Transcript A Sound Research Project – Linking Program Needs and

A Sound Research Project –
Linking Program Needs and
Desired Outcomes
Caile E. Spear, Dept. of Kinesiology, Boise State University
[email protected]
Gayle Bush, Kinesiology & Health Promotion, Troy University
Ping Hu Johnson, Dept of Health, Physical Education and Sports Science,
Kennesaw State University
Michele Pettit, Dept. of Health Education & Health Promotion, UW-La Crosse
AAHPERD -March 17, 2010
Program Objectives:
By the end of this workshop, participants will be
able to:
1. Iterate steps in developing a research
proposal
2. Write succinct research hypotheses
3. Identify the appropriate research design
4. Select appropriate statistical methods and
analysis
Steps to Develop a
Research Proposal
• Decide what you want to do
o
Based on:
 Interest
 Knowledge and expertise
 Available resources
 personnel, equipment, materials, $$$, etc.
• Identify project goal(s)
o
o
What do you want to accomplish?
What is the problem that needs to be solved?
• Develop hypotheses
• Conduct literature review
o
o
Search literature
Organize literature
• Select research design
o
Depends on type of research
 Needs assessment, intervention, evaluation
o
Study Population vs. Study Sample
 Sample selection
• Select statistical methods
Literature Review
Subject/Title Search
Author Search
Identify possible articles
review titles and abstracts
Locate and obtain articles
library, online, interlibrary loan
Organize literature
o Provide
background information
 What has been done
 What needs to be done - need for research
 Why the need for research – justification/significance
o Identify
o Assist




theory/theories to guide research
with
Selection of research design and statistical methods
Selection or development of instrument for data collection
Development and implementation of intervention activities
Development and implementation of evaluation activities
Choosing a Theoretical
Framework
• Many health education projects are based
on specific theories or models.
• A framework is critical in planning a health
education or intervention project.
• Having a valid, reliable, and objective model
gives a research study credibility and a
basis for planning and evaluation.
Health Belief Model
Constructs:
o
o
o
o
o
o
Perceived susceptibility
Perceived severity
Perceived benefits of action
Perceived barriers to action
Cues to action
Self-efficacy
• Example: For a person to adopt recommended
physical activity behaviors, his/her perceived threat of
disease (and its severity) and benefits of action must
outweigh his/her perceived barriers to action.
Theory of Reasoned Action/Planned Behavior
Constructs• Attitude
• Perceived behavioral control
• Subjective norm
Example:
• Obese people who have a positive attitude towards
exercise, feel they can exercise, and have friends
thinking exercise is important, have positive intent
and are more likely to exercise
Social Cognitive Theory
Modeling
Skill Training (reasoning) – psychomotor
social skills (refusal skills) - behavioral rehearsal
Self-Monitoring - a contract with oneself
Contracting- contracting with others
•
•
•
•
Include a reward
Specific behaviors
Goals
Signatures
Example- Smoking cessation support groups
Stages of Change
Transtheoretical Model
People progress through 5 levels based on readiness to
change:
• Precontemplation
• Contemplation
• Preparation
• Action
• Maintenance
• Example- In adopting healthy behaviors (regular
physical activity) or eliminating unhealthy ones
(smoking, excessive alcohol intake), people cycle
through 5 stages
Socio-ecological Model: Steps to a Healthier US and other
community based health initiatives
The Precede-Proceed Model
Health Behavior Models
1. Health Belief Model
http://www.healthierus.gov/steps/2006Slides/A2/hefelfinger.html
2. Theory of Reasoned Action
3. Theory of Planned Behavior
http://www.etr.org/recapp/index.cfm?fuseaction=pages.TheoriesDet
ail&PageID=522#condomUse
4. Social Cognitive Theory
http://usaoll.org/mobile/theory_workbook/social_learning_theory.htm
5. Precede-Procede Model
http://envirocancer.cornell.edu/obesity/intervention101.cfm
6. Socio-ecological Model
http://www.ahrq.gov/clinic/uspstf07/methods/tfmethods.htm
7. Transtheoretical Model (Stages of Change)
http://www.aafp.org/afp/20000301/1409.html
Hypothesis
Formulating the Hypothesis
• A Hypothesis is the
expected result;
• It must be “testable”
• The study must be
designed in such a way
that the hypothesis can
be either supported or
refuted.
Research Hypothesis
• The anticipated outcome of a study or
experiment
• Must be based on some theoretical construct,
or on results from previous studies, or perhaps
on the researcher’s past experience and
observations
• For example:
o
“Children who participated in a 6-wk pedometer-based intervention
have higher daily step counts than children in the control group.”
Hypothesis Testing
• A scientific process that examines a
hypothesis against an alternative hypothesis
using appropriate statistical reasoning.
