10-12-10 PPT - Chen
10-12-10 PPT - Chen
10-12-10 PPT - Chen
Research Method
EDHE6530
Dr. Pu-Shih Daniel Chen
Overview
• Review our progress
• Chapter 3 Research Methods
• Break
• How to review journal articles?
• A quick review of quantitative research
methods
A Quick Review
Chapter One
• Introduction/problem statement
• Purpose of the study
• Conceptual framework or theoretical
orientation
• Research questions and hypotheses
• Definition of terms
• Significance of the study
• Limitations, delimitations, & assumptions
Key to a good literature review
• Always keep your research questions in
mind
• Critically analyze the literature
• Integrate instead of summarize
• Use primary sources
• Distinguish between assertion and
evidence
Other tips
• Don’t ignore studies that differ from
majority or personal bias
• Read most recent ones first, oldest last
• Read important landmark works
• Use data-based, empirical studies
• Opinion pieces and descriptive research
helpful in introduction – set stage
• Not a literary production- be clear and
concise
Basic Research
• Theory Driven
• Discovery of knowledge
Applied Research
• IMMEDIATE practical problem
• ACTUAL problems in the field
• Limited generalizability
Research vs. Assessment
Research Assessment
• Generate knowledge that • Evaluate the progress or
can be applied broadly results of a initiative or
• Audience is not location- program
bounded • Internally focused
• Sample based • To improve the project or
process
• Population based
Quantitative vs. Qualitative
• Epistemology
• Treatment of Theory
• Role of the researcher(s)
• Data collection and analysis
• Quality controls
• Report of findings
• Types of studies
• How to choose between the two methods
Quantitative Methods
• Experimental Design
• Casual-Comparative Design
– Survey Research
– Secondary Data Research
• Descriptive Design
Qualitative Methods
• Ethnography
• Case Study
• Grounded Theory
• Narrative Inquiry
– Content Analysis
Ideological Approaches
• Researcher’s role
– Interpreting the phenomenon through a
particular perspective
– Identifying marginalized or endangered
people in society
– To advocate
• Challenging traditional views and usually
controversial
Mixed-Method
• Philosophically torn between positivist
and Phenomenology
• Pragmatism comes to save the day
• Focusing on “solving the problems”
instead of philosophical arguments
• Mixed-method is NOT half quantitative
and half qualitative
Discuss Your Research
Questions and Possible Method
Selections
Chapter 3
Research Methods
Purpose of the method chapter
• Make your research transparent
• Make it possible to other researchers to
replicate your study
• Make it possible for other researchers to
judge the quality of your study
Did you remember?
No Statistics Required
5’ 5’9” 7’
Example for CLT
Average Adult Male Height in the U.S.
5’4”
5’ 5’9” 7’
Example for CLT
Average Adult Male Height in the U.S.
6’6”
5’ 5’9” 7’
Example for CLT
Average Adult Male Height in the U.S.
5’10”
5’ 5’9” 7’
Example of CLT
Sample Height
X1 5’4”
Example of CLT
Sample Height
X1 5’4”
X2 6’6”
Example of CLT
Sample Height
X1 5’4”
X2 6’6”
X3 5”10’
Example of CLT
Sample Height
X1 5’4”
X2 6’6”
X3 5”10’
X4, X5,X6,…….,Xn 5’3”,6’9”…..
Average (X1-Xn) 5’9”
Example for CLT
Example for CLT
Average Adult Male Height in the U.S.
<20% 5’4”
5’ 5’9” 7’
What does p < .05 mean?
0.25
0.2
0.15
0.1
0.05
0 2 4 6 8 10 12 14 16 18 20
Null Hypothesis Significance
Testing (NHST)
• Originated from agricultural sciences
• Test the probability of null hypothesis
• If the probability for null hypothesis to
happen is less than a certain value (p<.05
or p<.01) we call it statistically significant
The Problems with NHST
• The size of p value does NOT indicate the
strength of the relationship
• Statistical significant does NOT imply
theoretical or practical significance
• Sample size has a great impact (if you
have enough power, you will always
reject null hypothesis)
• Does not work for population study
Solutions
• Reporting confidence interval (CI),
standard errors, and effect sizes
• Discuss theoretical and practical
importance using effect sizes
• Downgrade the importance of p value
Are they different
• T-test
• Analysis of Variance (ANOVA)
• Chi Square
• Wilcoxon signed-rank test
• Cluster Analysis
• Discriminant Function Analysis
What is the relationship
• Correlational Coefficients
– Pearson's product-moment coefficient
– Spearman's rank correlation coefficient
• Presumptions
– Randomness
– Linearity
What cause it to happen?
• Linear regression
• Multiple regression
– Multiple linear regression
– Logistic regression
– Ordered logistic regression
– Multinomial logit regression
What cause the whole thing to work
like that?
• Structural equation modeling
– MANOVA, MANCOVA
– Factor Analysis
– Path Analysis
• Hierarchical linear modeling
• Repeated Measure
– HLM growth model
– Time series
– GLM repeated measure
Common Mistakes
• Ignoring assumptions for a specific
method
• Conducting multiple t-tests without
Bonferroni correction
• Misusing Exploratory Factor Analysis
(EFA)
• Inferring casual relationship in a
correlational study
Other Mistakes
• Not reporting the quality measures of
your instrument
• Ignoring the problems of missing data
• Ignoring the problems of self-selection
bias
• Using statistical methods that you don’t
really understand
Remember