Continuum of Quantitative Research Designs

Types of Research Designs:
Experimental or Quasiexperimental

n Researcher actively brings about the desired effect

n For testing cause-and-effect relationships

n Criteria to infer causality

– Causal (independent) variable and effect (dependent) variable must be associated with one another

– Causal variable must precede the effect

– Relationship cannot be explained by anyother variable

Experimental Research Designs

n True experiment (pretest-posttest control group)

– Two groups of subjects

» experimental group

» control group

– Subjects are measured before the experimental treatment on the same variables as after the treatment

n After-only design (Post test only control group)

– Data are collected only after the experimental treatment

Experimental Research Designs

n Solomon four-group design

– two experimental groups

» both groups receive the experimental treatment

» one group has pretest and posttest

» one group has posttest only

– two control groups

» neither group receives the experimental treatment

» one group receives pretest and posttest

» one group has posttest only

Experimental Research Design:
Solomon four-group design

n Subjects are randomly assigned to 1 of 4 groups

– Experimental groups 1 & 2 (E1 & E2)

– Control groups 1 & 2 (C1 & C2)

E1 Pretest ---> Treatment ---> Posttest

C1 Pretest ---------------------> Posttest

E2 Treatment ---> Posttest

C2 Posttest

Quasiexperimental Research Designs

n Nonequivalent control group design

– Useful when random assignment is not possible

– Relies on the use of comparative groups

n After-only non-equivalent control group design

– Two non-randomly assigned groups

(1 control, 1 experimental)

– Neither group is pretested

n Time series design

– Used when comparable groups are not available

– Uses experimental subjects as their own control

Pre-experimental Designs

n One shot case study

n One group pretest-posttest design

Evaluation Research

Purpose is to evaluate a program, treatment, program, practice, or policy

n Formative

– Assessment of program as it is being implemented

n Summative

– Assessment of program at completion of implementation

Research Designs

n Experimental

– Independent variable is manipulated

– Researcher manipulates 2 or more variables

n Nonexperimental Research Designs

– Independent variable is not manipulated

– Researcher explores relationships

e.g., factors that contribute to pain after surgery

Types of Nonexperimental Designs

n Survey studies

– Descriptive

– Exploratory

n Interrelationship studies

– Correlational studies

– Ex post facto studies

– Prediction studies

– Developmental studies

» Cross-sectional

» Longitudinal

Descriptive/Exploratory Survey Studies

n Collect descriptions of existing data

– Opinions (satisfaction with health care)

– Attitudes (AIDS, cancer, higher education))

– Facts (age, gender, religion)

n Purpose is to describe variables

n NOT to determine causation

Descriptive=Exploratory=Survey

n Terms used interchangeably

Descriptive/Exploratory Survey Studies

Data Collection Methods

n Questionnaire - structured

n Interview

– Structured

– Unstructured

Interview/Questionnaire

n Structured or close ended

– Likert

– Multiple choice

n Unstructured or open ended

– Tell me about your research course.

Surveys

n Advantages

– Great deal of information

– Economical

n Disadvantage

– Superficial

– Breadth rather than depth

– Requires great deal of expertise

– Time consuming and costly

Interrelationship Research Studies

n Examine relationships among variables

n Does NOT determine cause and effect

n If you are reading a correlational study and researcher concludes that one variable causes another, this is wrong

Interrelationship Studies

n Correlational

n Ex-post facto

n Prediction

n Developmental

– Cross sectional

– Longitudinal

Interrelationship Research Studies: Correlational Studies

n Determine whether the variables covary (as one changes, the other changes)

n Determine the amount of relationship between two or more variables (strength, and if positive or negative)

Interrelationship Research Studies: Ex Post Facto Studies

n Also called causal-comparative or comparative studies

n Used when experimental studies cannot be used

– Subjects who have experienced X (the independent variable) are compared with a control group

Interrelationship Research Studies: Prediction Studies

n Retrospective (data were collected previously)

– e.g., patient chart data

n Use retrospective data to make predictions about another group

– Predict success on NCLEX

Interrelationship Research Studies: Developmental Studies

n Cross-Sectional

– Collect data on only one occasion

– May be one or more groups of subjects

n Longitudinal

– Collect data on more than one occasion

– May be one or more groups of subjects

– Each subject followed separately

Retrospective Studies

n Retrospective Studies

– Dependent variable has already been affected by the independent variable

– Investigator attempts to link present events to past events

– Similar to ex post facto studies

Prospective Studies

n Prospective Studies

– Explore presumed causes or presumed relationships and then move forward in time to the presumed effect

– Similar to longitudinal studies

Methodological Research

n Develop instruments

n In other words, study of the methds used in doing research

n Psychometrics

– Measuring concepts with valid and reliable tools

Meta-analysis Research method

n Synthsizes findings of many studies to draw conclusions regarding what is currently known about an area of focus

THE END