• Through the hypothesis testing, we infer the
findings from a sample to the population
(i.e., inferential statistics).
o
Using our sample statistic, we want to make a conclusion about
what is happening in the population.
Sampling
Study Population vs. Study Sample
• Study Population:
o share a common characteristic (age, sex, health
condition)
• Study Sample - a subset of the study population
• Sampling - methods of selecting a study sample
o Probability sample - allows for valid generalization
 simple - sampling unit (individual, natural group, etc.)
 systemic - nth
 stratified -proportional vs. nonproportional
o Non-Probability
Sample - limited generalizability
 Convenience
 Volunteers
 Grab samples
 Homogeneous samples
 Judgmental samples
 Snowball samples
 Quota samples
Research Design
Research Designs
• Non-experimental
o
o
No randomization
No comparison/control group
• Quasi-experimental
o
o
No randomization
Comparison/control group
• Experimental
o
o
Randomization
Control group
Source: Windsor et al., 1994
Inductive vs. Deductive Reasoning
Inductive
Deductive
• Hypotheses are
generated from
specific observations
and theories emerge
• A theory exists and
hypotheses are tested
using quantitative
methods
• Qualitative research
• Quantitative research
Source: Babbie, 2001
Qualitative vs. Quantitative Research
Qualitative
• Example:
o
RQ: What factors
contribute to binge
drinking among college
students?
Quantitative
• Example:
 RQ: Are gender and
Greek involvement
predictive of binge
drinking among
college students?
Statistics
Descriptive vs. Inferential Statistics
Descriptive
Inferential
• Describe a data set:
Demographics, Mean,
Range, Standard
Deviation, etc.
• Attempt to accurately
draw conclusions
about a larger
population based on
information collected in
a sample.
Examples of Inferential Statistics
• Correlation
• Regression
• T-tests
• ANOVA
Correlation
• Represents the strength of the
relationship or association between
two or more variables from the same
sample (values range -1 to 1)
• Example:
What is the relationship between
height and weight?
o RQ:
Regression
• Used to predict a variable (dependent/
outcome) from one or more predictor
(independent) variables
• Example:
o
RQ: Are attitude, subjective norm, and perceived behavioral
control predictive of college students’ intentions to quit
smoking?
o
This example utilizes the Theory of Planned Behavior which
has been used to examine individual behaviors and develop
programs.
“T” Tests (comparison of means)
• Used to draw conclusions/infer differences
in means (averages) between two
populations or sets of scores
• Examples
o
Repeated measures
o
Matched pairs
o
Post-test only between two groups with differing
interventions
Pre-Post Test Examples
• Pre – post test for knowledge, fitness levels,
attitudes, and specific behaviors.
• Examples:
o
Asthma 101 and Open Airways
o
Physical fitness: fall vs. spring
o
Attitudes and behaviors (the CATCH program related to diet and
exercise)
Tying Pieces Together
T test Examples
Example #1
• Example: Evaluation of a 1-day advocacy
training workshop for health educators
• Design: Non-experimental
• Research Question: Does a significant
difference exist between participants’
knowledge of advocacy before and after
the workshop?
• Methods: Pre/post-tests
• Statistical Analysis: Dependent t-test
Example #2
• Example: Evaluation of a comprehensive sex
education curriculum for 9th graders
• Design: Experimental
• Research Question:
Does the prevalence of unintended pregnancy differ
between students who complete a comprehensive sex
education curriculum and students who complete
an abstinence-based sex education curriculum?
• Methods: Post-tests
• Statistical Analysis: Independent t-test
ANOVA
• Used for more than two groups with repeated
measures such as a pre-mid-post test, or numerous
post tests after an intervention
Example:
1. Pre-test
****Intervention-9th grade sex education curriculum
2. Post-test
3. Nine month follow-up test
ANOVA (Cont.)
• Example: compare four physical education classes with
differing curricula or exercise programs
• Within-Group Variation–the amount of variation
among observations within each group (class, school,
gender, etc.)
• Between-Group Variation–the amount of variation
between all the group means
Summary
• HEDIR discussion on efficacy of abstinence
program
• Issue-can results be replicated
• Why?-many programs, what works in our
community
• Background of problem
– Teen pregnancy
– Variety of programs
– Efficacy
• Literature review -research-based,
theoretically based, factual, developmentally
appropriate, populations, short-term and
long-term outcomes
• Research question -Students in program
greater intent to remain abstinent vs regular
program
• Operational definitions - type of sex,
abstinence-only, abstinence-based
• Data analysis
• Results
Project Ideas
• Think-Pair-Share
– Premise lit review & theory selection done
• Identify research question
• Generate hypothesis
• Sample
• Methods
• Data collection
• Data analysis
– Who needs to be on board