2. Objectives • Define experimental research • Distinguish between experimental, quasiexperimental, and preexperimental designs • Discuss three experimental designs • Discuss two quasiexperimental designs • Discuss two preexperimental • Discuss four types of nonexperimental research designs • Discuss internal and external validity of experimental designs • Identify 6 threats to internal • Identify 3 threats to external validity

3. Quantitative Research Design It is the overall plan for gathering data. The plan that the researchers plan to adopt to develop out the strategies information that is accurate and interpretable. Must be appropriate to test the study hypothesis(es) or answer research problems

4. Research designs: Two broad categories • Quantitative: • Experimental • Nonexperimental • Qualitative

5. Experimental research designs: • True experimental designs • Quasiexperimental designs • Preexperimental designs

6. True experimental research designs: • Pretest-posttest control group • R O1 X O2 (experimental group) • R O1 O2 (comparison group) • Posttest-Only Control Group • R X O1 (experimental group) • R O1 (comparison group) • Solomon four-group

7. Solomon four-group • R O1 X O2 (experimental group1) • R O1 O2 (comparison group1) • R X O2 (experimental group2) • R O2 (comparison group2)

8. Quasiexperimental designs • Nonequivalent Control Group Design • O1 X O2 (experimental group) • O1 O2 (comparison group) • Time Series Design • O1 O2 O3 X O4 O5 O6

9. Preexperimental designs: • One-Shot Case Study • X O • One-Group Pretest-Posttest Design • O1 X O2

10. Nonexperimental designs • Survey studies • Correlational studies • Comparative studies • Action studies • Developmental studies • Evaluation studies • Meta-analysis studies • Methodological studies • Needs assessment studies • Secondary studies

11. Experimental Research: • Must have three elements: • Manipulation of the independent variable • Control: Introduce controls over the experimental situation, including the use of a control group • Randomization: assignment to control and experimental groups randomly

12. True experimental designs: • Studies in which the researcher has a great deal of control over the research situation. • Minimal internal validity threat (The degree to which changes in the dependent variable (effect) can be attributed to the independent variable (cause)

13. Pretest-Posttest Control Group Design • Most frequent design Symbolic representation: 1. R O ¹ X O ² (experimental group) 2. R O ¹ - O ² (comparison group)

14. Characteristics: • Most frequently used experimental design • Controls for all threats to internal validity • Disadvantages: The external threat of the reactive effects of the pretest. Results can be generalized only to situations with pretest administration before treatment

15. Posttest-Only Control Group Design • R X O¹(experimental group) • R O ¹ (comparison group)

16. Features of Posttest-Only Control Group • Easier and superior to pretest-posttest design • Random assignment of subjects into groups in the posttest-only group ensures equality

17. Solomon Four Group Design: • Subjects are randomly assigned to one of four groups: two experimental and two comparison groups • R O¹ X O ² (experimental group 1) • R O¹ O ² (comparison group1) • R X O ² (experimental group2) • R O ² (comparison group 2)

18. Solomon Four-Group Design • Most powerful experimental design because it minimizes threats to internal and external validity • It controls for all threats to internal validity and for reactive effects of the pretest. Any differences between the experimental and the comparison groups can be more confidently associated with experimental treatment

19. Disadvantage of Solomon Four-Group Design • Requires a large sample • Complicated analysis

20. Example of Solomon four groupdesign • Effects of caring, measurement, and time on miscarriage impact and women's well-being.Swanson KM.BACKGROUND: Responses to miscarriage range from relief to devastation, yet there have been no randomized controlled studies that demonstrate significant effects of counseling with women who miscarry. OBJECTIVE: To test the effects of caring-based counseling, measurement, and time on the integration of loss (miscarriage impact) and women's emotional well-being (moods and self-esteem) in the first year after miscarrying. METHOD: ANCOVA was used in this randomized, longitudinal Solomon four-group experimental investigation. Enrolled were 242; 185 completed. Outcomes included: self-esteem, overall emotional disturbance, anger, depression, anxiety, confusion, overall miscarriage impact, personal significance, devastating event, lost baby, and isolated. RESULTS: During the first year after loss (a) caring was effective in reducing overall emotional disturbance, anger, and depression; and (b) time passing led to increased self-esteem and decreased anxiety, depression, anger, confusion, and personal significance of loss. CONCLUSION: Caring, measurement, and time had some positive and significant effects on the integration of loss and enhancement of well-being in the first year subsequent to miscarrying.

21. Quasiexperimental Designs • When the design lacks: Random assignment of subjects. Comparison group The researcher uses an existent intact groups for the experimental and comparison groups

22. Quasiexperimental Designs/cont Two designs: 1. Nonequivalent Control Group: Similar to pretest-posttest control group design but with no random assignment of subjects to experimental and comparison groups O1 X O2 (experimental group) O1 O2 (comparison group)

23. Quasiexperimental Designs/Nonequivalent Control Group • Problems of this design: • Threats to internal validity are history, testing, maturation and instrumentation change • Biggest threat is selection bias (no one can tell if the two groups were similar to start with???)

