Desain Penelitian

Research
Designs
(desain penelitian)
Copyright © 1999
by W.B.Saunders Company. All rights reserved.
Quantitative Research
Increased
• Descriptive Control
• Correlational with
Design
• Quasi-experimental
• Experimental
Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Research Design
• Definition: Blueprint or detailed
plan for conducting a study
• Purpose, Review of Literature, &
Framework
Study Design
Design Characteristics
• Maximizes control over factors to
increase validity of the findings
• Guides the researcher in planning &
implementing a study
• Not specific to a particular study but
is linked to other steps of the research
process
Concepts Relevant to Design
• Causality: A B
Pressure Ulcer
• Multicausality:
Years Smoking
Heart
High-Fat Diet
Disease
Limited Exercise

• Probability: likelihood of
accurately predicting an event
• Bias: slanting findings away
from the truth
• Manipulation: implementation
of a treatment
• Control
• Implemented throughout the design
• Improve accuracy of findings
• Increased control in quasi-
experimental research
• Greatest in experimental research
Varying Control in Studies
Quasi-experimental Experimental
Lower Control High Control
Type of Sample Selected:
Convenience ----------|--------> Random

Type of Sample Selected:
Heterogeneous -----|----> Homogeneous

Measurement
of Dependent Variable
Crude Precise
Measurement------|----> Measurement

Control
of Independent Variable
Limited or Highly
no Control ----------|--------> Controlled

Group
no comparison group ----> comparison
group ----> alternative treatment
control group ----> no treatment control
group

Selection of Groups
No randomization ---------> Randomization

Setting Selected
Highly
Natural -----------|----------> Controlled

Design Validity
• Measure of truth or accuracy of a
study examined with critique &
controlled in conducting a study
• Statistical conclusion validity
• Internal validity
• Construct validity
• External validity
Types of Design Validity
• Statistical Conclusion Validity --
is concerned with whether the
conclusions about relationships or
differences drawn from statistical
analyses are an accurate reflection
of the real world.

Statistical Conclusion Validity
• Low statistical power: Type II error,

no significant difference when
there is.
• Violated assumptions
• At least interval data
• Random sample
• Normal distribution
• Reliability of Instruments:
consistency in measurement
• Physiological measures: Precision,
such as with temp., BP
• Likert scale: Homogeneity--Reliability
with Cronbach’s alpha (.7, .8, .9)
• Observations: Stability--Interrater
reliability

• Reliability of Treatment
• Protocol for treatment
• Training of researchers
• Interrater reliability: # of items
completed in protocol/ # of items
on protocol (%)

• Internal Validity --
is the extent to which the effects
detected in the study are a
true reflection of reality,
rather than the result of
extraneous variables.
Internal Validity
• History: event impacting subjects
& data collected
• Maturation of subjects: wiser, tired,
older influencing responses
• Testing: learn from being in study
• Instruments’ accuracy

Internal Validity
• Subject selection nonrandom
& potentially biased
• Mortality: those dropping out
of study
• Diffusion or imitation of treatment:
control or comparison group
has access to treatment
• Construct Validity--
examines the fit between the
conceptual definitions and
operational definitions of
variables. Link of framework
concepts & variables studied.
Construct Validity
• Validity of instruments
• Physiologic measure: Accuracy
• Scales: construct validity
• Mono-operation bias: 1 instrument
to measure a variable
• Mono-method bias: 1 type of
instrument, such as attitude scales
Construct Validity
• Subjects guessing outcome of study
& acting in a different way
• Subjects performing for experiment
to be seen in a good light or trying to
get an effective treatment
• Treatment interactions with more
than one treatment
External Validity---
is concerned with the extent to
which study findings can be
generalized beyond the sample
used in the study.

