Experimental and Quasi-Experimental Designs Notes

Research Design

The purpose of a research design is to provide a blueprint or plan for:
- Testing the hypothesis and/or research question(s).
- Identifying the independent and dependent variables.
- Describing how the study will be conducted.
- Describing data collection and analysis processes.
Each research design has specific characteristics to maintain control, which need to be identified and managed.

Quantitative research focuses on describing and testing relationships between variables and/or examining cause-and-effect relationships between variables.
Example: Nursing research looking at variables (interventions) that impact health and wellness, which is testing a relationship.
Research designs that support this type of exploration and evaluation:
- Experimental
- Quasiexperimental

Experimental research is an objective, systematic, controlled investigation to examine the probability and causality among selected independent and dependent variables to predict and control phenomena.
Also referred to as pretest-posttest control group design or classic experiment.
Three key properties:
- Randomization
- Control
- Manipulation

Randomization is the assignment of participants (subjects) to the control group or experimental group randomly.
Each subject has an equal chance of being assigned to either group.
Randomization eliminates any systematic bias that could impact the dependent variable.
Assignment can occur using a computer program or a table of random numbers.
Group assignments must be concealed – no researcher input.
A randomized control trial is considered the best research design.
- Minimum bias is present.
- Level II of evidence hierarchy.

Manipulation: the design identifies the independent variable and how it will be changed or “manipulated” within the experiment.
Focuses on the differences between the control and experimental groups.
Example: A group of students learning about identifying skin cancers; both groups completed a pretest before the in-class lecture.
- Control group – only receives the lecture.
- Experimental group – receives the lecture plus an online video module.
- Both groups took a post-test three (3) days later to assess learning.
- IV – educational strategy; DV – learning outcome.

Antecedent variables: variables that occur before the study but may impact the dependent variable (DV) and confound results.
Example: age, socioeconomic status, health status.
Using randomized control and experimental groups will limit this.
However, the researcher should acknowledge this in the results.

Intervening variables: a condition that occurs during the course of the study and is not part of the study but may affect the dependent variable and, therefore, the study’s results.
Example: side effects of medications or treatment may influence patient views.

Useful when exploring major problems and/or when the number of participants is limited.

Advantages:
- Most appropriate for testing cause-and-effect relationships.
- Provides the highest level of evidence for single studies.
Disadvantages:
- Not all research questions are amenable to experimental manipulation or randomization.
- Participant mortality, especially control group participants.
- Difficult logistics in field settings.
- Hawthorne effect.

Quasi-experimental Research Design intends to test cause-and-effect relationships; however, full experimental control is not possible.
Control is limited due to the nature of the independent variable or the nature of the available participants.
Lacks randomization.
Can result in threats to internal validity:
- Selection bias
- Maturation effects
- Testing effects
- Attrition

Looks exactly like experimental design but participants are not randomized.
Example: teaching two sections of students about diabetes – the researcher is limited by the enrollment into each section – subjects are not random.

Nonequivalent Control Group design is commonly used in nursing research when randomizing the patient population may not be possible due to limited access.
Threats to internal validity, such as selection bias, testing effects, and attrition, are possible.
A weakness of quasi-experimental design is that it creates challenges in establishing clear cause-and-effect relationships.
Researchers may use a priori to rule out extraneous variables.
A priori reasoning or knowledge is an understanding of content that comes from theoretical deduction or learning rather than from observation or experience.

Advantages:
- Practical and more feasible, especially in clinical settings.
- Some generalizability.
Disadvantages:
- Difficult to make clear cause-and-effect statements.
- May not be able to randomize.

	Experimental	Quasi-Experimental
Randomization	Ensures a random participant selection process	Participants cannot be randomly assigned
Control	Holding the conditions of the study consistent	Uses certain mechanisms to ensure the conditions are as consistent as possible
Manipulation	of the independent variable	of the independent variable
Research Question	"What happens if…?"	"What happens if…?"
Participant Assignment	Participants can be randomly assigned to groups	Participants cannot be randomly assigned to groups
Pretest Data	If collected, uses pretest-posttest design; if not collected, after-only design	If collected, uses pretest-posttest design; if not collected, after-only design

What experimental design is used? Is it appropriate?
How are randomization, control, and manipulation applied?
Is there a reason to believe an alternative explanation exists for the findings?
Are all threats to validity, including mortality, addressed in the report?

What quasi-experimental design is used? Is it appropriate?
What are the most common threats to the validity of the findings?
Are there plausible alternative explanations for the findings? Are they addressed?
Does the author address threats to the study’s validity?
Are limitations addressed?