Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Experimental Research

6,616 views

Published on

Industrial Research & Statistical Analysis
ME 204

Published in: Education
  • Be the first to comment

Experimental Research

  1. 1. Marilou K. Peralta Master in Management Engineering Pangasinan State University
  2. 2. PRESENTATION CONTENT 1. What is Experimental Research? 2. Variables and Threats to Experimental Research 3. Types of Experimental Research Designs 4. Pros & Cons of Experimental Research Designs
  3. 3. RESEARCH DESIGN The research design is a logical model that shows the strategies for sample selection, development of measurement tools, data collection as well as methods of data processing and analysis.
  4. 4. QUANTITATIVE RESEARCH RESEARCH EXPERIMENTAL True Experimental Design Controlled experimentation Solomon four experimental design Delayed effects experimental designs NON-EXPERIMENTAL Correlation studies Surveys
  5. 5. Quasi-experimental designs  One shot study  Pretest-posttest design  Contrasted groups  Time series
  6. 6. EXPERIMENTAL RESEARCH It is a systematic and scientific approach to research in which the researcher manipulates one or more variables and controls, and measures any change in other variables The purpose is to study the cause and effect relationship
  7. 7. EXPERIMENTAL RESEARCH An experiment is a set of observations conducted under controlled circumstances in which the investigator manipulates conditions to ascertain what effects such manipulation has on the outcome.
  8. 8. REQUIREMENTS OF A GOOD EXPERIMENTAL DESIGN Requirements of a good experimental design 1. Subjects are randomly assigned to the experimental groups (EG) and control groups (CG) -presence of a comparison group; random assignment
  9. 9. 2. The independent variable(IV) can be manipulated such that any change in the dependent variable (DV) is attributable to changes in the independent variable (manipulable IV)
  10. 10. EXPERIMENTAL RESEARCH VARIABLES: Independent Variable – is presumed to cause changes to occur in another variable, experimental or treatment variable Dependent Variable / Measured Variable – what changes because of another variable, the effect or outcome variable
  11. 11. EXPERIMENTAL RESEARCH EXTRANEOUS VARIABLES - (EV) are effectively controlled to avoid confounding or spurious INTERNAL VALIDITY: indicates whether the independent variable was the sole cause of the change in the dependent variable EXTERNAL VALIDITY: involves the extent to which the results of a study can be generalized (applied) beyond the sample Source: http://web.mst.edu/~psyworl d/extraneous.htm
  12. 12. VALIDITY The internal validity of any research undertaking is strengthened by the correct choice of research design and the soundness and appropriateness of decisions that pertain to sampling, instrumentation, data collection and analysis.
  13. 13. THREATS TO EXPERIMENTAL VALIDITY History Maturation Testing Instrumentation Selection Regression Experimental mortality
  14. 14. EXPERIMENTAL RESEARCH DESIGN 1. PRE-EXPERIMENTAL DESIGN One-Shot Case Study Design One-group Pretest Posttest Design Static Group Comparison 2. TRUE EXPERIMENTAL DESIGN Posttest only Control Group Design Pretest Posttest Control Group Design Solomon Four Group Design 3. QUASI-EXPERIMENTAL DESIGN Non-equivalent Control Group Time Series 4. FACTORIAL DESIGN Multiple Time Series
  15. 15. 1. PRE-EXPERIMENTAL DESIGN One-Shot Case Study Design One-group Pretest Posttest Design Static Group Comparison pre-experimental designs represent the simplest form of research designs  considered “pre-,” indicating they are preparatory or prerequisite to true experimental designs no randomization procedures are used to control for extraneous variables Source: http://srmo.sagepub.com/view/encyc-of-research-design/n330.xml
  16. 16. ONE-SHOT CASE STUDY DESIGN Selected experimental group TREATMEN T POST-TEST
  17. 17. ONE-SHOT CASE STUDY DESIGN Another example of one such design might be a one-off survey of unemployed people in a specific local area to assess their health status and the impact of unemployment on health.