24. Quasiexperimental Designs/cont 2. Time-Series Design The researcher periodically observes or measures the subjects O1 O2 O3 X O4 O5 O6 Example: assessing pain level of a group of people with low back ache over three weeks Then give a specific exercise to reduce the pain. And pain level is measured again to determine if low backache still persists

25. Time Series Design/cont • Problems with this design: • Threats to validity are history and testing

26. Preexperimental designs • Very weak design • Types: • One shot case study: exposure of one group to an experimental treatment and observed after the treatment X O • Example: A group of diabetics attend a clinic class (x) and tested on their knowledge of diabetes (o) after the class is completed. • No comparisons made • No one can tell whether they had the knowledge before the intervention • Threats to internal validity are history, maturation and selection bias

27. One Group Pretest-Posttest Design • Provides a comparison between a group of subjects before and after the experimental treatment O1 X O2 Example: a group of clients are given a pretest of their knowledge. Then they are given the class and be post tested at the end of the class. Threats to internal validity: history, maturation, testing, and instrumentation change.

28. Nonexperimental research • Is descriptive in nature because there is no manipulation or control of variables • The researcher must control for extraneous variables through careful selection of study sample • Extraneous variables threaten the validity of the study

29. Nonexperimental research/cont • Types: • Survey (political polls): self report data from samples with the purpose of describing populations on some variables of interest • Most important are the sampling techniques, which depend on: sampling methods (probability sampling and sample size)

30. Survey • Methods of conduction: • Questionnaires and interviews conducted by: Phone, mail, or personal contact with subjects

31. Survey/cont • Advantages: • Producing accurate information with relatively small samples • Quick with minimal cost data • Disadvantages: • Self report responses may be unreliable because of socially acceptable responses

32. Correlational studies • Examines the strength of relationship between variables. X is related to Y or the relationship between more than two variables. • Magnitude and direction of relationship are indicated by a correlation coefficient It can be positive (+) or negative (-) and ranges between -1.00 - +1.00 (negative to positive correlation).

33. Correlation coefficient reported through statistics such as Pearson’s product-moment correlation r and or the Spearman rho • Example: age and assertiveness, it is not a causation relationship

34. Comparative studies • Examines the differences between intact groups on some dependent variable of interest. • Similar to experimental design but the researcher does not manipulate the independent variable (it is inherent characteristic).

35. Comparative studies/cont • Example: a choice of two groups: • 1. Children exposed to abuse (independent V) and compared to children who were not exposed to abuse their childhood • 2. The dependent variable may be the self esteem (effect)

36. Comparative studies/Cont • Two types: • Retrospective: (ex post facto studies or after the fact) The dependent variable (effect) is identified in the present (a disease condition) and an attempt is made to determine the independent variable (cause of the disease) that occurred in the past.

37. Prospective: • The independent variable (high cholesterol) is identified at the present time, and then subjects are followed in the future to observe the dependent variable (Incidence of coronary artery disease). • Prospective studies may use an experimental approach whereas retrospective studies would never use this type of design.

38. Disadvantages • Prospective studies are costly • May have subject dropout

39. Methodological studies • Studies that are concerned with the development, testing, and evaluation of research instruments and methods

40. Validity of experimental Designs • Extraneous variables: • Extraneous variables are those variables that the researcher is not able to control, or does not choose to control, and that may influence the results of a study. • Other names for extraneous variables are: confounding or intervening variables or study limitations. • In experimental studies, extraneous variables are considered threats to internal and external validity.

41. Internal validity: • The degree to which changes in the dependent variable (effect) can be attributed to the independent variable (cause). • Threats to internal validity are factors other than the independent variable that influence the dependent variable. These factors constitute rival explanations or competing hypotheses that might explain the study results.

42. External validity: the degree to which study results can be generalized to other people or other settings. • Questions to be asked in external validity are: • 1. With what degree of confidence can the study findings be transferred from the sample to the entire population? • 2. Will these study findings hold true with other groups in other times and places?

43. Threats to internal validity: • Selection bias: subject differences to begin with rather than the manipulation of the independent variable. (selection of a group who is different ‘in motivation’ more than in the control group to change • History: When some event other than the experimental treatment occurs during the course of the study (Introducing a teaching program and a newspaper article is published in the same topic area of the independent variable)

44. Threats to internal validity/cont: • Maturation: Changes that occur within the subjects during the experimental study influence the study results. (people may become older, taller, or sleepier from the time of pretest to the posttest. (studies of growth among children with the introduction of an independent variable as the cause and at the same time changes may occur in these children during the course of the study)

45. Testing: • It refers to the influence of the pretest or knowledge of baseline data on the posttest • Subjects may remember the answers they put on the pretest and put the same on the posttest e.g., in a study of weight reduction program, knowledge of your pretest result (weight) may make you make more effort in reducing weight regardless whether you take the program or not

46. Instrumentation change: • When mechanical instruments or judges (in observational studies) are used in the pretest and posttest phases of the study. • A change of the instrument itself or the difference in the ratings of the judges as they became more tired or make less or more observations

47. Mortality: • Subject dropout: dropouts are of a particular characteristics different from those who remained in the study • The longer the study lasts the more subjects drop out will occur • Subjects mortality is another factor.

48. Threats to external validity: • Hawthorne effect: when the study participants respond in a certain manner because they are aware that they are being observed (also a threat to internal validity)

49. Experimental effect: A threat that occurs when the researcher characteristics or behaviors influence the subjects behaviors. Eg., facial expressions, age, gender etc. • Used in experimental design research • Rosenthal effect: the influence of the interviewer on respondents answers, e.g.., the way the researcher dresses might influence the study participants

50. Reactive effects of the pretest: • Also called measurement effect: the subjects are sensitized to the treatment through taking the pretest