Designs


 * Essential Characteristics of Experimental Research **

I. Comparison of Groups a. Usually involves two groups of subjects: Experimental group and a Control or comparison group i. __Experimental Group__- receives the treatment ii. __Control/Comparison Group __ – allows the researcher to determine if the treatment has had an effect or if one treatment is more effective than another.

II. Manipulation of the Independent Variable a. The researcher determines the forms of the independent variable and which group will get which form. b. Independent variables are established by: i. 1 form of the variable versus another form ii. Presence versus absence of a particular form iii. Varying degrees of the same form c. Examples of independent variables that __can__ be manipulated: i. Teaching method <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">ii. Type of counseling <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">iii. Learning activities <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">iv. Assignment given <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">v. Materials <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1in; text-indent: -0.25in;">d. Examples of independent variables that __cannot__ be manipulated: <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">i. Gender <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">ii. Ethnicity <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">iii. Age <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1.5in; text-indent: -1.5in;">iv. Religious preference

<span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 0.75in; text-indent: -0.5in;">III. Randomization <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1in; text-indent: -0.25in;">a. Takes place before the experiment begins <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1in; text-indent: -0.25in;">b. It is a process of assigning or distributing individual groups <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1in; text-indent: -0.25in;">c. Random assignment allows the researcher to form groups that are equivalent <span style="color: black; font-family: 'Times New Roman','serif'; font-size: 13.5pt; line-height: normal; margin: 0in 0in 10pt 1in; text-indent: -0.25in;">d. It eliminate the threat of extraneous or other variables that might affect the outcome of the study

<span style="color: black; display: block; font-family: 'Times New Roman','serif'; font-size: 14pt; line-height: normal; margin: 0in 0in 10pt; text-align: center;">Weak/Pre-Experimental Designs <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The One-Shot Case Study Design
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">A Single group is exposed to treatment and the dependent variable is observed.
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Weakness of this design is the absence of any control.

<span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The One-Group Pre-Test-Posttest Design
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">A single group is measured or observed before and after treatment.
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Weakness of this design: Nine threats to internal validity: history, maturation, instrument decay, data collector characteristics, data collector, bias, testing, statistical regression, attitude of subjects, and implementation.

<span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The Static-Group Comparison Design
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Two already existing groups are used and are not randomly assigned.
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Weakness of this design it is more vulnerable to morality and location and the possibility of differentiated subject characteristics.

<span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The Static-Group Pretest-Posttest Design
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Two already existing groups receive both a pretest and posttest.
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Weakness of this design is the possibility of a testing threat.

** True Experimental Designs ** The essential ingredient of a true experiment is random assignment of subjects to treatment groups. *Random assignment: every individual who is participating in the experiment has an equal chance of being assigned to any of the experimental or control groups.

** Randomized Posttest-Only Control Group Design ** · Involves two randomly assigned groups, one group receives the experimental treatment while the other does not · Both groups are post-tested on the dependent variable · One of the best designs for experimental study · Testing is not a threat because subjects are not tested twice · Threats to internal validity: mortality, drop outs, implementation, data collector bias, location, and history

** Randomized Pretest-Posttest Control Group Design ** · Two groups of subjects are used by random assignment · Both groups are measured twice (Pretest and Posttest) at the same time · ** Pretest treatment interaction threat: ** The possibility that subjects may respond or react differently to a treatment <span style="color: black; font-family: Arial; font-size: 12pt; line-height: normal; margin: 0in 0in 10pt 0.25in; text-indent: -0.25in;">because they have been pretested, thereby creating a threat to internal validity · If groups are not similar, the researcher can use a Matching Design to reorganize the groups · A pretest is necessary to show change over time

** Randomized Solomon Four-Group Design ** · Attempt to eliminate the possible effect of a pretest · Random assignment of subjects to four groups · Two groups are pretested and two groups are not · One of each of the groups receives the treatment and the other two are control · Provides the best control of the threats to internal validity · Requires a large sample · Requires a considerable amount of energy and effort by the researcher

** Random Assignment with Matching ** · Pairs of individuals are matched on certain variables · The variables are based on previous research, theory, or experience · Members of each matched pair are then randomly assigned the experimental or control group · A pretest on the dependent variable is often used to match the subjects · Methods of Matching o ** Mechanical: ** process of pairing two persons whose scores on a particular variable are similar, a check <span style="font-family: Arial; font-size: 12pt; line-height: normal; margin: 0in 0in 10pt 0.75in; text-indent: -0.25in;">should be made to ensure the groups are equivalent o ** Statistical: ** Each subject is given a “predicted score” on the dependent variable. This is based on the <span style="font-family: Arial; font-size: 12pt; line-height: normal; margin: 0in 0in 10pt 0.75in; text-indent: -0.25in;">correlation between the dependent variable and the matching variable. Statistical matching does not <span style="font-family: Arial; font-size: 12pt; line-height: normal; margin: 0in 0in 10pt 0.75in; text-indent: -0.25in;">necessitate a loss of subjects

<span style="color: black; display: block; font-family: 'Times New Roman','serif'; font-size: 14pt; line-height: normal; margin: 0in 0in 10pt; text-align: center;">Quasi-Experimental Designs


 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Researcher has little or no control over the allocation of the treatments or other factors being studied
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Difference in this approach is the lack of random assignment
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Use of time series analysis: interrupted and non-interrupted

<span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">Types of quasi-experimental designs: <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The removed-treatment design <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The case-control design <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The non-equivalent control groups design <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The interrupted time-series design <span style="font-family: 'Times New Roman',Times,serif; font-size: 140%;">The regression discontinuity design

<span style="display: block; font-family: 'Times New Roman',Times,serif; font-size: 140%; text-align: left;">The most commonly used quasi-experimental design is the [|nonequivalent groups design]. In its simplest form it requires a pretest and post-test for a treated and comparison group. It's identical to the [|Analysis of Covariance] design except that the groups are not created through random assignment.