Research Design for Quantitative Studies. Characteristics of Well Designed Research Clearly defined research question Clearly defined population under.

Research Design for Quantitative Studies

Characteristics of Well Designed Research Clearly defined research question Clearly defined population under study Clearly defined sampling procedure Clearly defined variables of interest Clearly defined statistical method Statement of the null hypothesis, Ho: Statement of the alternative hypothesis, Ha: Collection of sample under appropriate conditions Analysis and report

Research Design

Overall plan

– answer research questions

– test research hypotheses

Dimensions of Research Design

InterventionComparisonsControls for extraneous variablesTiming of data collectionResearch settingCommunication with subjects

Types of Research Design

Experimental Research

Quasi Experimental Research

Nonexperimental Research

Experimental Research

ManipulationControlRandomization

Experimental Designs

Basic Solomon Four-Group Factorial Repeated Measures Clinical Trials

Basic Experimental Designs

Posttest-onlyPretest-Posttest R=Randomization M=Measurement

X=Treatment

R X M-2

R M-1 X M-2

Solomon Four-Group Design

Data Collection

Group Before After

Experiment-pretest X X

Experiment-no pretest X

Control-pretest X X

Control-no pretest X

Factorial Design

Permits the testing of multiple hypotheses in a single experiment

Permit evaluation of main effects as well as interaction effects

Subjects randomly assigned to some combination of treatments

Independent variables known as factors Each factor has two or more levels

Factorial Design

Type of Stimulation

Auditory(A1) Tactile (A2)

15 min (B1) A1 B1 A2 B1

30 Min (B2) A1 B2 A2 B2

45 Min (B3) A1 B3 A2 B3

Repeated Measures Design

Also known as crossover design Exposure of same subjects to more than one

experimental treatment Participants must be randomly assigned to

different orderings of treatment Ensures highest possible equivalence among

subjects exposed to different conditions May have carry-over effects

Randomized Clinical Trials

Distinctive application of experimental design

Involves testing of clinical treatmentRandom assignment of subjectsFocused on outcomesLarge and heterogenous sample

Advantages of Experimental Research

Powerful method of testing cause-effect relationships

Disadvantages of Experimental Research

Many variables not amenable to experimental manipulation

Ethical considerations may prevent experimental manipulation

May be impracticalHawthorne (placebo) effect

Quasi-Experimental Research

Involves manipulation of independent variable

Lacks either or both– Randomization

– Control Group

Quasi-Experimental Designs

Non-equivalent control group design

Time series design

Nonequivalent Control Group Designs

Involves experimental treatment and two or more groups of subjects

Comparison group rather than control group

Quasi-experimental versus pre-experimental

Time Series Designs

One Group Data collection over extended time

period Experimental treatment during course

of data collection

Advantages and Disadvantages of Quasi-Experimental Design

Practical and feasible

Cannot make cause-effect inferences

Nonexperimental Research

No manipulative control of independent variables

Non-Experimental Designs

Ex Post Facto/Correlational research– Retrospective studies

– Prospective studiesDescriptive research

Advantages and Disadvantages of Non-experimental Research

Cannot infer causal relationshipsEfficient means of collecting large

amounts of data about problemRealistic

Additional Types of Quantitative Research

SurveyEvaluationNeeds AssessmentOutcomes Meta-analysis

Techniques of Research Control

External factors– constancy of

conditions– protocols

Techniques or Research Control

Intrinsic factors– Randomization– Homogeneity– Blocking– Matching– Statistical control

Threats to Internal Validity

History Testing Instrumentation Selection Maturation Mortality

External Validity

Generalizability of results– Validity– Reliability

Threats to External Validity

Hawthorne effect Novelty effect Interaction of history and treatment

effect Experimenter effect Measurement effect

Nursing Studies Are---

Usually nonexperimental Most commonly use correlational design More often retrospective than prospective More likely to use analysis of covariance and

similar techniques to control intrinsic factors Typically limited in terms of external validity

Tips for Enhancing Research Conrol

Strategies to control external factors Maximize constancy of conditions Strategies to control study participants Strong research design Measurement of demographic variables Identify extraneous variables via review

of literature