External Validity
• Interactions of:
• Selection of subjects & treatment
• Setting & treatment: closing clinic,
nonsupportive staff
• History & treatment: events
influence treatment

Elements of Strong Design
• Controlling the Environment:

selection of the study setting
• Natural setting
• Partially controlled setting: clinic

• Highly controlled setting: lab
• Controlling the Equivalence of

Subjects & Groups
• Random subject
selection
• Random assignment
to groups

• Controlling the Treatment (Tx)

• Tx based on research & practice
• Protocol developed for implementation
• Document Tx was implemented
• Checklist to determine extent Tx
was implemented if incomplete
• Re-evaluate Tx during study

• Controlling Measurement
• Reliability
• Validity
• # Measurement methods
• Types of instruments

• Controlling Extraneous Variables

• Identify & eliminate by sample
criteria, setting, design
• Random sampling
• Sample: heterogeneous, homogeneous,
matching
• Statistical control

Problems with Study Designs
• Inappropriate for purpose

& framework
• Poorly developed
• Poorly implemented
• Inadequate treatment, sample,
measurement methods

Critiquing a Study Design
1. Was the type of design identified?

2. Was the study design linked to the
purpose &/or objectives, questions,
or hypotheses?
3. Were all variables manipulated
or measured?

4. If the study included a treatment, was it
clearly described & consistently
implemented?
5. Were extraneous variables identified &
controlled?
6. What were the threats to design validity in
this study?

7. Was a pilot study performed prior

to this study? What was the reason
for the pilot & the outcome?
* Study feasibility
* Refine design or treatment
* Examine validity & reliability
of measurement methods

Identifying a Design
in a Quantitative
Research Report

Study Purpose
• Describe Variables
• Examine Relationships
• Determine Differences
• Test a Treatment
• A Combination of these
Types of Designs
• Descriptive Study Designs
• Correlational Study Designs
• Quasi-Experimental
Study Designs
• Experimental Study Designs
Identifying a Design
Is there a treatment?
No Yes
Is the primary purpose Is the treatment tightly controlled

examination of relationships? by the researcher?
No Yes No Yes
Descriptive Will the sample be Quasi-Experimental Will a randomly assigned

Design studied as a single Study control group be used
group?
No Yes
No Yes
Is the original sample
Correlational randomly selected?
Design
No Yes
Experimental
Study
Identifying a Descriptive Design
Examining sequences across time?
No Yes
One Group? Following same subjects across time?
No Yes No Yes
Comparative Descriptive Data collected Single unit

Descriptive Design across time of study
Design
No Yes No Yes
Cross-sectional design Studying events partitioned Longitudinal Case

across time? Study Study
No Yes
Trend Analysis Repeated measures of each subject
Yes No
Cross-sectional design Longitudinal design with

Copyright © 1999 by W.B.Saunders Company. All rights reserved. with treatment partitioning treatment partitioning
Descriptive Study Designs
• Typical Descriptive Design

• Comparative Descriptive Design
• Time Dimensional Design
• Case Study Design

Diagramming the Design
• Clarifies the variables to be
measured or manipulated
• Indicates focus of the study:
description, relationships,
differences, &/or testing a
treatment
Diagramming the Design
• Identifies data collection process:
time for study, treatment
implementation, measurement
of variables
• Provides direction to
data analysis
Typical Descriptive Design
Clarification Measurement Description Interpretation

Description
Variable 1 of
Variable 1
Description
Variable 2 Interpretation
Phenomenon of
Variable 2 of Meaning
of Interest
Description
Variable 3 of
Variable 3
Development of
Description Hypotheses
Variable 4 of
Variable 4

Comparative Descriptive Design
Group I
{variable(s) Describe
measured}
Comparison of
Interpretation
Groups on
of Meaning
Selected Variables
Group II
{variable(s) Describe Development
measured}
of Hypotheses

Correlational Study Designs
• Descriptive
Correlational Design
• Predictive Design
• Model-Testing Design

Identifying Type of
Correlational Design
Describe Predict Test
relationships relationships theoretically
between/among between/among proposed
variables? variables? relationships?
Descriptive Predictive Model

correlational correlational testing
design design design

Descriptive Correlational Design
Measurement
Description
Research
of variable Interpretation
Variable
of Meaning
1
Examination of
Relationship
Research Description
Variable of variable Development
2
of Hypotheses

Predictive Design
Value of Value of Predicted

Value
Independent Value of
of
Intercept
+ Independent
Variable + Variable = Dependent
1 2 Variable

Study Groups
• Groups in comparative
descriptive studies
• Control group
• Comparison group
• Equivalent versus
nonequivalent groups
Pre-Experimental Designs
• One-group post-test–only design

