Aim
To be able to carry out common statistical procedures and interpret the results
 Assumes no previous experience in statistics
 Limited to few procedures, introductory
 Reduced I/O to ease learning
 Link to StatTools index,
list of procedures required in a busy clinical research environment.
Module Contents
Statistical Procedure
Survey statistics : Statistics with 1 group
 Population Mean
 Population Proportion
 Correlation, parametric and nonparametric
 Regression
Comparison between two groups
 Compare two sets of measurements, parametric and nonparametric
 Compare two proportions using different models
 Risk Difference
 Risk Ratio
 Odds Ratio
 Compare two regressions : covariance
Metaanalysis
Prediction
 True and False, Positive and Negative, Rates
 Likelihood Ratios, Test Positive and Test Negative
 Use of Test Results, Pretest and Posttest Probabilities
 Receiver Operator Characteristics (ROC)
Theories
Basic introductions only
 Normal distributions : z, t
 95% confidence interval : one tail, two tail
 Mean, Standard Deviation
 Effect and its Standard Error, Differences and Ratios
 Null Hypothesis, statistical significance
 One and Two tail models
 Sample size
 Nonparametric statistics
Graphics
Macroplot :
 Advantages and Disadvantages
 Why Macros
 Forest Plot
Plots :
 Pie Chart
 Bar Chart
 X/Y Scatterplot
 Plots associated with statistical procedures
 Data Plot
 Forest Plot
 ROC Plot
 Regression Line Plot
Assessment
Two levels
 Part I Basic Level
 Part II Advanced Level
Grading :
 D : Satisfactory completion of Part I by deadline
 C : Completion of Part I with <4 resubmissions
 B :
 Maximum 30% of those with C
 Completion of Part I and II
 Priority depends on number of resubmissions, and time of final submission
 A :
 Maximum 30% of those with B
 Completion of Part I and II with first submission
 Priority depends on number of resubmissions, and time of submission
Timing :
 Before end of March, 2018
Teaching
Web Site, contains all necessary information
Guided studies and exercises
Email to Prof. Chang, Q&A page
Workshops
Assessment templates
How to Study
Principles of Studying Statistics
Knowledge : Accumulation of information and understand its structure and relationship
 Acquire by seeking and memorising information
 Refined by considering comparing different opinions
Skill : Able to produce results
 Acquire by copying successful tactics and strategies
 Refined by practice
Wise saying :
 To hear is not as good as to see
 To see is not as good as to do
 To do is not as good as to make mistakes
Statistics :
 Require knowledge as a background for understanding
 Requires making decisions on choosing procedures and interpretation of results
 Handles approximations and ambiguities
 Therefore more skill than knowledge
Advice to Students
Accept Statistics is mostly skill
It cannot be taught, it can only be learnt by repetitive problem solving, performing and interpreting statistical procedures
Do not try to remember theories. Do the procedures, and look up the theories
whenever something difficult to understand is encountered.
Use modern facilities. Send email to ask questions instead of spending hours searching through
text books
