Experimental Research
Quasi- Experimental Designs: Identical to true experimental design but does not have random assignment and results in groups that are non-equivalent. These designs are inferior to true experimental designs internal validity due to variety of selection related threats such as history or maturation threats
Two Group Experimental Designs
Factorial Designs: Allows for manipulation of two or more independent variables. Each independent variable is called a factor and each sub-division of a factor is called a level. Allows the researcher to examine the individual effect of treatment on each dependent variable (main effect) and the joint effect (interaction effect). Most basic design is 2 x 2- each number represents a factor and the value of each factor represents the number of levels in the factor
Hybrid Experimental Designs
One or more independent variables are manipulated by the researcher (as treatments), subjects are randomly assigned to different treatment levels and the results of the treatments (dependent variable) are observed. Strength is the internal validity to link cause and effect through treatment manipulation.
Basic Concepts
Treatment and control group: subjects are administered one or more experimental stimulus called treatment (treatment group) and the other subjects and not given anything (control group)
Treatment manipulation helps control the "cause" in the cause-effect relationship .Must be checked using pretest and pilot tests prior to the study.
Random selection and assignment: Random selection process of randomly selecting a sample from population. Random assignment is randomly assigning subjects to experimental or control groups
Threats to internal validity include history threats, maturation threat, testing threat, regression threat, mortality threat and instrumentation threat
click to edit
Pretest-posttest control group: Subjects are randomly assigned to treatment and control groups, subjected to pretest measurement of dependent variables, treatment group is administered treatment (independent variable) and dependent variable is tested again (posttest)
Posttest only control group :Simpler version of pretest-posttest because the pretest measurement is omitted
Covariance designs : Measures of dependent variables influenced by covariates (variables that are not central interest to experimental study but should be controlled to eliminate the potential effect on dependent variable)
Solomon four group design : Sample is divided into two treatment groups and two control groups. One treatment and control group receive the pretest, the other groups do not. Intended to test the biasing effect of pretest measurement on posttest measures
Switched replications design A two group design implemented in two phases with three waves of measurement. Treatment group in first phase is the control group in the second phase and the control group in the first phase is the treatment group in the second phase. All subjects receive treatment either during the first of second phase
Randomized block design : variation of posttest-only or pretest-posttest control group design where the subject population can be grouped into homogenous subgroups (blocks), which replicates the experiment. It can reduce the "noise" in data that attributes to the differences between the blocks
Regression discontinuity design
Proxy pretest design
Non equivalent switched replication design
Separate pretest-posttest samples design
Nonequivalent group design
Non equivalent dependent variable