Research Designs
Andrea M. Landis, PhD, RN
UW LEAH
Learning Objectives
Discuss concepts important to research design
Identify threats of internal validity
Review different types of non-experimental, experimental,
and quasi-experimental research designs
Research Design: Definition and
Characteristics
The vehicle for hypothesis testing or answering research
questions
A blueprint for conducting a study
Maximizes control over factors that could negatively effect
the validity of study findings
Guides the researcher in planning and conducting a study
Links the steps of the research process in the study
Concepts Important to Research
Design
Causality
Cause is not directly observable but must be observed
The cause is necessary for the effect to occur
Multicausality – recognition that a number of interrelating
variables can be involved in causing a particular effect
Probability – Addresses the likelihood that something will happen
in a given situation
Bias
To slant away from the truth or the expected
Failing to consider or include both sides of the question or hypothesis
Control – A check or comparison. Methods to keep the study
conditions constant during the study
Forms of Control
Manipulation – Researcher exercises by specifying the IV
Elimination or Inclusion – Holding certain aspects of
intervening and extraneous variables constant
Statistical – Controlling extraneous variables by including
them in the statistical analysis
Randomization – Distribution of effects of extraneous
variables via change with assignment of subjects to groups
based on probability
What is the difference between random sampling and random
assignment?
Concepts Important to Research
Design
Study Validity – truth or accuracy of the study findings.
Internal Validity – extent to which the effects detected in the
study are a true reflection of reality.
External Validity – extent to which the findings of the study
can be generalized to the general population
Threats of Internal Validity
History: due to intervening events between pre- and post
test
Maturation: produced by changes in members in one group
which occurred at a different rate than in the comparison
group between data points
Testing: created by repeated measurement
Instrumentation: produced by a change in the measuring
instrument between the pre- and post-test
Selection: differences between the kinds of people in an
experimental group in comparison to the other(s). Due to
lack of random placement of subjects into two groups.
Threats of Internal Validity
Mortality: due to differences in those who dropped out of a
particular treatment group versus the comparison group(s)
Ambiguity About the Direction of Causal Influence: occurs when
cause and effect variables are measured at the same time (e.g.,
correlational studies)
Diffusion or Imitation or Treatments: spurious communication of
the treatment to the control group(s)
Compensatory Equalization of Treatments: when the control
group receives the treatment inadvertently because it is seen by
administrators or health care providers to be best for patients
Reactive Effect: produced by a data collector or a subject’s
response to being in a study which improves subject performance
or behavior
Types of Research Designs
Non-experimental – both randomization and manipulation
absent
Experimental – both randomization and manipulation
present
True or classic experiment
Quasi-experimental – manipulation present, but not
randomization
One-group (pretest – posttest) design
Major Categories of Non-experimental
Designs
Descriptive
Designed to document conditions, attitudes, or characteristics
of individuals or groups
Exploratory
Focuses on the relationships among these factors
Predictive
Aimed at the development of systems to predict criteria of
interest by utilizing information from one or more predictors
Explanatory
Aimed at testing of hypotheses formulated to explain
phenomena of interest. Involves theoretical model testing.
Methods of Non-Experimental
Research
Retrospective (ex post facto)
Involves examining data that have been collected in the past, often
obtained from medical records or survey
Prospective
Variables are measured through direct recording in the present
Longitudinal
Follows a cohort of subjects over time, performing repeated
measurements at prescribed intervals
Cross-sectional
Researcher studies a stratified group of subjects at one point in time
and draws conclusions about development within a population by
comparing the characteristics of those strata.
Perspectives in Qualitative Research
Designs
Phenomenology
Seeks to draw meaning of experiences through narrative subject
materials. Words like “lived experience” often describe
phenomenological studies.
Ethnography
Study of the social milieu of a specific cultural group or people.
Researcher often immersed in subject’s way of life.
Grounded Theory
Researcher uses data to develop a theory that will explain what
is observed. Researcher collects, codes, and analyzes data
simultaneously.
Epidemiological Research
Concerned with the study of the distribution of disease,
injury, or dysfunction in human populations
Observational Epidemiologic Studies
Gather measures about disease frequency: prevalence (existing
cases), incidence (new cases)
Analytic Epidemiologic – Used when enough is known about
a condition to allow testing of hypotheses about the
association of specific risk factors (exposures) and outcomes
Case-control studies – groups of individuals are selected on the
basis of whether they have the disorder under study
Cohort studies – group of individuals followed over time to
determine if they will develop a disorder
Nontraditional Designs: Examples
Methodological Designs
Used to develop research approaches or the R/V of instruments
to measure constructs used as variables in research
Secondary Analysis
Studying data previously collected in another study
Meta-Analysis Designs
Involves merging findings from many studies that have
examined the same phenomenon
Levels of Evidence