  18. 18. ONE-SHOT CASE STUDY DESIGN Example: the effects of counseling sessions on the attitudes of identified bullies in school. Experimental Group Treatment (X) Posttest (O2) Bully students Counseling Observation Pretest – O1 Posttest – O2 Treatment – X Randomization – R Control Group - C
  19. 19. ONE-SHOT CASE STUDY DESIGN Example: You want to assess the effects of counseling sessions on the attitudes of identified bullies in school. INTERNAL 1. History VALIDITY: – History, parents Maturation, may have Testing, other Instrumentation, counselling Selection, being done Regression, at other Experimental mortality Experimental Group Treatment (X) Posttest (O2) Bully students Counseling Observation place 2. Instrumentation Pretest – O1 Posttest – O2 Treatment – X Randomization – R Control Group - C
  20. 20. ONE-SHOT CASE STUDY DESIGN Example: You want to determine whether praising primary school children makes them do better in Mathematics. INTERNAL VALIDITY: History, Maturation, Testing, Instrumentation, Selection, Regression, Experimental mortality Selection – possible that students selected were already good in Mathematics Experimental Group Treatment (X) Posttest (O) Primary school children History Praising – the school Attentiveness/Performance had organized a motivation course
  21. 21. ONE-GROUP PRETEST-POSTTEST DESIGN Selected experimen tal group PRETEST TREATMEN T POST-TEST (Before and After design)
  22. 22. Another example Testing the effectiveness of a DRUG A on capacity to recall words EG : # words recalled---exposure to DRUG A -----# words recalled CG: # words recalled---no exposure to DRUGA -----# words recalled
  23. 23. ONE-GROUP PRETEST-POSTTEST DESIGN Example: You want to determine whether praising primary school children makes them do better in Mathematics. Experimental Group Pretest(O1) Treatment (X) Posttest(O2) Primary school children Praising Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  24. 24. ONE-GROUP PRETEST-POSTTEST DESIGN Example: You want to determine whether praising primary school children makes them do better in Mathematics. Maturation: period between pretest and posttest is long so subjects may have matured because of developmental changes. Testing: period between the pretest and the posttest is too short and there is the possibility that subjects can remember the questions and answers. INTERNAL VALIDITY: History, Maturation, Testing, Instrumentation, Selection, Regression, Experimental mortality Experimental Group Pretest(O1) Treatment (X) Posttest(O2) Primary school children Praising Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  25. 25. STATIC GROUP COMPARISON experimental group TREATMEN T POST-TEST control group POST-TEST
  26. 26. STATIC GROUP COMPARISON Example: Determine whether praising primary school children makes them do better in Mathematics. Group A Treatment (X) Posttest(O2) Praising Group B Posttest(O2) Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  27. 27. PRE-EXPERIMENTAL DESIGN One-Shot Case Study Design One-group Pretest Posttest Design Static Group Comparison PROs: As exploratory approaches, pre-experiments can be a cost-effective way to discern whether a potential explanation is worthy of further investigation suitable for beginners CONs: Lower validity, difficult to assess the significance of an observed change in the case Very little control over the research, higher threat to internal validity it is difficult or impossible to rule out rival hypotheses or explanations
  28. 28. 2. TRUE EXPERIMENTAL DESIGN Posttest only Control Group Design Pretest Posttest Control Group Design Solomon Four Group Design Manipulation – control of independent variable by the researcher through treatment/intervention Control – the use of control group and extraneous variable Randomization – every subjects have equal chance of being assigned to experimental and control group Random assignment helps ensure that there is no pre-existing condition that will influence the variables and mess up the results.
  29. 29. POSTTEST ONLY CONTROL GROUP Randomly selected experimen tal group TREATMEN T POST-TEST Randomly selected control group POST-TEST
  30. 30. POSTTEST ONLY CONTROL GROUP Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  31. 31. PRETEST POSTTEST CONTROL GROUP DESIGN Randomly selected experimental group PRE-TEST TREATMEN T POST-TEST Randomly selected control group PRE-TEST POST-TEST
