|
|
|
|
![](../../teaching.jpg) |
![](logo.webp)
This page aims to help scientists to design and analyze their experiments. It is mainly focused on Biomedical research, although not exclusively, and it is used and created by the students of my courses. You may find helpful my free online book, my Youtube channel, and my slides.
The topics covered in this blog are basic statistics, experiment design, and sample size calculation.
If your experiment is important, then the animals involved, funders and the general public deserve that you design and analyze it carefully.
Statistics is like a calculator. You need it to extract information from your experiments. Statistics is not more important than the science behind your experiment, but if you want to use this calculator, you must follow its "user guide instructions". On the other side, using this calculator will prevent you from missing important effects (false negatives) and being fooled by noise (false positives). You may feel comfortable using the standard tools you have always used for many experiments and that they truly apply to your experiment. In some other cases, the needed statistical tools go beyond the tools you use daily. You should look for further statistical help if you feel uncomfortable with the design or analysis of your experiment results. Remind that training is essential to keep up fit and be able to use statistical tools correctly. Nature did not select us because we were good at solving statistical problems. Statistics is a by-product of our evolution and is not easy for anyone. Don’t feel bad if it escapes your intuition. Mastering it is a matter of being exposed to it, time, patience and learning.
Hypothesis tests
Sources of systematic errors (causing biased results)
Statistical significance and power
Type 1 (false positive) and Type 2 (false negative) errors
Type 1 (false positive) and Type 2 (false negative) errors (2)
P-values and confidence intervals
Bonferroni correction to avoid the inflation of false positives by many simultaneous tests
Multiple testing and how to avoid the inflation of false positives
Chi-squared test (independence of two categorical variables)
Fisher's exact test (independence of two categorical variables)
1-way ANOVA (equality of the mean of several groups)
Normality tests
Survival analysis
Basics
Interpretation of hazard ratios, Kaplan-Meier curves, and proportional hazards
Clinical and biomedical applications
Case-control studies
Case-control studies (2)
Case-control and cohort studies
Use of Statistics in Genome Wide Association Studies (GWAS)
Use of Statistics in genetic crossovers
Use of Statistics in clinical trials
Statistical significance and clinical relevance
Bayes theorem and clinical tests
Stratification in clinical trials
Non-inferiority clinical trials
Other topics
Bayes theorem
Boxplots to summarize the data
Correlation coefficient: interpretation and alternatives
Spearman's rank correlation
Homocedasticity and heterocedasticity
Assessing the consistency of measurements
Sample size calculation in clinical trials(difference of means and proportions)
Examples from the master in Neuroscience 2023
|
|
|