Module 2
Module 2
Study design
Provides a plan for answering the research question
Includes plan for study setting, participant selection, study procedures, variable measurement, data
collection, data management, data analysis
Considerations
Intervention
Comparisons
Potential confounding variables (other variables that could be the explanation for what you are
finding; if so, can you control them?)
o Controlling study context
o Controlling participant factors (randomization, homogeneity, matching, statistical control)
Blinding
Data collection times –
o Cross-sectional (all at one time)
o Longitudinal (longer, change over time)
Relative timing – prospective vs retrospective
Location
Classifications
Experimental
Quasi-experimental
Nonexperimental
Experimental designs
Requirements:
o Intervention
o Control
o Randomization
Experimental group
Control group
Examples:
Advantages
o Testing cause-and-effect relationships
Disadvantages
o Control vs practical significance
o Generalizability
o Need to randomize participants
Quasi-experimental designs
Requirements:
o Intervention
o Control
Examples:
Advantages
o Practical
o More acceptable to participants
o Can use when it’s unethical to randomize participants
Disadvantages
o More difficult to make causal inferences
Nonexperimental designs
Requirements:
o None
Types:
o Correlation studies - Study relationships between variables that are not manipulated;
correlation does not prove causation
Cohort design
Prognosis or etiology questions
Start with exposures and follow over time to see what outcome is
Case-control design
Look at past to see what exposures were
o Descriptive studies – Observe, describe, and document a phenomenon
Advantages
o Very practical
o Efficient way to collect large amounts of data
Disadvantages
o Cannot make causal inferences
o Self-selection
Sample size
Power – How many subjects are needed for research
o Sample size
o Attrition – Loss of people
Setting
Laboratory vs real setting
Data collection
Existing records vs original data
o Pros: Easier, can look back in charts
o Cons: Can’t add additional information (documentation, variable, etc.)
Original data
o Self-report
Interviews (get more data most likely) vs questionnaires (good for sensitive topics)
Open-ended vs close-ended
Investigator developed vs validated tools
o Observation
Categories
Rating scales
Observational sampling
o Biophysiological measures
In vivo: In a living person (BP, SpO2)
In vitro: Removed from person and tested later (blood work)
Variable data:
Inferential statistics
Hypothesis testing
o Develop null and research hypotheses
o Choose a level of significance
o Determine which statistical test is appropriate
What is the research question/hypothesis?
Identify the independent and dependent variables
Identify what levels of measurement the variables were measured
Identify whether a parametric or non-parametric analysis is appropriate
o Run analysis to obtain test static and p value
o Make decision about whether to reject or fail to reject the null hypothesis
o Make a conclusion