Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
5/5/2009
1
Diane E. Beck, Pharm.D.
Interim Associate Dean for Curriculum and AssessmentInterim Associate Dean for Curriculum and Assessment & Professor
College of Pharmacy University of Florida
Introduction
Research that uses a systematic
h d l l i
2
methodology to explain educational problems and issues.
5/5/2009
2
Compare inductive and deductive approaches to research. Identify when to use an inductive versus an deductive approach.
Explain the term “causation” Describe why research designs that investigate causation are
“confirmatory”.
Describe the common educational research designs. Discuss when it is appropriate to use the design and skills needed
b th h t th d iby the researcher to use the design.
For Projects Proposed by Med Ed Fellows: Identify the type of design needed and “why” the design is
appropriate.
Identify attributes of effective curricular resources.
Goal:Identify irrefutable facts & i l t th
Deductive Inductive
Goal:New insights from data ll t d/ b ti& universal truths
Make Predictions Based on General
Principles
Drawgeneralizations from a limited
collected/observations
4
Common in Educational Research
Typical of Clinical/ Laboratory Research
number of observations
5/5/2009
3
Deductive Inductive
Develop a Conceptual Framework via Literature
Search5
Probable Cause
EffectCause
Independent Variables
Dependent Variables
Explained by one of the following:• Theory: Connected hypotheses and variables identified by
the authors. (Explains & Predicts)• Conceptual Framework: Visual model described by other
6
authors; structure from their “lens”• Theoretical Rationale: Rationale proposed by other
authors based on studies about relationships• Hunch
5/5/2009
4
Quantitative Research Designs
Qualitative Research Designs
Combined Qualitative and
Quantitative D iDesigns
Deductive
Inductive
Both
Investigators ask a narrow, specific question, collects quantitative data, and uses statisticsto analyze data.
Investigators ask broad, general questions, rely on the views of participants, collects data consisting of “words”, and analyzes the data for “themes.”
Both quantitative and qualitative methods are used
Is the research confirmatory?
Do you want to evaluate
Causation?
Use Inductive/
Qualitative methods
NoYes
Can you randomize?
Use randomized
Use quasi-experimental
Use correlational
methods
Example 1: e planator
Example 1:Interviews;
focus groups
Example 2:
Yes No
NoYesNote: If you want to use a quasi‐experimental method, but can not assign subjects to a control group, a do ed
experimental methods
Example: RCT
experimental
Methods*
Example: Pre-test post-test
explanatory correlations
Example 2:
Surveys
a p eObservation
Example 3: Surveys – open
ended questions
8
g p,use “non‐experimental methods” such as case‐control, cross‐sectional, time series, and cohort studies.
Deductive Inductive
5/5/2009
5
Is the research confirmatory?
Do you want to evaluate
Causation?
Use Inductive/
Qualitative methods
NoYes
Can you randomize?
Use randomized
Use quasi-experimental
Use correlational
methods
Example 1: e planator
E xample 1:: Interviews;
focus groups
Example 2:
Yes No
NoYes
randomized experimental
methods
Example: RCT
experimental
Methods*
Example: Pre-test post-test
explanatory correlations
Example 2:
Surveys
a p eObservation
Example 3: Surveys – open
ended questions
9
Note: If you want to use a quasi‐experimental method, but can not assign subjects to a control group, use “non‐experimental methods” such as case‐control, cross‐sectional, time series, and cohort studies.
“The Gold Standard” if you want to evaluate “causation” Confirmatory
When to Use: You have a hypothesis to test (supported by a
theory, etc). You can randomize subjects to treatment andYou can randomize subjects to treatment and
control groups.
Skills Needed by the Researcher: Quantitative Research Methods Statistics
5/5/2009
6
The Most Common Design in Education RCT not possible when there are logistic issues/not
practical to randomize
Examples: ○ Non-equivalent control groups
○ Pre-test post-test design
When to Use: You have a hypothesis to test (supported by a theory, etc).
You can NOT randomize subjects to treatment and control groups.
Skills Needed by the Researcher: Quantitative Research Methods
Statistics
Is the research confirmatory?
Do you want to evaluate
Causation?
Use Inductive/
Qualitative methods
NoYes
Can you randomize?
Use randomized
Use quasi-experimental
Use correlational
methods
Example 1: e planator
E xample 1:: Interviews;
focus groups
Example 2:
Yes No
NoYes
a do edexperimental
methods
Example: RCT
experimental
Methods*
Example: Pre-test post-test
explanatory correlations
Example 2:
Surveys
a p eObservation
Example 3: Surveys – open
ended questions
12
Note: If you want to use a quasi‐experimental method, but can not assign subjects to a control group, use “non‐experimental methods” such as case‐control, cross‐sectional, time series, and cohort studies.
5/5/2009
7
Examples: Frequently used to correlate academic success with
admission criteria.○ Does not involve an intervention
When to Use: You want a confirmatory approach but DO NOT want to
prove causation
When you want to show how variables are “associated”
Skills Needed by the Researcher: Quantitative Research Methods
Statistics
Is the research confirmatory?
