Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
25 Cards in this Set
- Front
- Back
Control in Research |
Controlling those threats of internal validity through control procedures, researchers are more certain of the influences of one variable (IV) and another (DV) |
|
What is the goal of experimentation? |
To maximize experimental variance, control variance due to confounds and minimize error variance |
|
Three Main Control Techniques Used to Reduce Threats to Internal Validity |
1. Control as related to participant assignment
2. Control as related to experimental design
3. Control as related to the logic of experimentation |
|
Participant Selection |
Random selection/sampling How you choose participants of your study from a mass population |
|
Participant Assignment |
Relevant as a control procedure as a threat of internal validity Assure study groups are equivalent or relative characteristics at the start of the study *** Participation assignment is NOT the same as participation selection *** |
|
Ways to assign participants to groups to control for potential confounds |
1. Elimination Procedure
2. Equating/Matching Procedure
3. Random Assignment |
|
Elimination Procedure |
ONLY ONE TO ELIMINATE THE CONFOUND takes pre-identified confounding variable and limits all participants to only one level of the confound variable Goal is to ensure the two groups do NOT differ Identify the strongest possible confounding variables and limit our participants to one level of the variable ONLY |
|
Equating/Matching Procedure |
assign participants in a way that the two groups are equal with respect to the potential confound Controls the confound |
|
What are the problems with the first two methods of assignment? |
Both control procedures start off with the research identifying the potential confounds and specifically controlling for them |
|
Random Assignment |
SOLUTION to the major problems caused by the other two methods of assignment Controls for known and unknown potentially confounding variables Confound variance is randomly distributed among the groups |
|
5 Types of Experimental Design |
1. General Design
2. Control Group Designs
3. Randomized Control Group Design
4. Randomized Pre-Post Control Group Designs
5. Solomon Four-Group Design |
|
What do Experimental Designs do? |
Determine the effects of the IV and DV Rule out alternative explanations Requires strict control |
|
EXPERIMENTAL DESIGN General Design |
Group --> Treatment (IV) --> Measurement (DV) Threat: no controlled procedures/ does NOT rule out any threats |
|
EXPERIMENTAL DESIGN Control Group Designs |
Group A --> T --> M (post-test measures) Group B --> no T --> M Haven't rules out confounds |
|
EXPERIMENTAL DESIGN Randomized Control Group Design |
R Group A --> T --> M R Group B --> no T --> M Adding R controls for all possible confounds |
|
EXPERIMENTAL DESIGN Randomized Pre-Post Control Group Design |
R Group A M --> T --> M R Group B M --> no T --> M measure groups on things prior to intervention can interact with the level of IV to have a differential impact on DV |
|
EXPERIMENTAL DESIGN Solomon Four-Group Design |
Combines two basic experimental designs Allows us to assess whether there is an interaction between the treatment and the pretest Randomly assign individuals into 4 treatment groups |
|
Criteria for TRUE Experiments |
Participants MUST BE randomly assigned At least two levels of the IV Controls for major threats to internal validity MUST institute controlled procedures |
|
Paramount Imperative of Experimentation |
F-Ratio should be ~1 before you begin If between > within = threat to internal validity |
|
Between Subjects Design |
experimental designs in which comparisons are made between groups of participants Each participant is in ONLY one study group, receives ONLY one level of the IV and contributes ONLY one score on DV |
|
Completely Randomized Design |
Simple random between-subjects design number of groups is dependent on how may levels of IV there are |
|
Multilevel Between-Subjects Design
|
Each group receives a distinct level
Amount of levels determines amount of treatments |
|
What are the Three Hypotheses for Multilevel Completely Randomized Design? |
1. Are the results due to chance? 2. Are the results due to a confound? 3. Are the results due to the IV? |
|
HYPOTHESE Due to Chance |
A significant F = reject the null -- difference between groups Single factor ANOVA Post-Hoc tests determine where the difference is |
|
HYPOTHESES Due to a Confound |
any alternate explanations due to the difference random assignment and control group |