Presenter: Dr Shahid Ullah (FCEB) 

 DateTopic 

Tuesday 12th June 2012

Lecture recording is not available due to the technical error.

Slide (PDF 1MB)

Causality

Righteous, sensible, and durable inference from complex observational data in health sciences requires the careful formulation, implementation, and interpretation of statistical models in terms of causal effects. The research questions that motivate most studies in the health sciences are causal in nature. For example, what fraction of deaths from given disease could have been avoided by a given treatment or policy? Does 2km walk test with high body mass improve the prediction of V02max? Not surprisingly, the target of such questions is the elucidation of cause-effect relationships among variables of interests, for example, outcomes, exposures and confounders. The statistics lecture is designed to help you to develop a framework for accurate cause-effect relationships among those variables with appropriate model diagnostics. The lecture will also give you the syntax to perform variety of statistical models in R software. 

R logoTo download R, please visit http://www.r-project.org/. Please visit our learning resources page (Learn R) for more information.

Tuesday 22nd May 2012

Audio/Video

Slide (PDF 1MB)

Statistical tests for health sciences

Choosing the right test to compare measurements is a bit tricky. Many statistical tests are based upon the assumption that the data are sampled from a Gaussian distribution. These tests are referred to as parametric tests. Tests that do not make assumptions about the population distribution are referred to as nonparametric tests. They also depend on the nature of the outcome and explanatory variables analysed. The lecture is designed to help you to develop a framework for choosing the correct statistical test to your research question(s)/hypothesis. The lecture will also give you the syntax to perform variety of statistical tests in R software.

R logoTo download R, please visit http://www.r-project.org/. Please visit our learning resources page (Learn R) for more information.

Tuesday 24th April 2012

Audio/Video 

Slide (PDF 903KB)

Study Design

There are many different approaches to research, and many different study designs are used to answer different research questions. In fact, the study design depends on the question being asked, and the priority goes to the study which should be comparative and seek to avoid all potential causes of biases. The most valuable thing a researcher can possess is knowledge of the principles of good study design. There is no point to analysing data from a study that was not correctly designed to answer the research question. In reality, there's a valid point in refusing to analyse such data lest faulty results be responsible for implementing a program or policy contrary to what's really needed.

 

Tuesday 20th March 2012 

Audio/Video

Slide (PDF 1MB)

Introduction to Statistics for Health Sciences & Free Software 

Concepts and procedures in statistics are inherent to publications in Science. Most medical and health science journals have policies that govern the reporting of statistical procedures and results. In an effort to improve the calibre of statistical information, academics and postgraduate students are definitely needed to consult with smart statisticians to avoid critical statistical comments from journal reviewers and PhD examiners. Statisticians can contribute to good research by improving the design of studies as well as suggesting the optimum analysis of the results. The aim of the lecture is to give you the clear guidelines of using statistics in medical and health sciences research. The lecture will also give you the free software to analyse the data in a smart way.

 

R logoShahid Ullah introduced a free statistical/graphical software, R. To download R, please visit http://www.r-project.org/. Please visit our learning resources page (Learn R) for more information.

 

Thursday 10th November 2011

Audio/Video 

Slide (PDF 1MB)

Sample Size Calculation and Power Analysis

Reporting of the sample size calculation has greatly increased in the past decades in the health and medical sciences. It has been challenging that how many participants we should include to detect a clinically relevant treatment effect. The oversized trials, which expose too many people to the new therapy, or underpowered trials, which may fail to achieve significant results, should be avoided. In fact, the number should not be too small (under powered) or too big (huge cost/time involved). We should select an optimum number of cases. Learn the statistical methods in this 1 hour biostatistics lecture for determining this magic number (sample size).

 

gpicon-128.pngShahid Ullah introduced a sample size calculator G*Power3 during this lecture. G*Power3 is available from http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/. Please refer to http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/user-guide-by-design to find out how to use G*Power to calculate sample size and power by study design.