Do you want to evaluate
Causation?
Use Inductive/
Qualitative methods
NoYes
Can you randomize?
Use randomized
Use quasi-experimental
Use correlational
methods
Example 1: e planator
E xample 1:: Interviews;
focus groups
Example 2:
Yes No
NoYes
a do edexperimental
methods
Example: RCT
experimental
Methods*
Example: Pre-test post-test
explanatory correlations
Example 2:
Surveys
a p eObservation
Example 3: Surveys – open
ended questions
14
Note: If you want to use a quasi‐experimental method, but can not assign subjects to a control group, use “non‐experimental methods” such as case‐control, cross‐sectional, time series, and cohort studies.
5/5/2009
8
Use in Education: Frequently used in education; does not involve an intervention.
Usually hypothesizes variablesy yp
When to Use: When you DO NOT want to evaluate causation
When you want to document variables and their relationships
Surveys can also be “exploratory” – esp if open ended questions are used
Skills Needed by the Researcher: Survey Design – the survey is your “tool”; validity is essential
Quantitative Research Methods
Statistics
A survey involves more than giving subjects a “questionnaire”questionnaire See Recommended
References and seek an expert on item writing
5/5/2009
9
Is the research confirmatory?
Do you want to evaluate
Causation?
Use Inductive/
Qualitative methods
NoYes
Can you randomize?
Use randomized
Use quasi-experimental
Use correlational
methods
Example 1: e planator
Example 1:Interviews;
focus groups
Example 2:
Yes No
NoYes
randomized experimental
methods
Example: RCT
experimental
Methods*
Example: Pre-test post-test
explanatory correlations
Example 2:
Surveys
a p eObservation
Example 3: Surveys – open
ended questions
17
Note: If you want to use a quasi‐experimental method, but can not assign subjects to a control group, use “non‐experimental methods” such as case‐control, cross‐sectional, time series, and cohort studies.
Use in Education:A. Interviews (focus groups, individual interviews); observation; think
aloud; and simulated recall are methods used to gather information or “observations” from individuals to identify related factorsor “observations” from individuals to identify related factors.A. Grounded theory research
B. Ethnographic Research
C. Narrative Research
When to Use: A. Little is known about the problem – “Exploratory”
B There is a need to understand what factors or variables contributeB. There is a need to understand what factors or variables contribute to the problem or issue.
C. Note: These are not confirmatory
Skills Needed by the Researcher: Qualitative research skills : In-depth recording, analysis of words using
special software & triangulation of data; Inductive thinking abilities
5/5/2009
10
Explore the life of an individualBiography
PhenomenologyUnderstand the essence of
experiences about a
Inductive (Qualitative) Research Designs
Ethnographic Explore the shared l f
Grounded Theory Research
Phenomenology experiences about a phenomenon (eg, learning)
Develop a theory grounded in data
from the field
Case Study
g pResearch culture of a group
of people
Develop an in-depth analysis of a single case
or multiple casesRefs: Creswell JW. Qualitative Inquiry and Research Design, 1998 and Educational Research 3rd Ed, 2008
Is the research confirmatory?
Do you want to evaluate
Causation?
Use Inductive/
Qualitative methods
NoYes
Can you randomize?
Use randomized
Use quasi-experimental
Use correlational
methods
Example 1: e planator
E xample 1:: Interviews;
focus groups
Example 2:
Yes No
NoYes
Sometimes Both Inductive and Deductive Methods are
Inductivedeductive
a do edexperimental
methods
Example: RCT
experimental
Methods*
Example: Pre-test post-test
explanatory correlations
Example 2:
Surveys
a p eObservation
Example 3: Surveys – open
ended questions
20
Note: If you want to use a quasi‐experimental method, but can not assign subjects to a control group, use “non‐experimental methods” such as case‐control, cross‐sectional, time series, and cohort studies.
Used in the Same Project
5/5/2009
11
Use in Education: A combination of inductive and deductive methods are
used such as:○ Begin with an inductive design and then do a deductive designg g g
○ Begin with a deductive design and then do an inductive design
When to Use: Methods are usually not used for evaluating “causation”
Used when you want to identify and document variables and their relationships
Surveys can also be “exploratory” – esp if open ended questions Su eys ca a so be e p o ato y esp ope e ded quest o sare used
Skills Needed by the Researcher: Survey Design – the survey is your “tool”; validity is essential
Quantitative Research Methods
Statistics
Study Design Issues related to Internal
validity Issues related to External
validity
Instrument Design
Potential “Threats to validity “ in Educational Research include:• History• Maturation• Mortality• Selection
22
Instrument Design Issues related to Construct
validity
Data Analysis Issues related to Statistical
Conclusion validity
Instrument Design is very important in educational research
5/5/2009
12
Notes about IRB Approval: Always safer to submit – even if it will be
“exempt”
Go the easy road…..Many institutions with a Health Science Center have an IRB office that handles medical research and another that handles “othermedical research and another that handles other research.”
