| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Board of Management and Trustees of the British Journal of Anaesthesia [2007]. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
Statistics III: Probability and statistical tests
1 Consultant
Department of Anaesthesia
Stockport NHS Foundation Trust
Stepping Hill Hospital
Stockport, SK2 7JE
UK
2 Specialist Registrar
Department of Anaesthesia
Royal Lancaster Infirmary
Ashton Road
Lancaster, LA1 4RP
UK
Tel: +44 161 419 5869 Fax: +44 161 419 5045 E-mail: a.mccluskey4@ntlworld.com
Key Words: The laws of probability dictate how typical a sample dataset is of the population from which it is drawn. • Which statistical test to use to analyse a dataset depends on a number of considerations including the type of data being analysed (e.g. interval or categorical), whether interval data are normally distributed or not and whether data are independent or paired. • Student's unpaired and paired t-tests are used to compare two groups of normally distributed independent and matched groups, respectively. • Analysis of variance (ANOVA) and repeated measures ANOVA are used to compare three or more groups of normally distributed independent and matched groups, respectively. • There are non-parametric equivalents of all the above tests. • Categorical data are compared by drawing-up a contingency table and applying either Fisher's exact or 
2 tests.
| The first 150 words of the full text of this article appear below. |
 |
Probability theory |
|---|
The probability of an event may be determined empirically (by observation) or mathematically (using probability theory). Probability theory is fundamentally important to inferential statistical analysis. Predicting population parameters from sample data is based on the assumption that the sample data are ‘typical’ of the population data. The laws of probability govern just how typical the data are. For example, we may toss a coin 20 times to determine the likelihood of obtaining heads on a single throw. Common sense tell us that, provided the coin is unbiased with heads just as likely to fall as tails, the ratio of heads:tails should be 1:1 and therefore the ‘expected’ outcome after 20 tosses would be 10 heads. However, the actual outcome may well be different. If we were to repeat the experiment by tossing the coin 1000 times, it is likely that the ratio heads:tails would be very close to 1:1 and
|
The binomial distribution |
|---|
|
Statistical inference |
|---|
One group of interval data
Estimation of the population mean from sample data
One sample t-test
Wilcoxon rank sum test
Comparing two groups of interval data
Unpaired (independent) normally distributed data: Student's unpaired two-sample t-test
Paired normally distributed interval data: Student's paired two-sample t-test
Non-parametric interval data
Three or more groups of interval data
Categorical data
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. McCluskey and A. G. Lalkhen Statistics IV: Interpreting the results of statistical tests CEACCP, December 1, 2007; 7(6): 208 - 212. [Full Text] [PDF]
|
|