  32. 32. PRETEST POSTTEST CONTROL GROUP DESIGN Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  33. 33. SOLOMON FOUR GROUP DESIGN Experimental Group A PRE-TEST TREATMEN T POST-TEST Experimental Group B Control Group A PRE-TEST POST-TEST POST-TEST POST-TEST Control Group B
  34. 34. SOLOMON FOUR GROUP DESIGN Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  35. 35. TRUE EXPERIMENTAL DESIGN Posttest only Control Group Pretest Posttest Control Group Design Solomon Four Group Design PROs: Greater internal validity, Control over extraneous variables is usually greater CONs: Ethical Problems, less external validity It may be unethical or impossible to randomly assign people to groups
  36. 36. 3. QUASI-EXPERIMENTAL DESIGN Non-equivalent Control Group Time Series Multiple Time Series is simply defined as not a true experiment since it does not have randomly assigned groups the comparison/control group is predetermined to be comparable to the treatment group in critical ways. Matching, comparing the same participants over time and pre-existing groups are used. These designs are frequently used when it is not logistically feasible or ethical to conduct a randomized controlled trial. Source: http://education-portal.com/academy/lesson/quasi-experimental-designs-definition-characteristics-types-examples. html#lesson
  37. 37. NON-EQUIVALENT CONTROL GROUP DESIGN Subjects are tested in existing group or intact group rather than being randomly selected O1 X O2 O1 O2 This design should only be used when random assignment is impossible Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  38. 38. NON-EQUIVALENT CONTROL GROUP DESIGN Subjects are tested in existing group or intact group rather than being randomly selected Selection Maturation Regression O1 X O2 O1 O2 Testing This design should only be used when random assignment is impossible Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  39. 39. TIME SERIES A single group is pretested repeatedly until pretest scores are stable , exposed to treatment and, then, repeatedly post tested O1 > O1 > O1 > O1 >O1 >X > O2 >O2 >O2 >O2>O2 Pretest – O1 Posttest – O2 Treatment – X Random Assignment – R Control Group - C
  40. 40. QUASI-EXPERIMENTAL DESIGN Non-equivalent Control Group Time Series Multiple Time Series PROs: can be very useful in generating results for general trends Greater external validity(more like real world conditions) Much more feasible given time and logistical constraints CONs: Without proper randomization, statistical tests can be meaningless Not as many variables controlled Selection could be biased Source: https://explorable.com/quasi-experimental-design
  41. 41. FACTORIAL DESIGN Used to examine the effects that the manipulation of at least 2 independent variables (simultaneously at different levels) has upon the dependent variable. PROs: more precision on each factor than with single factor experimentation, broadening the scope of an experiment, possible to estimate the interaction effect CONs: could be complex, the experiment can be very large with a number of factors each at several levels Source: http://www.stat.ncsu.edu/people/nelson/courses/st512/Factorial%20Experiments.pdf
  42. 42. FACTORIAL DESIGN Example: Driver frustration under low, medium, and high density traffic conditions and under traffic flow controlled by a police officer or a traffic signal was investigated. The measure of frustration was the number of horns honked by drivers before receiving the right-of-way at a controlled intersection. The mean number of horns honked in each condition were: Source: http://www.csupomona.edu/~dhorner/webpages/UOW-Res/094/Exr-2x3-table1-ans.pdf
  43. 43. FACTORIAL DESIGN Source: http://www.csupomona.edu/~dhorner/webpages/UOW-Res/094/Exr-2x3-table1-ans.pdf
  44. 44. EXPERIMENTAL RESEARCH DESIGN 1. PRE-EXPERIMENTAL DESIGN One-Shot Case Study Design One-group Pretest Posttest Design Static Group Comparison 2. TRUE EXPERIMENTAL DESIGN Posttest only Control Group Design Pretest Posttest Control Group Design Solomon Four Group Design 3. QUASI-EXPERIMENTAL DESIGN Non-equivalent Control Group Time Series 4. FACTORIAL DESIGN Multiple Time Series
  45. 45. procedures materials measures participants

×