RESEARCH DESIGNS OVERVIEW
EXPERIMENTAL DESIGNS
When the Research Question Deals “with what happens if”
Experimental Designs-random assignment pretest data cannot be collected postest only design pretest data can be collected pretest-postest design
QUASIEXPERIMENTAL DESIGN no random assignment to groups pretest data cannot be collected postest only design pretest data can be collected pretest-postest design
Causality TEMPORAL SEQUENCE CONCOMITANT VARIATION ABSENCE OF COMPETING EXPLANATIONS
EXPERIMENTATION RANDOMIZATION CONTROL MANIPULATION
LOGIC OF A TRUE EXPERIMENT When one starts a study with two groups who are equal with respect to important variables, administers a treatment to one group (experimental group) and not to the other (control group) holding all other conditions constant, any differences in the dependent variable following treatment probably were the result of that treatment.
Experimental Designs O=measurement of dep. variable X=application of indep. variable R=randomization
Pretest-postest experiment R O X O R O O
Solomon Four group R O XO R OO R XO R O
Postest Only RXORORXORO
Repeated Measures ROOOO X OOOO ROOOO OOOO
FACTORIAL DESIGN experimental if random assignment to cell
Advantages of Experiments - test cause and effect -control threats to internal validity
Disadvantages of Experiments -assumes all of the relevant variables have been identified -some variables not manipulable -randomization may pose difficulties in field studies -Hawthorne effect may not be avoidable
quasiexperimental designs lack either randomization and or a control group. (always have manipulation)
Nonequivalent control group OXOOOOXOOOO
Postest only-nonequivalent control group X O O
Time Series-non equivalent groups OOOOXOOOOOOOO
Time series design OOOO X OOOO
Crossover design OOX OO XOOOO
Advantages of Quasi- Experiments -practical, feasible,generalizable -adaptable to field studies -for some hypotheses they may be the only way to test -may save some experiments that suffer from attrition
Disadvantages -less confidence in causality -need a priori controls -must rule out alternative explanations -need more replication
EVALUATION RESEARCH
objectives: --to discover whether and how well objectives are being fulfilled --to discover the reasons for specific successes and failures --to direct the cours of experiment with techniques for increasing effectiveness --to uncover principles underlying a successful program --to redefine goals and objectives and means of obtaining them
Nonexperimental designs
DESCRIPTIVE/EXPLORATORY SURVEY collect detailed data on the status quo determine differences between varibles
Descriptive/Survey Research CHARACTERISTICS ATTITUDES FACTS OPINIONS BELIEFS BEHAVIORS
Descriptive Surveys Use: questionnaires or interviews Large samples Superficial information emphasizes breadth rather than depth. Researcher needs knowledge of techniques. Can be costly.
Descriptive Correlational Studies Tests covariance between variables
advantages increased flexibility when investigating complex relationships among variables efficient and effective method of collecting a large amount of data about a problem potential for practical application in clinical settings foundation for future experimental research studies explores relationships between variables that are inherently not manipulable
Disadvantages: the researcher is unable to manipulate the variables generalizability is decreased (preexisting groups) no causal relationship
EX POST FACTO DESIGNS “natural experiment” independent variable not manipulated
Types of Ex Post Facto Studies prospective retrospective longitudinal cross-sectional
advantages: similar to correlational advantages realistic and natural
disadvantages: no causality must rule out alternative explanations finding similar comparison groups difficult
Methodological Research Instrument Construction and Testing (psychometrics) Perfecting Research Techniques
METAANALYSIS-synthesis of research findings