• Post-test–only design with
nonequivalent groups
• One-group pretest–post-test
design

Quasi-Experimental Designs
• Untreated control group design

with pretest & post-test
• Nonequivalent dependent
variables design
• Removed-treatment design
with pretest & post-test
Pretest–Post-Test Control
Group Design
Measurement Manipulation of Measurement
of dependent independent of dependent
variables variable variable(s)
Randomly selected PRETEST TREATMENT POST-TEST

experimental
group
Randomly selected PRETEST POST-TEST

control group
Treatment: Under control of researcher
Findings: comparison of pretest and post-test scores

comparison of experimental and control groups
comparison of pretest-post-test differences between samples
Example: Your self (1990). The impact of group reminiscence counseling on a depressed elderly
population
Uncontrolled testing mortality
threats to validity: instrumentation restricted generalizability as control increases

Post-Test–Only
Control Group Design
Measurement Measurement
of independent of dependent
variables variable(s)
Randomly selected TREATMENT POST-TEST

experimental
group
Randomly selected POST-TEST

control group
Treatment: Under control of researcher
Findings: comparison of experimental and control groups
Example: Clochesy, Difani & Howe (1991). Electrode site preparation techniques: a
follow-up study
Uncontrolled instrumentation
threats to mortality
validity: limited generalizability as control increases

Advantages of Quasi-
Experimental Designs
• More Practical: Ease
of implementation
• More Feasible: resources,
subjects, time, setting
• More Generalizable:
Comparable to practice
Advantages of
Experimental Designs
• More Controls: design
& conduct of study
• Increased internally valid:
decreased threats to
design validity
• Fewer Rival Hypotheses
Outcomes Research
• Agencies Supporting Outcomes
Research
• Agency for Health Services
Research (AHSR)
• Agency for Health Care Policy
& Research (AHCPR)

Outcomes Research
• Structure of Care
• System Characteristics
• Provider Characteristics
• Patient Characteristics

Outcomes Research
• Process of Care
• Ordering tests
• Treatments: pharmacological
& nonpharmacological
• Referral
• Hospitalization
• Counseling & education

Outcomes Research
• Outcomes
• Clinical end points
• Functional status
• General well-being
• Satisfaction with care

Critiquing A Design
• Identify elements that were
controlled in the study
• Identify possible sources of bias
• Are there elements that could
have been controlled to improve
the study design?
Critiquing A Design
• What elements of the design were
manipulated and how were they
manipulated?
• How adequate was the manipulation?
• What elements should have been
manipulated to improve the validity of
the findings?

Critiquing A Design
• Based on your assessment of

the adequacy of the design,
how valid are the findings?

Critiquing A Design
• Is there another reasonable

(valid) explanation (rival
hypothesis) for the study
findings other than that
proposed by the researcher?

Questions

Desain Penelitian

Uploaded by

Copyright:

Available Formats

Desain Penelitian

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Desain Penelitian

Uploaded by

Copyright:

Available Formats

Research

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Type of Sample Selected:

Convenience ----------|--------> Random

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Type of Sample Selected:

Heterogeneous -----|----> Homogeneous

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

No randomization ---------> Randomization

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• Low statistical power: Type II error,

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• Controlling the Environment:

• Partially controlled setting: clinic

• Controlling the Equivalence of

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• Controlling the Treatment (Tx)

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• Controlling Extraneous Variables

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• Inappropriate for purpose

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

1. Was the type of design identified?

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

7. Was a pilot study performed prior

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Is the primary purpose Is the treatment tightly controlled

Descriptive Will the sample be Quasi-Experimental Will a randomly assigned

One Group? Following same subjects across time?

Comparative Descriptive Data collected Single unit

Cross-sectional design Studying events partitioned Longitudinal Case

Trend Analysis Repeated measures of each subject

Cross-sectional design Longitudinal design with

• Typical Descriptive Design

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Clarification Measurement Description Interpretation

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Descriptive Predictive Model

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

Value of Value of Predicted

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• One-group post-test–only design

Copyright © 1999 by W.B.Saunders Company. All rights reserved.

• Untreated control group design