STATISTICAL RESEARCH AND TRAINING CENTER
J and S Building, 104 Kalayaan Avenue, Diliman, Quezon City
Training Course on
Statistical Analysis with STATA
May 16 - 20, 2011
Introduction to Statistical Analysis
Jose Ramon Toots G. Albert, Ph.D.
Senior Research Fellow, Philippine Institute for Development Studies Past President, Philippine Statistical Association Fellow, Social Weather Stations email: jalbert@[Link] ; jrgalbert@[Link]
Agenda
Answering Questions Commonly Used Statistical Methods What is Statistical Analysis?
Descriptive Statistics Inferential Statistics Statistical Models
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Answering Questions
Many times, we need to answer questions:
What is the pattern of poverty in the Philippines? What is the current level of livestock production? How do SMEs operate? Why are the poor poor?
Research entails answering questions with the use of scientific methods
Statistical Research and Training Center Statistical Analysis with STATA May 16 20, 2011
Answering Questions
The Research Process
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Answering Questions
Types of Research: Qualitative research : based on info from understanding of the behavior of people and institutions, their values, rituals, symbols, beliefs, emotions. Quantitative research: based on info from surveys, experiments, and administrative records.
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Nature of Statistical Analysis
Purposes of Statistical The world before Analysis
analysis The world after analysis
USE DATA to Describe Data collection Explain Predict Make Decisions
Data interpretation
Data organization & preliminary analysis
Statistical Analysis with STATA May 16 20, 2011
Statistical Research and Training Center
7
I HOPE THIS ISNT STATISTICS
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Statistical analysis
Come on! It cant go
wrong every time...
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Data Description
Collect Data
Present Data
e.g. Survey e.g. Tables and Graphs
10
Characterize Data
e.g. Mean
xi
A Characteristic of a: Population is a Parameter Sample is a Statistic.
Statistical Research and Training Center Statistical Analysis with STATA May 16 20, 2011
Explaining/ Predicting Relationships
11
When explaining relationships with statistical models, we make conclusions and decisions about a large class of data (called a POPULATION) using SAMPLE results. This is called INFERENCE and involves
ESTIMATION HYPOTHESIS TESTING
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Method / Model
T-test Median-test Binomial-test Chi square goodness of fit test
Statistical Research and Training Center
12
Dependent Variables
Quantitative interval Quantitative interval or ordinal categorical (2 categories) categorical
Independent Variables
None (one population) None (one population) None (one population) None (one population)
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Method Dependent / Model Variables Independent Variables
13
2 independent sample One DV (Quantitative One IV with two T-test interval) levels (independent groups) Wilcoxon-Mann Whitney test Chi square goodness of fit test Fishers exact test
Statistical Research and Training Center
One DV (Quantitative One IV with two interval or ordinal ) levels (independent groups) One DV (categorical) One IV with two levels (independent groups)
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Method Dependent / Model Variables
Oneway ANOVA
14
Independent Variables
One DV (Quantitative One IV with 2 or interval) more levels (independent groups) One DV (Quantitative One IV with 2 or interval or ordinal) more levels (independent groups) One DV(categorical) One IV with 2 or more levels (independent groups)
Kruskall Wallis
Chi square test
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Method Dependent / Model Variables
Paired t-test
15
Independent Variables
One DV (Quantitative One IV with 2 levels interval) (dependent/ matched groups)
Wilcoxon signed rank One DV(Quantitative One IV w/ 2 levels test interval or ordinal ) (dependent/ matched groups)
Mcnemar test
One DV(categorical)
One IV w/ 2 levels (dependent/ matched groups)
Statistical Analysis with STATA May 16 20, 2011
Statistical Research and Training Center
Common Stat Methods
Statistical Method Dependent / Model Variables
one-way repeated measures ANOVA Friedman test
16
Independent Variables
One DV (Quantitative One IV w/ 2 or more interval) levels (dep/ matched groups) One DV(Quantitative One IV w/ 2 or more interval or ordinal) levels (dep/ matched groups)
repeated measures logistic regression
One DV(categorical)
One IV w/ 2 or more levels (dep/ matched groups)
Statistical Analysis with STATA May 16 20, 2011
Statistical Research and Training Center
Common Stat Methods
Statistical Method Dependent / Model Variables
factorial ANOVA
17
Independent Variables
One DV (Quantitative two or more IVs interval) (independent groups) One DV (categorical) two or more IVs (independent groups)
Factorial logistic regression
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Dependent Method / Model Variables
Correlation Simple linear regression Nonparametric correlation simple logistic regression One DV (Quantitative interval)
18
Independent Variables
One IV
One DV One IV (Quantitative (Quantitative interval or ordinal) interval or ordinal) One DV (Categorical ) One IV
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Method Dependent / Model Variables
Multiple regression 1 DV (Quantitative interval)
19
Independent Variables
1 or more interval IVs and/or 1 or more categorical IVs 1 or more interval IVs and/or 1 or more categorical IVs
Analysis of covariance
multiple logistic regression discriminant analysis 1 DV (Categorical)
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Dependent Method / Model Variables
Multivariate Regression Multiple Analysis of Covariance (MANCOVA) Multiple Analysis of Variance (MANOVA) Two or More DVs (quantitative interval) Two or More DVs (quantitative interval) Two or More DVs (quantitative interval)
20
Independent Variables
Two or more IVs (emphasis on quantitative) Two or more IVs (emphasis on qualitative) Two or more IVs (all qualitative)
Statistical Research and Training Center
Statistical Analysis with STATA May 16 20, 2011
Common Stat Methods
Statistical Method Dependent / Model Variables
Factor Analysis Two or More DVs (all quantitative interval)
Two Sets of Two or More DVs (all quantitative interval)
21
Independent Variables
None
Canonical Correlations
None
Generalized Repeated Quantitative variables Two or more IVs Measures Model measured at different (some quantitative points in time some qualitative)
Statistical Research and Training Center Statistical Analysis with STATA May 16 20, 2011
STATISTICAL RESEARCH AND TRAINING CENTER
J and S Building, 104 Kalayaan Avenue, Diliman, Quezon City
Training Course on
Statistical Analysis with STATA
May 16 - 20, 2011
Moving on to Using Stata