Experimental Design Internal Validation Experimental Design I Definition
- Slides: 11
Experimental Design Internal Validation
Experimental Design I. Definition of Experimental Design II. Simple Experimental Design III. Complex Experimental Design IV. Quasi-Experimental Design V. Threats to Validity
Experimental Design I. Definition of Experimental Design Control over the sequence and proportion of the independent variable involving: 1) at least two conditions (i. e. an independent variable); 2) random assignment of subjects to conditions; and 3) the measurement of some outcome (i. e. dependent variable)
Experimental Design II. Simple Experimental Design 1. Post-test Control Group Designs (t-test) E R X -> O 1 C R Example -> O 2 2. Pre-Post-test Control Group Designs (t-test) E R O 1 X -> O 3 C R O 2 -> O 4
Experimental Design II. Simple Experimental Design 3. Soloman Four Group Design (t-test) E 1 R X 1 -> O 1 C 1 R -> O 2 E 2 R O 1 X 1 -> O 3 C 2 R O 2 -> O 4 4. Analysis of Variance (ANOVA) E 1 R X 1 -> O 1 E 2 R X 2 -> O 2 E 3 R X 3 -> O 3 Example
Experimental Design III. Complex Experimental Design (Factorial Designs) uses ‘Two Way Analysis of Variance’ 1. Completely Randomized Designs (CRD) (This example is a 2 x 3 CRD) C-E 1 C-E 2 C-E 3 R-E 1 O 12 O 13 R-E 2 O 21 O 22 O 23 Main Effects Interaction Effects
Experimental Design III. Complex Experimental Design (cont. ) 2. Incomplete Designs (IRD) Split Plot Design (This example is a 2 x 3 SPD) C-E 1 C-E 2 C-E 3 R-E 1 R-E 2 3. - O 12 O 21 O 22 O 13 - Repeated Measures Designs (RMD) Latin Square Design ( This example is a 4 x 4 4. LSD) O 1 O 2 O 3 O 4 O 1 O 2 O 3
Experimental Design IV. Quasi-Experimental Design 1. One Shot Case Study E O 1 X ->O 2 2. Non-Equivalent Control Group Design E O 1 X -> O 3 C O 2 -> O 4 3. Interrupted Time-Series Design E O 1 O 2 O 3 X O 4 O 5 O 6
Experimental Design V. Threats to Validity 1. History = confounding of IV over time 2. Maturation = age / experience contaminate 3. Testing = subjects come to understanding IV 4. Regression to the Mean = extreme scores regress 5. Selection of Participants = non-random assignment 6. Mortality = subject attrition 7. Diffusion of Treatments = lack of control group Back to the Beginning End Presentation
Two Sample t-test Problem: Suppose you wanted to know if students who work (the experimental condition) take fewer units than students who do not (the control condition). If a sample of 25 working students yielded a mean of 12 units with an unbiased standard deviation of 3 units and 25 who do not work took an average 15 units with an unbiased standard deviation of 4 units, could you conclude that the population of students not working take significantly more units? Step 1 State the hypotheses: Ho: = ; H 1: < Step 2: Specify the distribution: (t-distribution) Step 3: Set alpha (say. 05; one tail test, N>30, therefore t= 1. 65) Step 4: Calculate the outcome: Step 5: Draw the conclusion: Reject Ho: 3. 0 > 1. 65 Working students take significantly fewer units. Back
Multiple Sample Test (ANOVA) Problem: Suppose your instructor divides your class into three sub-groups, each receiving a different teaching strategy (experimental condition). If the following test scores were generated, could you assume that teaching strategy affects test results? 2= 3; H 1: Ho is false At Home 115 125 135 145 155 Step 3: Set alpha (say. 05; therefore F = 3. 89) 140 150 160 Step 4: Calculate the outcome: 145 155 165 175 185 140 150 Both C+H Step 1: State hypotheses: Ho: 1 = In Class 160 Grand Mean = 150 Step 2: Specify the distribution: (F-distribution) Source SS df MS F Bet 1000 2 500 1. 54 Within 3900 12 325 Step 5: Draw the conclusion: Retain Ho: 1. 54 < 3. 89 Type of instruction does not influence test scores. Back
- Experimental vs non experimental
- Experimental vs non experimental
- Experimental research design types
- Experimental vs non experimental research
- Experimental vs non experimental
- Ist 331
- A nested model for visualization design and validation
- Non-experimental design definition
- Dq design qualification
- Retrospective validation
- Integrative validation
- Validation definition