What are the statistics?
Statistics is the science and art of making sense of data. Biostatistics compromise of statistical techniques which are most useful for health-related data. Biostatics can be regarded as the "tool kit" which is the basis for understanding and assessing the evidence for medical treatments and health interventions.Statistics is a scientific discipline consisting of procedures for collecting, describing, analysing and interpreting numerical data. Biostatistics compromises of the statistical techniques most appropriate to health-related data.
"Scientific discipline"
"numerical data" as opposed to qualitative data.
It is possible, though perhaps not very useful, to study statistics simply as part of mathematics wiuth no concern for its application at all" "Statistics may be discussed without appearing to use any mathematics at all" - Martin Bland - the leading Biostatistics Teacher.Statistics are a brand of mathematics and are applied to real life data. Statistics are about taking data to information to knowledge to understanding.
- Data are not information
- Information is not knowledge
- Knowledge is not understanding
- Understanding is not wisdom - Clifford Stoll
Hospital data and death registry data are sometimes collected for another purpose so you have to be aware of the limitations you place on data. Hospital data may be used for the financial year update as well as keeping a patient history to track ailments throughout a lifetime.
Data sources
- Survey specific collection - relevant, targeted and detailed. Expensive, slow but quicker now with online survey tools.
- Routinely collected information for example hospital data, death registry data. There are some limitations including the relevancy, level of detail and currency.
- Aggregated from multiple sources - limitations include consistency of collection, the accuracy of links. For example, data from multiple hospitals may have inconsistencies in data collection due to procedural differences.
Data dredging
A traditional analysis proceeds by forming a hypothesis and investigating whether or not the data support the hypothesis
Data dredging is a much better way of collecting data and involves sifting through the data looking for any relationship that might appear to be there. Some of the data may be there purely by chance. Using data dredging you run the risk of finding coincidences.
1. Population and sample
- How to describe the concepts of the population sample and sample units.
- Critique a sampling method in terms of bias
- Explain the process of sampling and its role in providing information about a population.
Hint: SAS and SPSS are the most common statistical programs. It would be good to have these on your resume however, they do require an annual licence!
Why do we need statistics?
We use statistics as a tool, a method of communication and statistics do not stop with the calculations. The can express the question in testable format, can be analysed, interpreted and applied to gain a result. Statistics can be used to find how big of a difference between the results is significant.
Statistics can be misleading.
"There are three kinds of lies; liesm damned lies, and statistics" - attributed to Benjamin Disraeli"Statistics are like a drunk with a lampost: used more for support than illumination" - Winston Churchill
"I only believe the statistics that I doctored myself" - Winston Churchill
The art and science of statistics is all about allowing the data to tell a truthful story. Find ways to analyse the data without being misleading and it can be used to tell great truths.
What is the best way to analyse statistics? Stata!
Stata is a software used for statistical analysis. Used to manage and analyse your data as well as:- Data entry, saving data files, compute summary statistics
- Simple data interrogation e.g. checking variable ranges
- Simple hypothesis testing e.g. comparing means, tests of association
- Cheap, compared to other packages (eg SAS, SPSS)
- You get a permanent licence. You only need a new licence if you choose to upgrade to a later version. Unlike SAS, SPSS which require annual licences.
- Easy to use for simple analyses but capable of sophisticated, cutting edge analyses.
- Stata/IC is the small Stata version
Comments
Post a Comment