○ At UF, submit to IRB2 if your project does not involve data that falls under HIPAA: http://irb.ufl.edu/irb02/
23
Envision, Plan, Implement, and Write your project so that it will be accepted by Editors. Start now…….“Do it the right way”
24
5/5/2009
13
Common reasons for rejection in RIME proceedings: Problem statement - 17%
Relevance – 5%
Research Design – 6%
Sample and Sampling – 10%
Instrumentation and Data Collection – 14%
Results – 20%
Discussion and Conclusion – 3%
Title – 3%
Abstract – 2%
Writing/Presentation – 9%
25Bordage G. Academic Medicine. 2001;76;889-896.
26
5/5/2009
14
(Important if your Intervention involves Curricular Resources)
Develop resources based on “How People Learn”Learn” Reference: Bransford JD, Brown AL, Cocking RR (eds). How People
Learn: Brain, Mind, Experience, and School. National Academies Press, 1st Ed., 2000. Available at: http://www.nap.edu/openbook.php?record_id=9853&page=R1
Good URL to start: http://www.vanderbilt.edu/cft/resources/teaching_resources/theory/HPL.htm
Other URL: http://www.newhorizons.org/neuro/neu_review_bransford.htm
See Instructional Design Guides at: http://ets.tlt.psu.edu/learningdesign/
5/5/2009
15
1. Title Page with Abstract (150 words) Abstract should succinctly outline: specific aim
(research question and hypothesis), background/ significance experimental design and methodssignificance, experimental design and methods.
2. Specific Aims
3. Background Literature review that provides the background and
convinces the reader that the project is important/addresses a relevant problem. p p
4. Preliminary work – if applicable If you have already done work in this area, describe
it.
See Next Slide….
5. Experimental Design/Methods Research design, population & sample, power
Continued from Prior Slide
calculation – if possible, intervention, procedures, measurable outcomes, instrumentation (include validity & reliability), data collection & quality control, data analysis
6. Timeline for Proposed Activities
7 Literature Citations7. Literature Citations
8. Appendix IRB approval - Documents for IRB submission
5/5/2009
16
Establish a “conceptual framework” and determine whether inductive and/or deductive approaches are most appropriate.most appropriate.
Select the best design given practical considerations; if using confirmatory designs, anticipate “threats to validity.”
Instruments (eg. exams, surveys) are your “tools” make s re the ha e e idence of“tools”….make sure they have evidence of “validity.”
Before You Begin….Learn reasons why reviewers accept/reject educational manuscripts. Essential Reading: Academic Medicine 2001 Issue
1. Core Reference: A. Creswell JW. Educational Research: Planning, Conducting, and
Evaluating Quantitative and Qualitative Research. Merrill P ti H ll U S ddl Ri NJ 2008 ISBN 0 13Prentice Hall, Upper Saddle River, NJ, 2008, ISBN-0-13-613550-1.
B. Trochim WMK. Web Center for Social Research Methods. (See Knowledge Base link to purchase as a Text). Accessed August 1, 2008, http://www.socialresearchmethods.net/
2. More In-depth References:A. Campbell & Stanley, Experimental and Quasi-experimental
Designs for Research, 1963, Houghton-Miflin, Boston, MA, SISBN 0-395-30787-2
B. Cook and Campbell, Quasi-experimentation: Design & Analysis Issues for Field Settings, 1979, Houghton-Miflin, Boston, MA, ISBN 0-395-30790-2 3-11600
C. Shadish, Cook and Campbell, Experimental and Quasi-experimental Designs for Generalized Causal Inference, Houghton-Miflin, Boston, MA, ISBN-13 978-0-395-61556-0.
32
5/5/2009
17
3. Survey Research: A. Draugalis JR, Coons SJ, Plaza CM. Best practices for survey research
reports: A synopsis for authors and reviewers. Am J Pharm Educ. 2008;72:Article 112008;72:Article 11.
B. Fowler FJ. Survey Research Methods. 3rd ed. Thousand Oaks, CA: Sage Publications, 2002.
C. Scheuren F. What is a Survey? Available at: http://www.whatisasurvey.info/ Accessed February 1, 2009.
D. American Association for Public Opinion Research. Best Practices for Survey Research and Public Opinion Research. Available at: http://www.aapor.org/bestpractices Accessed: February 1, 2009.p p g p y ,
E. Doyle JK. Introduction to Survey Methodology and Design. Available at: http://www.sysurvey.com/tips/introduction_to_survey.htm Accessed: February 1, 2009.
33
4. What Editors Look for: A. Review Academic Medicine, September 2001
Issue (free access):
http://journals.lww.com/academicmedicine/pages/toc.aspx?year=2001&issue=09000
5. Development of a Conceptual p pFramework is “key to success”:
A. Reference: Bordage G. Conceptual frameworks to illuminate and magnify. Medical Education 2009: 43: 312–319.
34
5/5/2009
18
Thank You !!!You !!!
Diane E. Beck, Pharm.D.
Phone: 352-273-6279
University of Florida
College of Pharmacy